~-.- .7 ABSTRACT AUTOMATED NIETABOLIC PROFILING OF ORGANIC ACIDS IN HUMAN URINE BY GAS CHROMATOGRAPHY-MASS SPECTROMETRY by Stephen Carl Gates The impetus for this research was the need to deve10p a method capable of both identifying and quantitating the large number of compounds typically present in biological fluids. The organic acid fraction of human urine was selected to test the resolution of the system, since at least 60 to 80 peaks, most of which are only partially resolved, are visible in this fraction when it is separated by gas chromatography. The method chosen to measure the analyte pattern (“metabolic profile”) of urinary organic acids was separation on DEAE-Sephadex, followed by analysis of the derivatized acid fraction on a gas chromatograph-mass spectrometer-computer system. Data are collected with repetitive scanning of the magnetic field of the mass spectrometer accompanied by temperature -programmed gas Stephen Carl Gates chromatography. The repetitive seaming data are then analyzed off-line by the mass spectral metabolite program (MSSMET). This program utilizes a reverse library search of the mass spectral data. Gas chromatographic retention indices are used both to limit the amount of data searched and to aid in identification of the substances. The intensities of a small set of pre-selected ions are used to judge the degree of match between each library spectrum and the experimental spectra. Compounds are quantitated by measuring the height and area of the peak of the “designate” ion, which is the ion which has been selected as being most likely to be differentiating for that compound. The relative concentration of each compound is calculated from the ratio of the area (or height) of the designate ion of the compound to the area (or height) of the designate ion of the internal standard. This relative concentration can be converted to absolute concentration by the application of an appropriate correction factor determined experimentally for each compound. For each substance positively identified as matching a 1ibrary spectrum, MSSMET prints out information including name, concentration, degree of match to library spectrum, retention index, and retention time. Stephen Carl Gates MSSMET was tested on spectra from a variety of pure compounds and urinary organic acid samples. All samples were analyzed as the trimethylsilyl derivatives on 10 ft 5% OV-1'7. These tests indicated that the retention indices were extremely precise (better than 0.2%) and that the compounds could be identified accurately almost to the limit of detection of the compound (usually to about 10 to 20 ng injected). Quantitation was linear over a 500 to 1000-fold range; this range was not extended by the use of isotope dilution techniques. Precision of the repetitive scanning technique was approximately 3% on isotope ratio determination, 5% on relative area determination with pure compounds, and 8% on relative area determination with urine samples. Once MSSMET had been validated, the same techniques were applied to a variety of urine samples. These included urines from 9 “healthy” adults, 6 hospitalized children, and 5 children being treated for neuroblastoma. An average of 100 t 30 compounds were reliably identified and quantitated in each urine sample, with up to 32 more compounds found less reliably. Statistical analysis of the reliably-found compounds indicated that the distribution of compound concentrations in adults was generally log-normal. Levels of twenty compounds were found to be significantly different Stephen Carl Gates (p (.10) between the adult and juvenile groups, while levels of 13 compounds were significantly different in the neuroblastoma urines compared to the other two groups. Differences at each level of significance tested (0.10, 0.05, 0.01, and 0.001) were found to considerably exceed those expected by chance. The concentration of one substance (caffeic acid) was also found to be related to the prognosis for survival of the neuroblastoma patients, although this finding is very tentative. AUTOMATED METABOLIC PROF ILING OF ORGANIC ACIDS IN HUMAN URINE BY GAS CHROMATOGRAPHY—MASS SPECTROMETRY by Stephen Carl Gates A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1977 $10705! ACKNOWLEDGMENTS In a project of this magnitude, a great many individuals necessarily make significant contributions to its success. Foremost among those whom I would like to acknowledge for their help in this project is Dr. Charles C. Sweeley, my thesis adviser. Dr. Sweeley not only suggested the original idea behind this project, but also has provided a constant flow of useful ideas and suggestions throughout my graduate career. He has provided an extremely well-equipped mass spectrometry laboratory for me to use in my research and has been especially helpful in teaching me to ask the proper analytical questions. I have also appreciated the opportunities he has given me to pursue ideas related to, but not directly contributing to, the development of MSSME T. I also owe a special debt of gratitude to Nancy Dendramis for her help in the processing of urine samples and to Jack Harten for his helpin the use of the LKB-9000. I have greatly benefitted from the example set by Bernd Soltmann, who has helped everyone in the laboratory to become more expert on the technical details of mass spectrometry. Bernd, Norm Young and Frank Martin have been extremely helpful in keeping all of the equipment in the mass ii spectrometry facility repaired. Three programmers--Norm, Mike Smisko and Curt Ashendel--all made very significant contributions to the development of MSSME T, and I am appreciative of the strong foundations they built during the design and construction of early versions of this program. A number of individuals also helped in planning and implementing the statistical analysis of the MSSMET data. Bob Wilson in the College of Education and Bill Brown in the Computer Center both provided valuable advice in planning statistical studies. Dr. B.E. Blaisdell in Dr. Sweeley’s laboratory generously allowed me to use one of his statistical analysis programs and provided several useful criticisms of the MSSMET library. Furthermore, I would like to thank Dr. John F. Holland for his general guidance, as well as his advice in technical matters. I sincerely appreciate the exposure to Dr. Arthur Kohrman’s contagious enthusiasm for medical research, as well as his help in keeping a balanced perspective about the goals and needs of such research. I genuinely appreciate the efforts of the personnel at the Bronson Clinical Investigation Unit in collecting the adult reference urine samples, and the generosity of the Upjohn Company in donating the urine samples to us. iii . 'rr‘ W’s-w— Dr. William Krivit at the University of Minnesota has provided the urine samples from several groups of children; I sincerely appreciate his generosity in providing us with the samples and data on the subjects. My father, Harry Gates, deserves special thanks for his superb job in typing the final draft of the dissertation. His typing this dissertation has made life considerably easier for me. Finally, I would like to thank the one person most influential in my completion of this research—-my wife, Jackie Cresswell. She not only provided encouragement throughout the five years, but also put up with a husband who frequently seem ed to spend more time with MS. Met than he did with her. iv TABLE OF CONTENTS Page LIST OF TABLES ............................ x LIST OF FIGURES ........................... xii INTRODUCTION ....... . ................... 1 CHAPTER ONE : LITERATURE REVIEW ............ 5 Development of the concept of metabolic profiling. . . 5 Techniques for separating low—molecular- weight components of biological fluids ...... 9 Paper and thin—layer chromatography . . . 9 Column chromatography ............ 9 Gas chromatography .............. 11 Gas chromatograph-mass spectrometer— computer systems ............ 12 High resolution GC and GC-MS ....... 17 ‘ Computer processing of GC-MS data for metabolic profiling ................... 20 . Reverse search methods ........... 20 Retention indices ................ 22 Other GC—MS-COM techniques for profiling .................. 23 Disease diagnosis by metabolic profiling ......... 26 Non-statistical methods ............ 26 Statistical treatment of data ......... 30 TABLE OF CONTENTS (Cont’d.) Page CHAPTER TWO: MATERIALS ................... 37 Reagents .............................. 37 Chromatography packings and supplies .......... 39 Glassware ............................. 40 Instruments ............................ 40 Data systems . . . . . ...................... 41 Computer programs ...................... 41 Miscellaneous supplies .................... 42 CHAPTER THREE: METHODS .................. 43 Collection of urine samples ................. 43 Standard reference urine ........... 43 BCIU collection ................. 43 Juvenile reference and neuroblastoma ' urines ................... 44 Foy urine ..................... 44 Questionnaires .......................... 45 Isolation of organic acids from urine ........... 45 Preparation of columns ............ 45 Creatinine determination . . ......... 46 Preparation of urine sample . . . ...... 47 Gas chromatographic analysis ............... 50 Analysis of samples on the LKB-9000 GC-MS- COM system ....................... 53 Preparation of samples for preliminary studies . . . 56 Capillary stability ........ . ....... 56 Urine stability .................. 57 Silylating solvent ................ 5'? Recovery study ........ . . . . . . . . 58 Linearity and isotope dilution series . . . 60 Urine studies. ................. 64 Analysis of data by MSSMET . . .............. 65 Program initiation ............... 66 Reading the library file ............ 66 Location of compounds ............ 75 Location of ion peaks ............. 80 Baseline determination . . . ......... 84 Calculation of peak amount .......... 85 vi TABLE OF CONTENTS (Cont’d.) Printing of results ............... Selection of designate and confirming ions ........ MSSMET analysis of urine samples ............ Statistics ............................. MSSTAT ..................... FRGENL ..................... Clinical report form .............. CHAPTER FOUR: RESULTS .................... Evaluation of MSSMET and GC-MS- COM system . . . Data collection reliability .......... Data collection parameters ......... Library spectra ................. Calculation of retention indices and match coefficients by MSSMET . . . Mass chromatogram peak detection by MSSMET ................. Sensitivity and linearity .of system response . . . . .. ........... Precision of retention indices ........ Match coefficient reliability ......... Reproducibility ................. Quantitative precision ............ Quantitative accuracy ............. Tests of the urine separation procedure ......... Recovery ................... . Reproducibility ................ . Sample stability ................. Silylating mixtures ............... Clinical studies ......................... MSSMET analysis of urine samples . . . . Analysis of procedural blank ........ Statistical analysis of urine sample . . . . Retention indices . . . . ............ Match coefficients ............... Peak width in urine samples . . . ...... Relative peak areas and heights ...... Distribution of peak amounts ........ Outlier test . . ....... . .......... Page 91 91 99 100 100 10 1 102 103 103 103 104 107 107 107 108 108 108 108 117 123 123 123 124 124 132 132 133 134 134 136 141 143 143 150 150 TABLE OF CONTENTS (Cont’d.) Comparison of subject groups ....... Clinical report form ............. CHAPTER FIVE: CONCLUSIONS AND DISCUSSION Organic acid separation procedure ............ Reproducibility ................. Other studies of extraction procedure . . Procedural blank ................ Choice of GC column type and derivatizing agent . . . Use of repetitive Scanning data . . . . .......... GC conditions .......................... Data storage and transfer .................. Time required for GC- MS analysis ............ Performance of MSSMET ................. MSSMET baseline determination and peak detection .............. Speed ....................... Accuracy ..................... Retention indices ........................ Precision of retention indices ....... Accuracy of retention indices ....... Match coefficient ........................ K-factors ............................. Ion intensity variability . . . ................ Comparison of MSSMET to SIM .............. Ease of operation of MSSMET ............... Studies on urine samples .................. 1 Selection of subjects ............. Selection of MSSMET library for urine samples .................. Selection of designate and confirming ions from urine samples ....... GC-MS analysis of urine samples ..... MSSMET analysis of urine samples . Number of substances found ........ Statistical analysis of urine data ........ . . . . . Distribution of concentrations ....... Distribution of retention times ....... Distribution of retention indices ...... Interitem correlations ..... . ...... viii Page 157 158 171 171 171 174 176 178 180 184 185 186 187 188 189 189 191 191 191 197 199 203 206 209 211 211 212 214 220 221 226 228 229 230 234 236 o 4 g n o b c I a e I I. o I t O c c l O O O I a t v m I TABLE OF CONTENTS (Cont’d.) Outlier test ................... Comparison of subject groups ....... Clinical report form ...... . ....... Other statistical considerations ...... CHAPTER SIX: EVALUATIONS AND RECOMMENDATIONS APPENDICES Chemical separation procedure .............. Quantitative precision . . .................. Other recommendations ................... Speculat1on on long-term prOSpects . ......... Appendix A: Diet, health and drug questionnaire . . . Appendix B: BESTLIB (MSSMET library) ....... Appendix C: Compounds excluded from statistical calculations ............ Appendix D. Summary of urine samples analyzed. Appendix E: Relative areas calculated by MSSMET. Appendix F: Normalized relative areas ........ Appendix G: Substances omitted from summation during normalization ......... Appendix H: Clinical report form ............ Appendix I: Complete MSSMET “found” file . . . Appendix J: T-test of log10 of normalized data and tabulation of compound means, standard deviations, standard errors and coefficients of variation ....................... Appendix K: List of publications ............. REFERENCES ............................. Page 238 242 250 251 258 259 261 263 270 273 279 288 292 296 305 314 317 321 327 342 344 LIST OF TABLES TABLE 1. 2. 3. 10. 11. 12. 13. 14. 15. Silylating mixtures ...................... Composition of recovery study solutions ........ Recovery study samples .................. Composition of linearity and isotope dilution series Compounds monitored by SIM during linearity study . Precision of isotope ratio determination ........ . Accuracy of isotope ratio determination ........ . Recoveries of organic acids using barium hydroxide-DEAE-Sephadex method ........ . Analysis of procedural blank ............... Correlation of MSSMET output data ........... Test for outlying substances ................ Substances differentiating urines of adult subjects from urines of juvenile subjects .......... Substances differentiating urines of subjects with neuroblastoma from urines of all control subjects .......................... Comparison of three statistical tests .......... Comparison of Student t-test on log—transformed data with Wilcoxon test ................ Page 58 59 59 62 64 122 122 125 142 155 159 161 163 167 LIST OF TABLES (Cont’d.) TABLE Page 16. Comparison of literature and experimental retention indices .................... 193 17. Number of compounds found in urine by MSSMET . . 227 18. Agreement between peak area match coefficient and peak height match coefficient ......... 239 D1 Summary of urine samples analyzed ........... 293 E1 Relative areas calculated by MSSMET . . . ....... 297 F1 Normalized relative areas .................. 306 11 Complete MSSMET “found” file .............. 322 J1 T-test and mean values .................... 329 xi LIST OF FIGURES FIGURE 1. Diseases where organic acids have been detected at abnormal levels by GC—MS ..... 2. Generalized MSSMET flowchart ............. 3. Detailed MSSMET flowchart ................ 4. Typical MSSMET library entry .............. 5. Determination of retention index window ........ 6. Formula for calculation of match coefficient by MSSMET ....................... 7. Criteria for positive match to library entry ..... 8. Detection of mass chromatogram peaks ........ 9. Baseline determination I .................. 10. Baseline determination H .................. 11. Baseline determination III .............. A. . . 12. Determination of relative peak area ........... 13. Formula for calculation of peak amount by MSSMET 14. Typical MSSMET output ................... 15: Formula for calculation of ratio by MSSDSG ..... 16. Typical MSSDSG output .................... xii Page 27 68 70 74 77 79 81 83 87 87 87 89 90 93 96 98 LIST OF FIGURES (Cont’d.) FIGURE Page 17. Minimum criteria for LKB-9000 operating parameters . . . . ............... . . . . 106 18. Quantitative working curves measured by repetitive scanning . . . . . . . . . . ..... . . . 110 19. Quantitative working curve at high concentrations measured by repetitive scanning . ........ 112 20 Precision of retention index determination on pure compounds ........ . ......... 114 21. Dependence of match coefficient upon amount of sample injected . . ....... . ..... . . . . 116 22. Reproducibility of repetitive scanning GC—MS on urine samples ......... . .......... 119 23. Dependence of GC- MS reproducibility upon retention index .......... . . ..... . . . 121 24. Reproducibility of analytical procedure on urine samples . . . . . . . ......... . ..... 127 25. Stability of stored urine samples ............. 129 26. Stability of stored sample capillaries .......... 131 27. Precision of retention index determination on urine samples using hydrocarbons as standards . . . 138 28. Precision of retention index determination on urine samples using metabolites as. standards . . . . 140 29. Distribution of mean match coefficients . . . . . . . . 145 30. Distribution of individual match coefficients . . . . . 147 31. Distribution of designate ion peak widths . . . . . . . . 149 xiii LIST OF FIGURES (Cont’d.) FIGURE 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. Distribution of concentrations in a group of reference urine samples ............... Distribution of logarithms of concentrations in a group of reference urine samples ......... Determination of the quantitative correction factor (k) by two methods .............. Quantitative working curves produced by selected ion monitoring ............... Distribution of ions in a typical urine sample GC-MS analysis ........... . ........ Dependence of ion distribution upon region of GC-MS analysis ............ . ....... GC analysis of typical BCIU urine sample ....... GC analysis of typical juvenile control urine sample. GC analysis of typical neuroblastoma urine sample . Variability of retention times of retention index standards .................... . Concentrations of VMA and HVA in four types of urine samples ................. . . . Concentrations of caffeic and m—hydroxyphenyl- hydracrylic acids in four types of urine samples .......................... xiv Page 152 154 202 208 216 218 223 223 223 232 254 256 :- 2 a , c a u I I C i l‘ O o n u o a o I u u s o . o n o I - o p o . - a o a . . . , . . A s t u o n u . - l . . 1 . . . . u A I c I I n a a . n u . o a o LIST OF FIGURES (Cont’d.) FIGURE Page A1 Diet, health and drug history questionnaire . . . . 274 Bl BESTLIB (MSSMET library) ............. 280 C1 Compounds excluded from statistical calculations 289 G1 35 compounds omitted .................. 314 G2 12 compounds omitted .................. 315 H1 Clinical report form .................. 318 XV INTRODUCTION Man has long been interested in the possibility that his urine might provide clues to the state of his health or even of the world around him. From early recorded history, there are references to the uses of urine, and by Roman times a whole body of “knowledge” had developed concerning its medical importance. In medieval Europe the study of urine was viewed as a means of determining the reason for a person’s ill health or forecasting his fortunes, and professional roving “uroscopists” were often paid for this service. In some modern-day primitive societies, urine remains a fluid believed to have magical or beneficial properties. While most of these early uses of urine are rather non- scientific by modern standards, urine also has had its place in the development of science. For example, some of the earliest dis- coveries in analytical chemistry resulted from attempts to study this easily available fluid. Beginning with ammonium salts and elemental phosphorous, early chemists (and alchemists) isolated or derived a great number of chemicals from urine that were used extensively in early analytical and synthetic studies. A number of organic compounds, including creatinine, allantoin, hippuric acid, 1 2 leucine, cystine, xanthine, and uric acid were first discovered or isolated from urine. Despite these early uses for and discoveries about urine, however, the detailed study of urine did not really begin until the invention of modern separation techniques. In particular, paper, thin-layer and column chromatographic methods made it quite clear that the compounds studied up to that time constituted only a small minority of the constituents to be found in human urine. Wlth the application of gas chromatographic techniques to the separation of the low molecular weight components of urine, the number of known, albeit often unidentified, components grew even further. Despite the plethora of components to study and the lack of techniques for studying more than a few closely-related compounds at one time, certain substances were detected and identified as being positively correlated with human disease states. This led early investigators to hope that compounds could be found that would be diagnostic of a wide variety of clinical abnormalities. Unfortunately, looking for a small number of compounds assoc- iated exclusively with each disease has proven to be an approach with utility only for a relatively few, generally rare diseases, notably the inherited metabolic disorders. Beginning in the 1940’s, a number of laboratories therefore r‘w'V’vs V' ' ' 3 began to develop more general methods for measuring multi- component biological mixtures, especially urine, with the explicit goal of being able to apply these techniques to the diagnosis of human diseases. To date, these efforts have been only minimally successful, probably at least in part because only qualitative, rather than quant- itative, extraction and analysis procedures have been utilized. This problem of detecting a large number of components in urine is well illustrated by the problem of examining the low molecular weight organic acids by gas chromatography. Here, typical low resolution packed column gas chromatographic analysis of urine reveals approximately 30 to 50 clearly distinguishable peaks, but mass spectrometry of these peaks reveals that the actual number of components is over 125. Very few of these peaks are well resolved from neighboring compounds by the gas chromatograph, and it is almost impossible to quantitate more than a few of the most intense peaks with reasonable accuracy. Capillary (high resolution) gas chromatography, on the other hand, while adequately resolving components, presents considerable difficulties in providing enough sample over a long enough period to permit reliable identification by mass spectrometric or other techniques. Hence, the current research was undertaken, using organic acids as the model substances, to determine whether a satisfactory technique could be found which would have the ability to do the following: Utilize gas chromatographic-mass spectrometric data. Reliably identify components of complex mixtures. Resolve all components adequately. Quantitate all components. Detect a wide range of concentrations of components. Be amenable to almost complete automation. Be adaptable to a number of different types of urinary constituents. Provide results suitable for clinical studies. The results of this research are described in the following chapters and in the papers listed in Appendix K. CHAPTER ONE: LITERATURE REVIEW Development of the concept of metabolic profiling The concept that individuals might have a “metabolic pattern” that would be reflected in the constituents of their biological fluids was first developed and tested by Roger Williams and his associates during the late 1940’s and early 1950’s (51W1). Utilizing data from over 200,000 paper chromatograms, many run with techniques developed in his own laboratory for this purpose, Williams was able to show convincingly that the excretion patterns for a variety of urinary components varied greatly from individual to individual, but that these patterns were relatively constant for a given individual. He summarized his findings in 1951 as follows (51W1): “It appears that each individual we have studied has whenever tested exhibited a characteristic pattern of measurements which is distinctive for that individual alone. While there are in every case day-to-day variations in saliva and urine compositions and in taste thresholds, certain items, at least, stand out as grossly distinctive and the patterns as a whole remain nearly constant.” Williams went on to use his methods to examine samples from 5 6 a variety of subjects, including alcoholics, schiZOphrenics and residents of mental hospitals, and produced what he considered to be very suggestive evidence that there were characteristic metabolic patterns associated with each of these groups (51W1). Williams’ work, however, was apparently not duplicated by others, to whom his task must have seemed rather Herculean with few promises of tangible results. Hence, his ideas about the utility of metabolic pattern analysis remained essentially dormant until the late 1960’s, when gas chromatography and liquid chromatography had advanced sufficiently to permit such studies with considerably less effort. Once these techniques became available, however, the rate of progress became extremely rapid. Thus, for example, in 1970 at least three different groups published papers describing multicomponent analyses of biological fluids and referred to the possibility of “considerable differences in excretion patterns of carbohydrates in disease” (70Y1), “personal blood ‘profiles’ ” (70W1) and a “characteristic excretion profile” of organic acids in urine of patients with phenylketonuria (70B1). However, the phrase most often used to describe the chromatographic patterns observed in biological fluids has been “metabolic profile.” This concept was introduced by the Hornings in 1971 (71H2, 71H3). As originally defined, this term meant “multicomponent GC analyses that define or describe metabolic 7 patterns for a group of metabolically or analytically related metabolites” (71H3). Commenting on the potential usefulness of this type of technique, the Hornings continue by suggesting, “Profiles may prove to be useful for characterizing both normal and pathologic states, for studies of drug metabolism, and for human developmental studies.” This definition of metabolic profile has been adopted by some workers essentially unchanged (73W3). Other workers have preferred just the term “profile” to mean the same thing (76M1). Johnson (72G3) has taken a more statistical approach by defining a profile as, “... a vector of numerical values corresponding to measured characteristics or attributes of a given subject. In addition to clinical chemistry measurements, the profile may include measurements on demographic or physical variables such as age, weight, sex, exercise status, etc. Profile analysis is the study of several profiles for the purpose of characterizing the profiles of a given group of subjects or comparing the profiles of a different group.” A number of hospital laboratories have experimented with a related technique, “multiphasic screening,” (reviewed in 71M1) designed to measure multiple components of a single serum or urine sample. The principal difference between multiphasic and 8 profile techniques has been one of technology: multiphasic testing has utilized single tests for each of the components, while profiling has used a single chromatographic run to analyze for multiple components. The underlying similarity of the two techniques is reflected in the recent literature; thus, for example, multiphasic testing was used to develop a “profile” that could differentiate drug- abuse and hospital-staff populations (74M4). Reece (74R1) has used an additional term, “uniphasic synthesis,” for multiphasic screening when the test results are analyzed utilizing multivariate statistical techniques. In general, then, interest in this type of approach is recent enough so that the terminology is in a rapid state of flux. However, in the material that follows, I will use a combination of the Homing and Johnson definitions: metabolic profiling is a means of obtaining, by chromatographic, physical examination, and demographic survey methods, a set of numerical values that can be used to estimate the chemical and health status of a given individual. In discussing metabolic profiling in the following sections, I will arbitrarily limit this definition to exclude studies of small numbers of components (e.g., less than 5 compounds), high molecular weight components (greater than 1000 amu) and all inorganic compounds. 9 Techniques for separating low-molecular-weight components of biological fluids Although the terminology used to describe multicomponent analysis of biological mixtures is new, the techniques used are old. These have included paper chromatography, thin-layer chrom a- tography, gas chromatography, liquid chrom atography, mass spectrometry and a variety of more specialized techniques for unusual types of compounds; in short, almost all of the tools of the modern analytical biochemist have been used in this type of re search. Paper and thin-layer chromatography. Certainly the principal profiling effort utilizing paper chromatographic techniques was that of Williams, as described previously. While a great many other groups were successful in devising means of separating various fractions of urine and other biological fluids, these methods were not generally applied to producing human health profiles. However, paper and thin-layer chromatography have been, and continue to be, important tools in rapid screening procedures for gross excesses or deficiencies of individual components in biological samples (see, for example, the review by Scriver, Clow and Lamm (7383) on screening procedures for aminoacidopathies). Column chromatography. In contrast, column chroma- tography, particularly high-pressure liquid chromatography, 10 has proven to be a quite effective means for obtaining human health profiles. Thus, Young (70Y1) and Jolley and Freeman (68J 1) reported one of the early attempts to use a high-resolution chromatographic apparatus to evaluate the health status of several types of individuals based on the profile of urinary carbohydrates. Young showed that considerable differences in the carbohydrate profile were apparent for several disease states, but that the pattern for a given individual was reasonably constant from day to day. Similarly, Pitt (131. (70P1) and Scott e_t,_al. (70B1, 71M2), both groups from the same lab at Oak Ridge, described a system for analyzing as many as 150 ultraviolet-absorbing substances in body fluids. A minicomputer was used to resolve peaks and store data (7081). Many of the compounds detected by this system were identified by gas chromatography -mass spectrometry (GC-MS) as aromatic organic acids and amino acids (71M2). Later work expanded this system to multiple column operation (72P2), use of a fluorometric detector for indolic compounds (72C5), sequential use of different column types to achieve more rapid separations (7382), and more sensitive detection systems (7381), including fluorescence monitoring of organic acids (73K1). Other laboratories have also developed this type of system, but with computerized data processing (72 Cl) and other types of 11 detectors (74K2, 75K2). A highly sophisticated system of this type has also been developed for computerized analysis of clinical amino acid data (7681, 7682); this system will be discussed in more detail in the section on statistical analysis of data. Gas chromatography. Although proposed as a possible tech- nique in 1941 in a pioneering paper on liquid-liquid partition chromatography (41M1), gas chromatography (GC) was not success- fully applied to complex biological mixtures on a practical basis until the introduction of lightly-loaded liquid phases by VandenHeuvel, Sweeley and Horning in 1960 (60V1). When coupled with the flame ionization or electron capture detectors developed at the same time, and suitable derivatization methods (see reference 73C3 for a history of G0), the number of uses of GC for the analysis of complex organic mixtures rapidly surpassed those of liquid chromatography, for which no comparable “universal” detector was then available and for which the speed of analysis was much slower than that of GC. Early workers were quick to utilize the GC to aid in the diagnosis of disease. Thus, for example, by 1964 Williams and Sweeley (64W2) had published a general procedure for analyzing urinary aromatic acids, gross excesses of which could often be associated with specific diseases. Similar procedures were published in the same volume for other low molecular weight 12 substances (6481). Unfortunately, as numerous workers were quick to discover, the gas chromatograph by itself provided neither sufficient specificity of detector response nor adequate chromatographic resolution to permit unequivocal identification of most peaks in complex mixtures, and quantitative analysis of small to medium— sized peaks proved to be difficult at best. As a result, there were no reported Roger William s-scale attempts to find the more subtle metabolic patterns present in biological fluids. Gas chromatograph-mass spectrometer-computer systems. Since the mid-1960’s, two fairly distinct approaches have been de- veloped to overcome the shortcomings of using low-resolution GC for metabolic profiling: combined gas chrom atograph-mass spectrometer-computer systems (GC-MS-COM) and capillary GC- computer systems. The GC-MS-COM approach will be discussed first. Although individual GC eluates had been transferred manually to mass spectrometers for analysis for some time, and direct coupling of a GC and M8 had been demonstrated in 1959 by Gohlke (59G1) and in 1961 by Henneberg (61H1), the introduction of the molecular separator by Ryhage (64R1) and Watson and Biemann (64W1), both in 1964, as a means of direct transfer of chromato- graphic material to the mass spectrometer allowed rapid VEG‘L‘ l3 qualitative analysis of complex mixtures on a relatively routine basis. Thus, for example, Ryhage’s first published use of the molecular separator was to obtain a profile of fatty acid methyl esters in butterfat (64R1). This was quickly followed by a study analyzing neutral fecal steroids in human subjects (64E 1), and then a virtual flood of papers utilizing GC-MS systems for the analysis of a wide variety of biological, environmental and geOphysical samples. Most of these early papers described uses of the GC—MS that were limited to qualitative analysis of a few peaks in a very small number of samples; researchers were limited principally by the lack of automated data processing equipment. An alternative approach, this time emphasizing quantitative analysis of a very small number of peaks, was developed first by Henneberg in 1961 (61H1) and later, independently, in Ryhage’s laboratory by Sweeley star. (6681) in 1966. “Selected ion monitoring,” originally used for monitoring one and two ions, respectively, provides a means of obtaining highly precise measure- ments on a small number of GC peaks, including those unresolved from neighboring components in a mixture. However, even when expanded to include on-line computer control, more ions, com- puterized data reduction, and multiple-ion set-selected ion monitoring (see, for example, 73H2, 73H3, 73J1, 73W1, 75Y2) this general technique has been limited to measuring a very small 14 number of components of any one mixture, and hence, while extreme- ly useful for other types of studies, has not seen particularly wide use for metabolic profiling. A notable exception to this has been Maume’s group in France (73M1, 73M2) who have successfully examined mixtures of closely related steroids, catecholamines or amino acids by monitoring ions common to whole classes of these compounds. Much more useful for profiling purposes is the computer- based technique called “mass chromatography,” first described by Hites and Biemann in 1970 (70H1). In this technique, complete mass spectra are taken at frequent intervals and the entire data set stored in a computer. After the mass spectral data collection is finished for a given sample, the data are displayed by plotting the intensities of certain key ions for each of the scans during the run. These intensity versus time plots (“mass chromatograms”) can be dis— played for any ion of interest within the entire mass range of the mass spectrometer. Thus, the mass chromatograms are equivalent to the traces generated during selected ion monitoring, except that there is a much lower sampling frequency for any one ion in mass chromatography. This results in a much lower quantitative pre- cision, but a much more generalized quantitative ability for mass chromatography compared to selected ion monitoring, and hence a higher degree of usefulness for profiling studies. 15 An approach that is intermediate between selected ion monitoring and mass chromatography was suggested by Axelson et al. (74A1) and Baczynskyj e_t_al_. (73Bl). In this method, spectra are taken repetitively, but over a much shorter mass range. This tech— nique, however, has been far surpassed in popularity by the select- ed ion monitoring type methods. In instruments with electrostatic mass filters (quadrupoles, dodecapoles), this technique, selected ion monitoring and repetitive scanning can often all be accomplished with essentially the same data algorithms and hardware, so for these types of instruments the distinction among techniques is more custom- ary than meaningful. However, while any one of the three techniques can be, and has been, used for metabolic profiling in its broadest sense, the procedure of performing repetitive scanning of the full mass range, followed by analysis of selected mass chromatograms, has proven the most generally useful for analysis of mixtures con- taining a large number of unidentified components. For this reason, a large number of laboratories has developed GC-MS-COM systems utilizing repetitive scanning-m ass chromatography techniques for performing metabolic profiling. Principal advocates of this approach have been the Hornings at Baylor University and Eldjarn, Jellum and Stokke at the Uni— versity of Oslo. The Hornings described their first studies on metabolic profiling in a pair of now-classic papers published in 16 1971 (71H2, 71H3). These papers not only defined metabolic pro- filing but also advocated use of methylene units (a measure of GC retention time that is virtually identical to retention indices, discuss- ed below) for assisting peak identification and further proposed a series of techniques for sample isolation which are still followed in many laboratories. In a similar fashion J ellum and his coworkers have utilized a GC-MS-COM system in the repetitive scanning mode for meta- bolic profiling studies. However, in contrast to the Hornings, Jellum’s group has used this system as the final stage of a rather complete screening system for metabolic disorders (72J 1, 72P1, 74E 1). Hence, they have primarily relied on their system to ident- ify major unknown peaks, rather than to spot abnormal profiles; only samples which have passed through the entire screening pro- cedure and still found to be medically interesting are submitted for GC-MS analysis. Common to both the Baylor and Oslo systems is an interest in identifying abnormal compounds or compounds present at abnormal levels. Little or no quantitative data has been published by either of these groups of workers. This same pattern of interest in qualitative, rather than quantitative, results has been followed by a number of other laboratories (70A1, 71H4, 71L1, 73W3, 7431, 74132, 74112, 74143). 17 Recently, more sophisticated systems have begun to appear. Sjovall, for example, has developed a method for location of steroid spectra and partial structure determination based on repetitive scanning data (73Bl). This has been expanded to include some quantitation (74A1). McLafferty et_a_.1_. (74Ml) have described a microprocessor-based system for automated identification of compounds in mixtures, and Dromey eLal. (76D3) have developed a method of resolving GC-MS data into contributions from separately identifiable peaks. These papers are all particularly germane to the present work and will be discussed in more detail in the section on computer techniques. High resolution GC and GC-MS. Two other alternatives to the use of low resolution (packed column) GC-low re solution M8 for metabolic profiling have been developed which still use GC or GC-MS, but with a greater resolution of either the mass spectrometer or the GC separation. Burlingame (74K1) has develOped a high resolution MS-COM system which he has used to analyze major components of urine samples. He has also published a preliminary report on a system using high resolution GC-high resolution MS (HRGC-HRMS). However, most other researchers have concen- trated on the less formidable task of coupling high resolution GC columns to either a computer or a low resolution MS- COM system. Pauling’s group (71A1), for example has advocated use of capillary 18 columns with on-line data analysis using pattern recognition techniques and mass spectral identification of individual peaks when necessary. Other groups, frequently using capillary GC-low resolution MS-COM systems, have been more interested in qualitative analysis. Zlatkis in particular has emphasized this approach in a series of papers describing the analysis of ether extracts of urine (71Z1), urine headspace samples (73Z1), organic volatiles in air (74B2) and serum and plasma headspace volatiles (74Z1). Politzer e_t__a_l. (75P2) have even expanded this approach to examining volatile fractions of lung, brain and liver tissues; they have also published a review of GC-MS studies of underivatized volatile compounds in biological fluids (76P1). Maume and Luyten (73M2) have utilized a similar system in analyzing derivatized and underivatized steroids down to 10 ng injected, and down to picomole amounts when selected ion monitoring techniques we re used. Horning gal. (74H2) have developed a method for the preparation of extremely high resolution (100,000 theoretical plates) thermostable capillary columns which they used for the analysis of a variety of biological fluid extracts. Luyten and Rutten (74L2) have used retention indices on the capillary columns to aid in compound identification. Novotny _e_t_§.l. have compared stationary phases of different polarity for their suitability in profiling studies (74N3), and have also 19 developed a method for concentrating samples prior to analysis (75N1). Recently, Knights ELL-1- (75K4) have used a direct- coupled capillary GC-low resolution MS- COM system for the qualitative analysis of acidic urinary metabolites obtained by ether extraction. Hedfjall and Ryhage (7 5H2) have also published a method for obtaining much more rapid scans (1.4 second cycle time) from the mass spectrometer to accommodate the need for faster data collection rates when utilizing capillary columns. 20 Computer processing of GC-MS data for metabolic profiling Forward library search methods. DeSpite the recent successes in utilizing capillary and high pressure liquid chromatography sys- tems, most attempts at metabolic profiling to date have involved low resolution GC, repetitive scanning of the mass spectrometer and on-line data processing. This has typically been followed by either manual identification of spectra or computerized library search procedures, or a combination of both (see, for example, 71H4, 73W3, 75H3, 76L1). Whether manual or computerized, these Spectrum identification procedures have typically been of the “forward” type: that is, comparing each sample Spectrum of interest to a large library of reference spectra to find the best match. These forward searches have often utilized mammoth data bases and sophisticated pattern analysis algorithms (73H1, 73K2). While pre-ordering the library file can decrease the amount of time need for such library searches -- Dromey (76D2), for example, has proposed a method for identification of functional groups as an aid to such a search -- the forward library search method has proven to be very time-consuming and costly. Reverse search methods. Beginning in 1974, several papers appeared which suggested an alternative approach for library search procedures. This approach, called the “reverse search” w—‘rfl 21 by Abramson (75A1), has also been developed by McLafferty (74M1) and this author (7482). The principal feature of this method is that spectra from the sample are searched for a match to a given library spectrum, rather than searching a library for a spectrum similar to the one of interest in the sample. As noted by Abramson (74M1), in a forward library search, “... the presence of significant levels of interference may artificially suppress the relative intensity of relevant masses and produce a bad fit. Even more importantly, when data are compressed (e.g., saving only the two largest peaks in a 14 amu region), interferences of any nature may cause relevant masses to be excluded.” These disadvantages are then presumably avoided by a reverse search procedure. I There are several published variations of the reverse library search. Abramson’s procedure (74M1) compares all spectra in a GC-MS run to each library spectrum, and then sums intensities for each positive match, so that an area is calculated. It makes decisions about a “match” based on a comparison of normalized intensities; for a match to be declared, the average match of normalized intensities between library and sample spectra must be within plus or minus 16%. His criterion for the selection of masses to be compared is generally peak intensity, so that he usually 22 selects the ten most intense masses of the library spectrum for comparison purposes. McLafferty, however, has provided a more systematic method for the selection of comparison masses; as he points out, “the most abundant mass spectral peaks are not necessarily the most character- istic” (74M1). McLafferty has developed and commercially marketed a system employing “probability based matching of mass spectra,” which examines the following factors: the uniqueness of a particular ion (m/ e) relative to all of the m/ e values in several thousand reference library spectra; the abundance of the ion in the reference spectrum; the degree of “dilution” of the spectrum by other spectra; and the “window tolerance,” or degree of variability permitted compared to the reference Spectrum. These criteria are used to compute a “confidence index,” which must be above a certain value for a sample Spectrum to be declared a match against a reference spectrum (74M1). McLafferty has also published a study of the uniqueness of various masses (7 5P1). Recently, deJong (gal. (76D1) have applied information theory to the development of a somewhat different coding and library search algorithm, but their method has not been as thoroughly tested as that of McLafferty. Retention indices. The reverse library search procedures described by Abram son and McLafferty necessitate searching an 23 entire GC-MS run for matches to each library Spectrum. However, an alternative method to provide a more selective search was described by Nau and Biemann (73N1, 74N2). In their approach, GC retention indices (deve10ped originally by Kovats, 58K1) are used to help further identify compounds located by a forward library search procedure. These retention indices, usually calculated by measuring GC retention times relative to a series of straight-chain hydro- carbons co-injected with the sample (see the methods section of this thesis for details), have also been used by Biemann’s group to aid in identification of related compounds by correlating shifts in retention indices with addition of specific functional groups (74C4). Other workers have used retention indices in combination with mass spectral correlations to compute a combined match score; this approach has been used both for reverse 1ibrary searches by this author (7482, 76G1, 77G1) and for forward library searches by Blaisdell (77B1). Other GC-MS- COM techniques for profiling. Three other approaches to the computerized analysis of GC-MS data are particularly pertinent to the work described in this thesis. The first of these is the system developed by Sjovall (73R1, 74A1), in which mass chromatograms are searched for locations where a number of ions are peaking. At each such location, a search is performed to identify potential molecular ions, and the general type of 24 compound (type of steroid skeleton in their usual case) identified where possible. A spectrum may then be compared to library Spectra by a forward search procedure, or not, as desired. In addition, a compound amount is calculated by comparison of ion intensities to those of an internal standard. The printout from this system includes the retention time relative to cholestane as an aid in identification. A related technique is that of Biller and Biemann (74B3). In an attempt to obtain spectra free from contributions of closely eluting compounds, their computer program examines all mass chromatograms for peaks. A new data file is then created which consists of the intensities of each peak found; these data are stored at the two scans corresponding to the apex and the immediately preceding scan of each mass chromatogram peak. They term this technique the production of “reconstructed mass Spectra,” and indicate that it improves the reliability of forward library search procedures. Dromey .e_t_§_l. (76D3) have recently proposed a more sophisticated method for obtaining reconstructed mass spectra. This approach uses well-resolved peaks in mass chromatograms to resolve peaks of other masses occuring at approximately the same location in the GC-MS run, and hence it is able to obtain spectra free from background and neighboring peak contributions. 25 This approach, however, requires a much larger amount of time and computer memory to accomplish than does the Biller and Biemann technique. 26 Disease diagnosis by metabolic profiling Regardless of the methodology used for metabolic profiling, the same problems must be confronted once the data are collected: how are the data to be analyzed and of what clinical significance are the data. In the case of some diseases, particularly metabolic dis- orders, where a few compounds are in gross excess, the data analysis need not involve statistics and the clinical significance is usually easy to discern. However, in the more common case, where the disturbance of metabolism is more subtle, complex statistical approaches may be necessary and the interpretation of the results may be correspondingly more difficult. Non-statistical methods. A whole body of literature has developed concerning the detection of human disease based on the analysis of levels of one or a few compounds in urine or blood. Much of this literature has been listed in two recent bibliographies (68K1, 7101). A more specialized sampling of such work is summarized in Table 1, which shows diseases detected by GC-MS methods because of excesses in one or more acidic metabolites in urine. While not truly “metabolic profiling,” these early studies at least have given confidence to later workers that metabolic profiles will have some meaning in terms of specific disease states. 27 moi. 0mg. mmmr Hme Hum—mp HBNN. Hmmb HUNN. NUNN. H02. :55. Hfiwo 00505055 «I 7-1.. III: .Iru 0500 03.05050 0500 0505077505013: . 050 055.055.3505 65590.3 050.0 053025050550 .0305 050555050 05.0 050053 00 08300.3 5500 500 55.05%053 050.0 0530555051! 000 0501505 0500 05050 000 5500 0.5050 00.0 500 0503505505051... 0500 00.007... _ 550 0050505 0500 55059505055 500 02098555.? 500 5 550E053 6300“: «002 039% 5555305505m 550500 050125505 58 05030.5 0500500 03004 050050050 505000505000 55500 55.8.500qu 550500 050355 050305 00530006 55:50005035550505: 00.0 5.550395305055502: 000005 05.5 55$ 0302 5500053. 05053 ”—00.8 3.502 00000050 08502.0 050 0555 00.0005 02:06 an 0H0>0A 55.8.50. «0 030300 :00m 05m 050¢ 050qu 05055 000.0005 .H 0.3.03 28 Hmwb HNmb SE. was. .55 52. mMmr 03mm. Hmmr 523. N23. HAS. . “Own. 0000a0w0m 500 50080500500000050558054 0033000305 055050500300 0500 5.50 050 5503500305-! .50509 “003300.505 0500055050” 0500 55000550....» 000 550000.007 so 0500 5050.5 5050:5050 0050005 053090305 050555050 5500 0500 03000 5005 000 03000050051000.5050 0500 500500 505050 000 03005 0500 5350550505 633536500 500 035005055500 0500 5.50503 050 0305 0300 5505020 20 3 00 08085 no: 0080 0.500.850 03090305 030553 5000005 05300305002 003005030." H0505 050500 55000335.. 0. 5005000335 55055550595095 0.500.850 530505m 550.5 505005030m 330550090 00055530003355 05030300 50050m 000050 0&0530 5000050 000050 00.508 H 0.55m 29 3305. and: 373:. N02. OOGQHOHQm 0500 005330000 55054” 0500 5050 000 03003 .550550 0500 035005050 5050 55050 0500 035095050 03800 5002 00500 500500 503090505 050003 0550500 5505030 550500 5503055005 050 0550500 0503055502 05500500 035805053 0000059 600508 5 0505.5 30 _S_t_atistical treatment of data. Because only a few studies have been done with metabolic profiling, little has been published that deals specifically with statistical analysis of this type of data, and hence most of the literature in this area has deve10ped from related studies in hospitals. I Most laboratories, especially hospital laboratories, have report- ed only mean values, or at best, means and standard deviations for individual compounds. Thus, for example, Young has reported “norm- a1 laboratory values” for over 200 blood, serum and urine constituents (75Yl). A growing body of literature suggests, however, that under some circumstances means and standard deviations may be mislead— ing. Burnett (75B2) has suggested listing means and standard devi- ations with “outlier” values (those further than some predetermined number of standard deviations from the mean) removed, at least when reporting quality control data. He then recommends reporting an ad- ditional value, the “outlier frequency,” to indicate the number of such outlying values removed. Reed 21g. (72R1, 72R2) have suggested using estimates of normality which do not assume Gaussian distri- butions (i.e., using nonparametric estimates) and have provided tables for doing so. Gindler has similarly recommended several rapid non- parametric tests for method comparison and quality control studies ('7 5G1). We stgard and Hunt (7 3W2) have evaluated several common least-squares methods for method-comparison studies. 31 A more general review of the statistical treatment of clinical laboratory data has recently been provided by Sunderman (7581). This review carefully distinguishes between “normal values” and “reference values,” favoring the latter term for most uses. He summarizes the types of information that should accompany refer- ence values, and then provides a very useful review of the require- ments for establishing a “discrimination value,” or statistical cutoff point for distinguishing individuals in two different categories (e.g., healthy versus diabetic). Young has published a review of the computerized interpretation of clinical chemical data that discusses this and other problems of data treatment ('76Y2). Werner and Marsh (75W2) have also provided a review of practical considerations when establishing normal values. However, Harris (75H1) has recently suggested that the use of reference standards, even when stratified by age and sex, may frequently lead to an inability to detect other than extremely gross deviations from “normal.” He has suggested criteria for deciding when use of such standards is inappropriate (74H1) and has recommended that, where possible, previous values from the same individual be used instead (75H1). This requires a different type of statistical approach ('7 6H1). A few workers have begun to apply more complex statistical methods to clinical labortory data; these methods have been fairly 32 widely applied to multiphasic screening data, but only in a very limited sense to metabolic profiling data. A pioneer in this area has been Winkel (72W1), who has suggested that multivariate statistical approaches could be used successfully to relate various test results. Winkel points out that it is possible to have values for a single variable that are univariate normal but abnormal in multivariate space, and conversely, that some univariate abnormal values, when taken in combination with other variables, can be shown to be normal. Johnson (72G3) has related the concept of the metabolic profile to multivariate statistics by suggesting that, “The profile could be thought of as a point in N space. Several profiles provide several points in the same N space. The statistical problem is to describe the cluster of points.” Mayron 9%. have used multivariate statistics to develop a profile of drug-abuse populations based on routine hospital tests (74M3), and Reece (MRI) has predicted the use of this type of statistic as a principal feature distinguishing the hospital screening process of the 1980’s. Only a few, usually rather preliminary, studies have been completed with quantitative metabolic profiling. Young eLgl. (71Y1) reported a study of the effects of patients being given an artificial diet. This study, which measured levels of approximately 300 ninhydrin-positive, ultraviolet-absorbing or carbohydrate -wvv r.5 n 33 components of urine and serum, led Young to conclude that many of the compounds in the volunteers’ samples were of dietary, rather than endogenous, origin. He also noted that at least 4 days of the artificial diet were required to reach a stable set of values for many of the compounds. Interestingly, a few of the compounds, including creatinine, were excreted at the same rate regardless of the diet. Harris and DeMets (71H1) extensively studied a smaller number of compounds; they found serum ionized calcium to be constant, within their analytical precision, for single individuals over a period of 10 to 12 weeks even when there were considerable inter-individual variations. Witten e_t_a_l_. (73W3, 73W4) reported normal organic acid levels in young adults on a standard diet, and then determined the effect of ethanol ingestion on the profile. They found that levels of 2- and 3-hydroxybutyric, adipic, 3-methyladipic, p-hydroxyphenyl- acetic and 2,5-furandicarboxylic acids were affected by the intake of ethanol. However, it should be noted that these results were at best semi-quantitative, since an ethyl acetate-ether extraction procedure was used and quantitation was by peak area on low resolution GC. Bjorkman 521:3. (7632) similarly determined levels of a number of major acidic metabolites in the urine from newborn humans, and Chalmers and Watts (74C3) have examined unconjugated aromatic acids in phenylketonuria, followed by a study of volatile fatty acids in several metabolic disorders (74C1). Liebich et a1. (75L3) have 34 used a similar semi-quantitative method for measuring levels of low molecular weight aliphatic alcohols in normal and diabetic individuals. In a more quantitative study, Yamamoto M. ('7 6Y1) have found a seasonal variation in levels of urinary metanephrine, and a minor seasonal variation of vanilmandelic acid when these compounds were studied over a 5-year period. I Routh and Paul (76Rl) found an effect of aspirin therapy on the levels of several serum constituents. Lawson e_t_a_l. (76Ll, 76C1, 7602) reported qualitatively different excretion patterns of several organic acids in man, and qualitatively significant variations that depended upon the type of diet consumed by their subjects. Many of the compounds reported in this last study were quantitated in clusters because they were un- resolved by the GC system used. Robinson gt__a_l. (73R2) have also reported very preliminary data suggesting general abnormalities among mentally retarded subjects, although this report was never amplified sufficiently to judge its importance. Blau 913;. (73B2), in an interesting study of aromatic acid excretion in heterozygotes for phenylketonuria, found that heterozygotes could be distinguished 35 from normal subjects by the excretion of o-hydroxyphenylacetic acid. Probably the most ambitious study so far has been that of Robinson and Pauling (74R2, 75Dl). These workers have used an ion-exchange chromatography and capillary GC to profile several urinary fractions, principally head-space volatiles and free amines. A pattern recognition procedure is used to identify peaks and peak areas are normalized to a specially-selected subset of the peaks to reduce inter-sample variability. The data collected are compared to one another using the Wilcoxon test, a nonparametric statistical ranking procedure. This approach has been used to search for differences due to sex, ingestion of birth control pills, student grade point average, multiple sclerosis, Huntington’s disease, fasting versus non-fasting, breast cancer and Duchenne dystrophy, with significant differences reported in each case except that of grade point average. However, they point out (75D1) that, “We have not proved that, for most of our sample groups, the only systematic property that contributes to the pattern for the group is that for which the group is labeled. We also have not shown how early the patterns for disease develop. We do not know whether or not the patterns are present before the disease is extensively developed, and therefore are useful for preventive medicine.” A less ambitious, but much better documented, system for 36 the statistical analysis of profiling data is that reported by Schoengold gt_a_l. (7681, 7682). This system was developed to allow processing of amino acid data, and it is noteworth in that it allows extensive inter-individual comparisons utilizing a variety of standard statistical methods on data routinely collected in the clinical lab- oratory. The authors, in describing this system, persuasively argue that a great deal of valuable data collected in the clinical laboratory is not utilized. They have therefore developed a relatively low-cost minicomputer-based system that makes information retrieval and comparison easy, and hence encourages such usage. The most recent technique to be applied to profiling data is that of computerized pattern recognition. This approach, which requires a large data base and a correspondingly large amount of computer memory and processing time, has been proposed as a means of finding data patterns that are not apparent from traditional statistical analysis. Kowalski (75K4) has illustrated the use of such a procedure to distinguish patients suffering from two liver diseases on the basis of levels of 8 blood enzymes, but no one has yet published a similar study using pattern recognition on levels of low molecular weight substances. CHAPTER TWO: Reagents General solvents Dry redistilled methanol Dry pyridine Acetic acid (glacial) 0. 1M Barium hydroxide Pyridinium acetate MATERIALS Hexane, acetone and methanol were redistilled by constant-flow rotary evaporation from reagent grade solvents. Reagent-grade methanol (Mallinckrodt, St. Louis, Mo.) was dried by distillation from magnesium turnings containing a catalytic amount of iodine and was stored over molecular sieves. Pyridine (Mallinckrodt, St. Louis, Mo.) was dried by distillation from barium oxide after refluxing 1 hour and was maintained over potassium hydroxide pellets. Allied Chemical Company, Morristown, N.J. Prepared on a weekly basis by dissolving 31.5 g of Ba(OH) (Fisher Scientific Company, Fair Lawn, N. J.) in redistilled water to a final volume of 1.00 1. Each of the following solutions was prepared by dilution to 1 l in a volumetric flask. Each was prepared fresh weekly. 37 Hydroxylamine hydrochloride Saturated picric acid Creatinine standard Organic acids Protium forms Deuterium forms 38 redistilled glacial pyridine acetic 0.5M 40 ml 29 ml 1.0M 80 ml 58 m1 1.5M 119 ml 90 ml 75mg of NH OH-HCl (J.T. Baker, Philipsburg, N. J.) was dissolved in 1 ml redistilled H20. Picric acid (J.T. Baker, Philipsburg, N.J.) was added in excess to re- distilled H O and stirred on a magnetic s%irrer 1/2 hour. The solution was then allowed to saturate in the dark for 24 hours and filtered. The filtrate was stored in a dark brown bottle. A 1.00 mg/ ml solution was made by weighing 10.0 mg creatinine (Sigma Chemical Company, St. Louis, Mo.) on an analytical balance and then diluting to 10 ml in a volumetric flask. Obtained variously from Dr. Clyde Williams, University of Florida; Sigma Chemical Company, St. Louis; and Aldrich Chemical Company, Milwaukee, Wisc. Deuterated (at -d ) forms of homovanillic, hippuric, 5—hydroxyindoleacetic and indoleacetic acids and 2,5,6-d -3, 4-dihydroxyphenylacetic (a<-d3 ) acid were obtained from Merck, Sharp and Dohme Canada Limited, Quebec, Canada. Hydrocarbon mixture Chromatography packings and ’supplies Diethylaminoethyl (DE AE) - Sephadex‘ A-25 Dim ethyldichlorosilane 5% OV-17 gas chromatography liquid phase coated on Supelcoport 80/100 mesh Bis-trimethylsilyltrifluoro- acetamide (BSTFA) with 1% trimethylchloro- silane (TMCS) Liquid chromatography columns with 200 ml reservoirs 12 Foot x 2 mm id paperclip— shaped gas chromato- graphy columns to fit Varian 2100 gas chrom atograph 39 64 pl decane, 70 pl undecane, 70 pl dodecane, 46 pl tetradecane, 51 pl hexadecane, 50 mg octadecane, 66 mg eicosane, 50 mg tetracosane, and 50 mg octacosane were dissolved in 10 ml hexane. Hydro- carbons were obtained from Applied Science, State College, Pa. and Aldrich Chemical Company, Milwaukee, Wisc. P'harm acia Fine Chemicals, Piscataway, N.J. Pierce Chemical Company, Rockford, Il. Anspec Company, Ann Arbor, Mi. (Distributor for Supelco, Bellefonte, Pa.) Regis Chemical Company, Morton Grove, Il.; or Pierce Chemical Company, Rockford, Il. Kontes Company, Vineland, N.J. Glass shop, Department of Chemistry, Michigan State University, East Lansing, Mi. 10 Foot x 2 mm id coiled glass gas chromato- graphy columns to fit LKB-9000 Glassware Disposable micropipettes, 100 ul 50 ml Glass centrifuge tubes Lyophilizer jars with ground glass joints Silanized glassware Instruments Gilford 300 spectrophotometer Varian 2100 gas chrom atograph with dual flame ionization detectors and Varian A-25 recorder Lyophilizer, Model 10—0 10 LKB-9000 gas chromatograph mass spectrometer 40 Glass shop, Department of Chemistry, Michigan State University, East Lansing, Mi. Dade, Miami, Florida Pyrex, MSU Biochemistry Stores VirTis Company, Gardiner, N. Y. Subjected to 5 min treatment with 1 to 5% solution of dimethyldichloro— silane in hexane, followed by washes with hexane, dry methanol and acetone. Gilford Instrument Laboratories, Oberlin, Ohio. Varian Aerograph, Walnut Creek, CA. VirTis Company, Gardiner, N. Y. Operated at 0.1 to 0.5 Torr. LKB Produktur, Stockholm, Sweden. Data systems PDP 87e P'DP' 11/40 Computer programs PDP 8/ e P‘DP’ 11/40 41 Marketed by Systems Industries, Sunnyvale, CA. , based on an original system by Sweeley et a1. (7083). The PDP 8/e has 16,000 12- bit words of core memory, a 1.2 million-word disk, DE Ctape magnetic tape storage, a Textronix 4010-1 cathode ray display device and a Tektronix hard copy unit number 4610. The PDP 11/40 has 56,000 16-bit words of core memory, two 1.2 million-word removable disks and a 7-track magnetic tape drive. It also has a Tektronix 4010 display and shares the hard copy unit of the 8/ e. The PDP 11/ 40 is capable of direct data data transfers to and from the PDP 8/ e through an interface designed in this laboratory. Programs were written by N.D. Young for the Systems Industries operating system. The programs are all based on those described by Sweeley et a1. (7083). All programs were written to be used with the Digital Equipment Corp- oration timesharing system, RSX- 11D, version 6B. General programs and assembly-language portions of MSSMET were written by C. Ashen- del. Fortran portions of MSSMET were written by M. Smisko and S. C. Gates. All other programs were written in Fortran IV by S. C. Gates, except for FRGENL, which was written by Dr. B.E. Blaisdell. All of the programs on both systems 42 were written under the direction of Dr. C. C. Sweeley and Dr. J.F. Holland. Miscellaneous supplies Urine collection containers, Falcon, Oxnard, CA. plastic, No. 4013 l Dram vials Kimax, MSU Stores, East Lansing, Mi. Accutint pH 6.9-8.4 Anachemia Chemicals, indicator paper Montreal, Canada. CHAPTER THREE: METHODS Collection of urine samples Several collections were made of urine samples that were used in work discussed in this thesis. Standard reference urine. A mixed urine was obtained by collections from a variety of individuals; no attempt was made to collect health or dietary information on these subjects, nor were they asked to fast. The sample consisted of approximately 3 liters of urine collected over a three-hour period from adult males visiting or working in the MSU Biochemistry Department. The sample was kept at 4°C during collection and then aliquoted into approximately 300 test tubes (13 x 100 mm) with Teflon-lined screw caps. The remain- ing urine was stored in a 1 liter plastic bottle. All of these fractions were stored at -80°C until used. The aliquoted samples served as reference and quality-control samples throughout the project. BCIU collection. Approximately 200 morning fasting urines were obtained from adult “healthy” volunteer subjects by the Bronson Clinical Investigation Unit, a research unit of the Upjohn 43 44 Chemical Company and Bronson Memorial Hospital in Kalamazoo, Michigan. Each subject was asked to conform to a written protocol, a copy of which is included in Appendix A. These samples, referred to as the “BCIU urines,” were refrigerated at -20°C as soon as they were brought by the volunteers to the Bronson Hospital and were subsequently transported in dry ice to MSU, where they were defrost— ed, aliquoted and stored at -80°C until used. Subjects were asked to complete a diet and health questionnaire illustrated in Appendix A. Subjects in this study may also have been subjects in other BCIU protocols, but not within a 72 hour period prior to the urine collection for this study. All subjects had had complete physical examinations within 6 months prior to the date they donated urine. The BCIU urines were used as a reference set of adult urines. ngenile reference and neuroblastoma urines. These urines were the generous gift of Dr. William Krivit, Department of Pedi- atrics, University of Minnesota. All were collected from children hOSpitalized at the University of Minnesota Hospitals. Diet and health questionnaires were not collected for this group. Urines were collect- ed as early morning samples and subsequently shipped to MSU in dry ice. Information on each patient is summarized in Appendix D. Foy urin . This urine was the generous gift of Dr. Robert Foy, Sparrow HOSpital, Lansing, Michigan. It was collected from a new— born suffering uncontrolled seizures of unknown eitology. 45 Questionnaires When appropriate, a diet, health and drug questionnaire was completed by subjects in this study. This questionnaire, shown in Appendix A, is based on the one used by the MSU Health Center in 1974, plus a simple dieter’s checklist of foods. The questionnaire was pre-tested on a group of volunteers and a question about smoking habits added as a result of suggestions from these subjects. The questionnaires were encoded into computer files using a special program written for this purpose (MSSQST) to ensure consistent coding. Isolation of organic acids from urine Preparation of columns. DEAE Sephadex A-25 is swollen in an excess of freshly-prepared 1.0M pyridinium acetate for at least 48 hours. During this time, the supernatant is discarded and replaced with an equal volume of 1.0M pyridinium acetate, with stirring, at least twice. At the end of the 48-hour period, the supernatant is replaced with an equal volume of 0.5M pyridinium acetate. The DEAE-Sephadex is allowed to soak for an additional 24 hours, with at least two changes of solution. Columns are prepared with a lightly-packed 1 cm plug of 46 glass wool at the bottom; too much or too tightly -packed glass wool results in extremely slow flow rates. Water is added and the glass wool poked with a stirring rod to dislodge air bubbles trapped in the glass wool. Most of the pyridinium acetate solution is then decanted from the DEAE-Sephadex, and the remaining material swirled to produce a thick slurry. This slurry is slowly poured into the column, allowing frequent periods of settling. This process is continued until; a column packing with dimensions 8 cm by 1 cm is obtained. Once the column is completely poured, 21 m1 of 0.5M pyridinium acetate is added and allowed to drain through the column until the fluid level is just above the level of the packing. Columns are left at this stage while the creatinine determinations are completed and the urine samples prepared; frequently the columns are prepared one day in advance and allowed to remain at this stage overnight. Creatinine determination. Either at the time the urine is aliquoted or at the time the urine is defrosted for separation on the DEAE-Sephadex, the concentration of creatinine is determined(54H1). Fresh alkaline picrate is made by combining 7.5 ml 10% NaOH(aq) with 100 m1 of saturated picric acid solution. A 25 ml volumetric flask is used for each of 5 standards and the samples, and 5.0 ml of the alkaline picric acid solution is pipetted into each flask. To each of the 5 standard flasks is added 0 pl, 62 pl, 125 pl, 188 p1 and 250 pl, respectively, of the 1.00mg/m1 creatinine standard solution, using a 47 100 111 syringe. The solutions are shaken and let stand for 10 min. Similarly, 50 pl of each urine sample is added to separate flasks. At the end of the 10 min period, all samples are diluted to 10.0 ml with water and thoroughly shaken. The absorbance of each sample is determined using the 0)11 standard as a blank. A graph of absorbance versus concentration of creatinine is constructed, with the 5 points on the abscissa from the standard solutions representing 0, 1.0, 2.0, 3.0, 4.0, and 5.0 mg/ml, respectively. This graph should be roughly linear at low concentrations and match previous standard curves, or it is discarded. Concentrations of creatinine in the urine samples are read from the graph. Preparation of urine sample. The procedure of Thompson and Markey (75T2) for the quantitative separation of organic acids from urine is used in a somewhat modified form. First, the sample of urine, previously stored at -80°C, is brought to room temperature in a bath of warm water. Likewise, the tropic acid internal standard solution (1 mg/ml tropic acid in methanol, stored at -80°C) is brought to room temperature. For each urine sample to be prepared, 50 p1 of the tropic acid solution is added to a silanized 13 x 100 mm test tube and evaporated to dryness under a stream of dry nitrogen. Each urine sample is then shaken to achieve a homogeneous solution and an aliquot, calculated to contain 1.44 mg of creatinine 48 (approximately), is removed and added to the test tube containing the internal standard. Each tube is then sonicated to assure complete mixing of the tropic acid with the sample. To each solution thus obtained is added 3.00 ml of 0.1M Ba(OH)2 (aq). Each tube is mixed on a mechanical mixer to homogeneity and then centrifuged for no more than 30 seconds at maximum speed on a small table-top centrifuge. The supernatant solution is then removed by pipette and the precipitate washed and centrifuged twice, with an additional 1.0 ml 0.1M barium hydroxide solution added each time. Oxime derivatives of oxo-acids in the mixture are prepared by the addition of 200 pl of a hydroxylamine hydrochloride solution to the combined supernatants, followed by heating at 80°C for 20 min. After cooling to room temperature in an ice bath, each sample is adjusted to pH 7 to 8 using 2N HCl (aq) or 2N acetic acid (aq). The pH is measured with pH 6.9-8.4 limited range pH paper; urine is pipetted in small amounts onto the pH paper rather than dipping it in the sample. Once the sample has been prepared, it is slowly added to the top of a DEAE-Sephadex column with a Pasteur-type pipette. The stopcock of the column is opened to allow the sample to drain onto the column bed. Approximately 5 ml of redistilled water is pipetted onto the column to avoid disturbing the Sephadex, and then enough 49 water is added to bring the total amount of water to 50 ml. The eluate from the 50 ml water wash is discarded. Acidic metabolites are eluted from the column into a clean, silanized 250 ml round-bottom flask with 40 ml of 1.5M pyridinium acetate. The sample is frozen in a dry-ice-acetone bath, using either hand or rotary-evaporator rotation of the flask to achieve a smooth, even coating of sample on the inside of the flask. If necessary, it is stored at -80°C until the lyophilizer is available. The sample is then dried on a lyOphilizer at approximately 0.1 Torr (or lower) until about 5 ml of sample remain. At this point, it is melted and transferred to a silanized 50 ml conical centrifuge tube with a ground glass stopper. It is refrozen by submerging the stoppered tube in dry-ice-acetone (or liquid nitrogen). The tube mouths are covered with several layers of coarse-mesh cheesecloth attached with a rubber band to prevent losses of sample material and loaded into 1-liter lyophilizing flasks. They are lyophilized to complete dryness and removed immediately. The dried samples are silylated in the same tubes by the addition of 250 pl of the standard Silylating mixture (BSTFA:TMCS:pyridine, 200:2:50, v/v). The stoppered tubes are heated at 80°C for 1 hour; an inverted test tube rack is placed over the ground-glass stoppers to hold them in place, reducing the likelihood of stoppers popping out during heating. In addition, the tubes are shaken at least once during the heating process 50 to ensure complete silylation. Once the samples are silylated, they are immediately trans- ferred to silanized glass capillaries. The capillaries are most conveniently prepared from commercially-available 100 p1 disposable micropipettes broken in half and sealed with a flame at one end. Each capillary is filled about one-half full (approximately 20 pl) from a 500 p1 syringe and sealed with a flame until a small bubble begins to form in the glass at the heated end. (Melting point capillaries are much harder to seal quickly and hence should not be used.) Sample capillaries are placed in 13 x 100 mm test tubes with screw-cap tops, labeled and stored at 4°C until used. The 500 pl syringe is cleaned with dry redistilled methanol, followed by redistilled hexanes between each sample. Residual methanol must be avoided since it will react with the trimethylsilyl derivatives. Gas chromatographic analysis All samples are analyzed by either GC or GC-MS. In either case, the chromatographic conditions are essentially the same: analysis on 5% OV-17 coated on 80/ 100 mesh Supelcoport, use of glass columns, and temperature programming from 60 to 260° at 4° / min. Each time a GC or GC-MS column is packed, it undergoes the 51 same procedure. This begins with removal of all previous packing material, aspiration of concentrated H2804 through the column until it is filled, and removal of the sulfuric acid after 15 to 30 minutes. The column is then washed successively with redistilled water and acetone until free of sulfuric acid, and dried with a stream of nitrogen gas. It is then filled with a freshly-prepared 1% solution of dimethyldichlorosilane (DMDS) in hexane and allowed to sit 10 to 15 min. The solution is removed by aspiration and replaced by successive washes of hexane; dry, redistilled methanol; and hexane. The column is again filled with the DMDS solution and the same process repeated to ensure complete silanizing of the glass. The remaining hexane is then removed with redistilled acetone, and the column dried with a stream of nitrogen. If necessary, the column is heated at 100°C to ensure dryness. The column is then packed in the following manner. A 3 cm glass wool plug is placed in the detector end of the column and the 5% OV-17 added via a small funnel while an aSpirator is connected to the detector end. An electric vibrator is used to Speed the process. The OV-17 is added slowly to avoid differential migration of fines, until the column is fully packed. The column is heated at 100°C for 10 min, and more packing added until it will not settle further upon vigorous vibration. The remaining 3 cm Space is filled with silanized glass wool which has been stored in a desiccator. The 52 DMDS, OV-17 and glass wool are immediately returned to the desiccator for long-term storage. Once the column is packed, it is conditioned with the detector end disconnected for at least 48 hours at 280°C. This is accomplished using normal carrier gas flow rates and programming the temperature from ambient to 280° at 2°/ min, then holding at 280° or higher until the column is fully conditioned. Since Teflon ferules are used, the nuts holding the column must be tightened several times, especially before the column is cooled, to avoid having the column leak or drop out of position. For GC analysis, the columns are 12-foot paperclip-shaped glass columns designed to fit the Varian Model 2100 gas chromatograph used in these studies. Since the columns frequently drop out of position, they are supported by specially-designed aluminum plates. Analysis on the GC is usually accomplished with the injector and detector heaters set at 300°C, the attenuation at 2, the gain at 10.10 amps/volt, the recorder at 1 mV full scale, and only a very small bucking voltage. Gases used are helium as carrier at 40 ml/min, hydrogen at 30 ml/min, and air at approximately 300 ml/min. Usually, 2 pl sample injections are used. A solution of fatty acid methyl esters is used to check the overall reSponse of the GC prior to injection of urine samples. Septa are preconditioned at 300°C and changed frequently. The detector cylinder is cleaned after 53 each 2 to 3 injections by sonication in methanol and brisk brushing with a small test-tube brush. Analysis of samples on the LKB-9000 GC-MS-COM system A regular routine is adhered to in analyzing samples on the LKB-9000, as follows: Column preparation. Glass columns 2.0 mm ID by 10 feet and containing 5% OV-17, are prepared exactly as for gas chromatography alone. Since the septa often do not stay in the septum holder, a special cage is fashioned from paperclips to keep the septa in place. Ferules are tightened regularly and replaced frequently. Septa are checked daily for leaks. At the begirming of each day, electronic noise levels are determined with a special test routine on the PDP 8/ e designed for this purpose (INTEST). Noise widths are checked, and the baseline is adjusted so that it is just above zero volts. A test is also made for high intensity random noise spikes. If the noise levels are above 300 mV (intensity of 15 out of a possible 500,000), or if Spikes greater than 600 mV are detected, further analysis is halted until these problems are eliminated. The ion source is then focused to a resolution of at least 500 (10% valley definition). A mass versus Hall effect calibration with 54 perfluorokerosene is then performed and the resulting calibration data are stored for use during the remainder of the day. Calibration is from m/ e 51 to 700 at scan Speed 8, with the mass marker reading 43 when scanning has ceased. A standard reference capillary sample (described below) is then run to test all aspects of the system. In general, 6 to 8 p1 of this sample are injected under the following conditions: the GC is temperature-programmed from 50-260°C at 10°/min; the sample is injected and the computer program started when the temperature reaches 60°C on the dial of the temperature controller. Other conditions are: ion source temperature, 290°; separator, 290°; GC injector, 150° (it reaches 270° by the end of the run); gain 8 on multiplier; scans at constant 4-second intervals at scan speed 8 over range m/ e 49 to m/ e 550 (the upper scan limit is set so that the magnet decays just to m/ e 49 before beginning the next scan, lower limit remains the same as for calibration); accelerating voltage, 3.5 kV; trap current, 65 pA; box current, 30 pA; filament current, approximately 4A. The valve is opened and data collection begun at 5 min. At the end of the analysis, the data are examined for evidence of high column bleed, low response of hippuric or uric acids, low sensitivity, poor chromatographic resolution, abnormal GC retention behavior and other problems in comparison to previous runs with the same sample. All of these problems are eliminated 55 before proceeding with data collection. Once the LKB-9000 is considered ready, the GC column is injected twice with 8 pl of the BSTFA-TMCS Silylating mixture, and then cooled to 40°C. All data are removed from the data storage disk, and the computer is set up for data collection as for the standard sample, except that the run is set to end at 60 min (SR=60 command). A 0.5 pl aliquot of a mixture of straight-chain hydrocarbons in hexane is withdrawn with a 10 pl syringe, followed by a 0.5 pl air “Spacer,” and then 8 to 9 p1 of the derivatized urine sample to be analyzed. The sample capillary that has just been opened is immediately discarded even if sample remains. The sample is injected under the same conditions as the reference capillary, except that the temperature programmer is set to 4°/ min, and the separator valve is not opened until 6.5 to 8 min after injection, depending on the elution time of the last solvent peak. At the end of each run, the GC column is injected with two 5-pl aliquots of the BSTFA-TMCS solution to prevent any carryover from sample to sample. The GC column is then cooled to room temperature to begin the next run. While the column is cooling, the data collected by the computer are examined for problems and transferred to the PDP 11/40 for storage and MSSMET analysis. During the transfer process, which takes approximately 10 min, the data are converted to the standard mass spectral data (MSD) format used on the 56 PDP 11/40 in this laboratory (76A1). Each run is stored with a six- digit identification code (month-day-run number), the sample number (month-day-year-initials of chemist-serial number), and other pertinent information or comments. Preparation of samples for preliminary studies Several studies were undertaken to test the analytical features of the DEAE-Sephadex procedure, the GC-MS- COM system , and the complete method, including statistical analysis of data. The samples prepared for these studies are described below: Approximately 10 m1 of the standard urine sample was separated on a column 10 times larger than normal and 10 times more of each of the solvents was used. The 1.5M pyridinium acetate eluate was divided into 10 containers, lyophilized, and then the silylated urines combined in a single container. Approximately 100 capillaries, each containing 15 pl of sample, were prepared and sealed. These capillaries are referred to as the “reference capillary samples” in this thesis; they were used as reference standards to test the condition of the GC and GC-MS. Capillary stability. A single urine sample was prepared by a scaled-up version of the standard procedure; all quantities were 5 times normal. A sample of the standard reference urine was used. 57 The silylated sample was carefully mixed in a Single 50 ml centrifuge tube, and sets of capillaries were stored at -80°, —20°, -4°C and room temperature. Randomly selected samples from each temperature set were chrom atographed at the following times: 0 days, 1 day, 2 days, 3 days, 1 week, 1 month, 2 months, 3 months and 6 months. Urine stability. A freshly collected sample of urine was divided into 17 aliquots of approximately 5 ml each. All fractions were placed in standard 125 m1 plastic urine collection containers. Four containers of each were stored at -80°C, -20°C, 4°C and room temperature. The remaining aliquot was analyzed immediately utilizing the DEAE-Sephadex procedure. One-m1 aliquots of each of the samples were then analyzed at intervals of 1 day, 1 week, 1 month and 6 months from the starting date. Each sample was defrosted as rapidly as possible (if necessary), mixed thoroughly, an aliquot withdrawn, and any remaining urine discarded rather than being placed at the original temperature again. All samples were separated on the DEAE-Sephadex and analyzed on the Varian 2100 GC. Silylating solvent. Four identical tubes were prepared containing 50 pg each of 5-hydroxyindoleacetic acid, indoleacetic acid and tr0pic acid. To each was added a Silylating mixture consisting of the solvents shown in Table 1. 58 Table 1. Silylating mixtures Tube no. Silylating reagent Additional solvent 1 75 pl BSTFA/TMCS 25 pl dimethylformamide 2 75 pl BSTFA/TMCS 25 pl acetonitrile 3 75 pl BSTFA/TMCS 25 pl pyridine 4 100 pl BSTFA/TMCS none All additional solvents were redistilled and stored with a drying agent. All four tubes were heated at 80°C for 1 hour and sealed in capillaries until analyzed on the gas chromatograph. Recovery study. Solutions of several apparently pure com- pounds were prepared as shown in Table 2. A 400-pl aliquot of each solution (measured with a 500 pl syringe with no air bubble between solution and plunger) was added to a 5 ml volumetric flask, as was 200 p1 of a 1.00 mg/ ml solution of indoleacetic acid in n—butanol. All of the solutions of Table 2 were prepared using volumetric flasks. The mixture was diluted to 5.00 ml with redistilled dry methanol to form the “Spike” solution. The spike solution was divided as shown in Table 3; the standard reference urine was used in tubes 1 through 11. All tubes containing urine were run through the complete DEAE-Sephadex procedure as usual. Tubes 12 through 14 were dried and silylated directly. All samples were run on the LKB-9000 and analyzed using MSSMET. 59 Table 2. Composition of recovery study solutions Final volume Dry wei ht (mg) ‘ Tml) Compound name Ascorbic 23. 8 10.0 0 -Hydroxy- a -m ethylglutaric 8. 8 10.0 Succinic l 1 . 6 10.0 Citric 12 . 6 10.0 Tropic 22 . 5 10.0 Vanilmandelic 10 . 1 10.0 nL-Glycerophosphoric 13.0 10.0 Salicylic 6. 4 50.0 Hippuric 24. 2 10.0 Table 3. Recovery study samples Tube no. Amount of SLike solution Amount of urine ‘__ (pl) (m1) 1 - 4 0 1.0 5 - 8 200 1.0 9 - 11 1000 1.0 12 200 0.0 13 100 0.0 14 500 0.0 60 Linearity and isotope dilution series. Stock solutions at 1.00 mg/ ml in isoprOpanol were prepared from 5-hydroxyindoleacetic (5-HIAA), 3,4-dihydroxyphenylacetic (DHPA), indoleacetic (IAA), homovanillic (HVA) and hippuric (HIP) acids. A total of 10 stock solutions were prepared from these compounds: one each of the unlabeled compounds, and one each of the dideutero (at-d2) forms of all except DHPA, which was available as the pentadeutero (ac-d 2,5,6-d3) form. The DHPA required a small amount of added 2, water to achieve complete dissolution. The unlabeled IAA and 5-HIAA required purification by recrystallization from hot chloroform to remove colored impurities. The unlabeled DHPA, while appearing impure, could not be purified further, and so was left as received. The labeled 5-HIAA, IAA and DHPA also appeared to require purification, as judged by color and crystal Shape, but none was purified because of the cost of the substances. In addition, 4 stock solutions (1.00 mg/ ml in isopropanol) were made of compounds for which there were no deuterated standards: vanillic (VAN), p-hydroxycinnamic (PHC), ascorbic (ASC) and citric (CIT) acids. A 1.00 mg/ ml solution of tropic acid in methanol was also used. A stock “unlabeled mixture” was prepared by using a 500 pl syringe to remove 500 pl of each of the unlabeled compounds except tropic acid. These aliquots were combined into a Single silanized 61 test tube and gently taken to dryness under a stream of dry nitrogen gas. The tube containing the unlabeled mixture was labeled tube number 1. A separate “dilution mixture” was prepared from 500 pl of each of the labeled compounds and the tropic acid. This mixture was evaporated to dryness with nitrogen and then diluted to 20.0 ml with a 1:5 mixture of chloroform -methanol (both previously redistilled). A series of 15 15-ml ground-glass-stoppered centrifuge tubes was then prepared as follows, using silanized tubes in each case. A 2-ml volumetric pipette was used to transfer 2.00 ml of the dilution mixture to the tube containing the dried unlabeled mixture (tube 1 number 1). This tube was then sonicated and mixed mechanically until a homogeneous solution was obtained. A 1-ml volumetric pipette (the “transfer pipette”) was used to transfer 1.00 ml of this solution to tube number 2. The contents of tube number 2 were diluted with 1.00 ml of the dilution mixture, using the “diluting” volumetric pipette. This tube was then sonicated and thoroughly mixed. The transfer pipette was cleaned by aspirating 20 m1 of methanol and 10 ml chloroform (each interspersed with several sets of air bubble) through the pipette. It was then used to transfer 1.00 ml of the contents of tube number 2 to tube number 3, and the Process repeated until tube number 14 was finished. Tube number 15 contained only 1.00 ml of the dilution mixture. The complete set 62 of tubes was then as Shown in Table 4. Table 4. Composition of linearity and isotope dilution series Tube no. 10 11 12 13 14 15 Amount of each Amount of each unlabeled compound unlabeled compound pg) per 4 ul injection ng 250 10,000 125 5,000 62.5 2,500 31.2 1,250 15.6 625 7.81 312 3.90 156 1.95 78.1 0.976 39.0 0.488 19.5 0.244 9.76 0.122 4.88 0.061 2.44 0.030 1.22 0.000 0.00 63 The contents of all 15 tubes were dried under a nitrogen stream and 100 pl of a 9:1 solution of BSTFA-TMCS: pyridine added to each. The tubes were sealed with ground-glass stoppers and heated for 1 hour at 80°C. Each fraction was then thoroughly mixed mechanically and transferred to 4 to 5 silanized glass capillaries and sealed until analyzed. Four-pl injections from freshly-Opened capillaries were used when the samples were analyzed on the GC-MS; approximately 0.5 p1 of the hydrocarbon mixture was added to each aliquot at the time of injection. Two separate sets of analyses were performed on these samples: repetitive scanning while the GC was programmed from 160 to 280°C at 4°/min, and selected ion monitoring analyses under the same conditions. Repetitive scanning was terminated 26 minutes after each injection. After the separator valve was closed, about 5 pl of BSTFA-TMCS were injected while the column temperature was kept at 280°. All injected substances were monitored by the repetitive scanning method; the substances and ion monitored by SIM are shown in Table 5. 64 Table 5. Compounds monitored by SIM during linearity study. Group* Compound Ions — 757%) I tropic acid 280 octadecane 254 citric acid 273 H 3,4-dihydroxyphenylacetic acid 384,385,389,390 III hippuric acid 206,207,208,209 IV indoleacetic acid 319,320,321,322 V 5-hydroxyindoleacetic acid 407,408,409,410 *Note: each group corresponds to a different magnet field strength. Urine studies. Once all of the preceding studies were complete, the entire procedure was tested on a series of urine samples. These included 9 BCIU urines, 5 neuroblastoma urines and 5 infant control urines. Each sample was prepared and analyzed by the procedures discussed in this thesis, with no difference in treatment among groups except that 1.0 ml of each of the neuroblastoma and infant control urines was used, whereas the BCIU urines were aliquoted so that 1.44 mg of creatinine equivalent was applied to the DEAE Sephadex column. In addition, one of the infant control samples was run on 3 separate days on the LKB-9000 to obtain a measure of the intrasample variability. 65 Analysis of data by MSSMET At the end of real-time data collection, a urine sample is represented by approximately 700 to 800 mass spectral scans, each containing an average of 300 or more mass/ intensity pairs. The goal of the mass spectral metabolite program (MSSMET) is to reduce these 1/2 million data to a set of 100 to 200 concentrations and compound names. MSSMET exists in two versions. The older one was designed for use with a PDP 8/ I computer with DE Ctape storage. Although successful in identifying compounds, this version, which was completed in late 1973, was limited by the small memory size and the slow tape access speeds of the PDP 8/I. It was abandoned when a more powerful computer, the PDP 11/ 40, became available. The second version of MSSMET was completed for the PDP 11/40 in late 1975 and was subsequently modified through a number of revisions until development was finished in the summer of 1976. Virtually all of the data reported in this thesis were provided by this last revision of MSSMET; hence, it will be the version discussed here. Documentation for the PDP 8/ I version is available elsewhere (7482). The following section provides a description of the structure and operation of MSSMET in general. For a more specific descrip? tion, the reader is referred to a separate document (76G3) which 66 also includes a copy of the program. MSSMET has been designed as a general-purpose GC-MS analysis program. However, to date it has been utilized principally for the analysis of organic acids in human urine, and it is this use that will be described here. In brief, MSSMET utilizes mass chromatography, GC retention indices and a reverse library search procedure to locate compounds of interest. Once located, the compounds are quantitated relative to an internal standard and the results provided to the user. This may be accomplished with almost no operator intervention. Flow charts of the normal operation of MSSMET are shown in Figures 2 and 3. More specifically, MSSMET is a library-based search procedure that operates as follows: Program initiation. This is the only portion of the program which absolutely requires operator interaction with the computer. The program begins by requesting a variety of information, such as the name of the library file, the name of the data file, the amount of urine extracted, the amount of creatinine per ml of urine and the amount of internal standard added. Once these data are read into the computer, the program can be set to run under manual control or, as is more often the case, under completely automatic control as directed by the library. Reading the library file. The library contains operating Figure 2. 67 Generalized MSSMET flowchart. The same type of analysis is performed by MSSMET for each biological sample. Repetitive scanning data are searched for spectra of retention standards, quantitative standards and the metabolites of interest. Compounds that are judged to match the library spectra are entered in a “found” file. Details of this process are given in Figure 3. 68 CSTART D V LOCATE HYDROCARBON STANDARDS I LOCATE INTERNAL STANDARD I STORE PEAK AREA OF INTERNAL 0; STANDARD A > YES END ) _ NO , I LOCATE METABOLITE] CALCULATE QUANTITY RELATIVE TO INTERNAL STANDARD MATCHES YES a LIBRARY ,7 N OUTPUT OATA To "FOUND" FILE FIGURE 2 69 Figure 3. Detailed MSSMET flowchart. Q} 70 START INPUT AMOUNT OF QUANTITATIVE STANDARDS LI 4 7F I INPUT LIBRARY ENTRY TAKE ACTION IF REQUIRED END YES OF ENTRIES END .7 No ENTRY OPT'ON CHANGE VALUE TYPE? OF OPTION COMPOUND RE- CONVERT RI 3' TENTION RRT CONVERT RRT TO RET. TIME YPE PAT To RET. TIME :V: CALCULATE SCAN NUMBER, "WI NDOW" COLLECT MASS CHROMATOGRAMS OF CONFIRMING IONS WITHIN WINDOW FIND PEAKS OF ALL CONFIRMING IONS FIGURE 3 7| COMPUTE PEAK HEIGHTS AND AREAS I CALCULATE MATCH OF EACH PEAK 0F DESIGNATE ION TO LIBRARY ENTRY __.J .__J QUANTITATIVE YES STORE PEAK SSTANDARD AREA, HEIGHT I /__.| YES CALCULATE RELATIVE AREA, HEIGHT I RETENTION YES STORE NAME, STANDARD RETENTION TIME YES OUTPUT TO "FOUND" FILE POSITIVE MATCH? I FIGURE 3 (Cont'dl) 72 parameters (“options”) and sufficient information about the compounds of interest to allow them to be found in the GC-MS data of the urine sample. The options include such information as the criteria for peak detection, what kinds Of outputs to generate and criteria for what constitutes an acceptable match between library and urine sample spectra. The compound information is of a fixed format and always includes the following: Compound name, Retention time, Designate ion and k-factor, and Confirming ions paired with intensities. A sample library entry of this type is shown in Figure 4. The retention time may be expressed in minutes and seconds (actual retention time), relative to a single standard (relative retention time), or relative to a set of hydrocarbons or other standards (retention index, originally defined in 58K1). The designate ion is that ion which is considered to be most likely to be differentiating of that compound, and the confirming ions are a set of up to 8 ions, always including the designate ion, that is used to judge whether a peak of the designate ion represents the compound of interest. The k-factor is used to convert the relative area of the designate ion to absolute concentration (mg/ ml or mg/ mg creatinine). A copy of the complete library used for the analysis of 73 .082 88883800 05 m0 003808008 030308 080 000008 05 000 80003-3 05 m0 038 05 0:0 :08 0008.80.00 05 mo 0 \8 ”8008 80580008 00 “030E fl .0080: 808800 080 0.453 080 80.3: 05 8 8098:: MR ”8:02:08 6859800 80.303809 0 8020 80308.88... 000.80 >380 “0 0&3 080000 05. Ahoy 80008 030609 a 80% 0088008 808038000 50008 808888 05 08.0 .3226 533 3008.3 05 4va 30800.80. 80300000 800g 05 088:0 0008503: 00028 03088 80.80.89 .8303ng a mo 03g 05 00.8080 8033 2.80390: 8... 0“ 008E 0am. .805: HHSmmS 05 8 08:00 0.80 00880 H0 00% 03h. .85 880: 0.02002 3280. .e 8:08 74 0mm v mmDoE .mmmeamm .«mm.o¢v .mmmeoead.ku .aoo.u.¢wd mama QHhmocqyzmzm)xomc>IIz mud * .Hm u0\ .oma zzx .om Ih\ 75 organic acids in urine samples is included as Appendix B. Location of compounds. All types of compounds are located in the same manner. After reading the library options and the first compound entry, MSSMET calculates an “expected retention time” for the compound. This is converted to a scan number in the data file and a “window” is then determined by adding and subtracting a pre-specified number of scans from the expected retention time. This window is the region of the GC—MS run within which the com— pound is expected to elute and must be no narrower than the widest expected peak. Mass chromatograms--plots of ion intensity versus scan number (70Hl)—- are collected within this window for all of the confirming ions, as shown in Figure 5. Peaks in each mass chromatogram are detected and measured to obtain peak areas and heights (see below). For each peak of the designate ion found within the window, a “match coefficient” is calculated to judge the degree of match between the library and urine sample spectra. The formula for this match coefficient is given in Figure 6; it is a modification of a formula proposed by Grotch (73G1). Confirming ions must peak within a certain number of scans (specified by an option) of the designate ion in order to be included in the match coefficient. In addition, the deviation of the actual retention time for each compound from the value expected is computed. Based on whether Figure 5. 76 Determination of retention index window. Positions of the hydrocarbon standards injected with the sample are determined from the mass chromato- gram of m/e 85. A “window” is centered at the retention index specified by the library entry for the trimethylsilyl derivative Of m—hydroxyphenylacetic acid (Figure 4), and mass chromatograms of the confirming ions are collected within this window. The ratios of these ions are compared to the library entry to determine a match coefficient. Ions peaking at scans 382 and 407 are from the trimethylsilyl derivatives of tropic acid and p-hydroxyphenylacetic acid, respectively. 77 5 IO II I2 I4 I6 I8 20 24 28 I III .............................. , ........ TII )- I: (I) Z In 5.. E LI.I .>. .- 4 .J m a IL LIA] LA fi 28 I >- /\ t 0') 5 ._ 2 E g v v .2“ f < _l “J m T I] 3S0 385 4350 355 400 405 78 83030009008 2.8083800 8088 80808 808: 05 08 28080800000 8088 80.8 808: 05 08 00 00.80808 08 0005 M00388 8808 808 08 00388 80.8 808 @808 0088080 08 0808050000 8088 08880m .20va 80880 3. 00008088 080 00 803808008 8 08 888808 08H. 8.30080 88088088 08 8880.: 05 8003008 8080088 08 0.80808 00 80888000 8088 8 0088080 8.32002 .9800: E 823E000 8088 00 80388080 808 8888b .0 0.308 79 200 c_ :0_ f. 00 3000.87me 0: 005: 2 :0 5... B 0.0827 “"0 0:2 05:05:00 00 008:2...2 Em_0_t000 SEEHQE DA .__. + u. . l-—-I 2W7]; OO_ -_ 92 II I“ Q -—. .__. I H— 2W L m manor... 80 the peak has a sufficiently high match coefficient and whether it elutes close enough to the expected retention time, the peak is designated as being in one of three categories: positively identified as matching the library entry (“+” category), identified as definitely not matching the library entry (“-” category), or of uncertain identity (“?” category). The range of values accepted for each of these categories is specified by the set of options. The use of these categories when retention indices are used to measure retention time (as is usually the case) is shown in Figure 7. Location of ion peaks. Central to any automated chromato- graphic procedure is the ability to accurately locate and quantitate peaks. MSSMET algorithms for accomplishing this are necessarily complicated to allow for a variety of special situations. As shown in Figure 8, the basic algorithm consists of identifying three regions in every peak detected in each mass chromatogram. In each region, a particular question is being asked: Region I. Has a peak definitely started? Region II. Has intensity reached a peak value and started to decrease? Region III. Has the peak ended? In order for a peak to be considered to have started, there must be either a certain number of points (specified by an option), each above the previous in intensity, or the slope of the curve must 81 Category Match coefficient I RI deviationf‘“ + 81 - 100 0 - 12 80 0 - 12 => 1 81 - 100 13 - 16 - 0 - 79 16 * Absolute value of deviation of retention index from library value. Figure '7. Criteria for positive match to the library entry. The match coefficient (MC) and the measured deviation of the retention index from the library value (IA RI! ) are used together to determine whether a given peak of the designate ion represents the compound of interest. Both the MC and the [4 RI] must fall within certain limits for the compound to be considered a positive match (+) to the library spectrum. Slightly wider limits determine the boundaries of the questionable match (?) region; all remaining possibilities are defined as negative matches (-). Only positive matches are placed in the “found” file. 82 .080380 8.88.5: 8.320qu 05 88 000 0.88 080808.888 803000.00 8808 05 80 0038> 82 .8808 8 80008808 8 0080080800 88 0.88 08808 8880 05 J08 0.88 880880 200080 0808: 080 080800. 0080880080 08 300805 05 80 080 080. 88 0080880080 08 300805 080. 80 080 080 8 80 .08800 80 808888 888800 8 808 8888 0080.805 3 088000. 3808088 808 080. 88 00080 0080080800 08 8808 088. .8880: 8808 080 80 8000888 888800 8 80 038> 080800.20. 8 30808 0088080 8080. 088 008088 088 0.8 .8. 888080 0080080800 08 8808 088. 0080800 08 08808 8880808088 88058000800 80 800,858 8888888 8 .80 08080 8828888 8 8080.80 88 888880 0080080800 08 0808 < .0280 30888800 .80 888080 8888800 08 8.808 05 8080.08? 80800 0» 000.008 088 08808 880 .0808008 0005 88 £808 05 80 080.8008 m 00803080080 888808 800800 8808 8.82002 088. .0808 8888088880880 0088 80 80300828 .8 8:088 83 m MKDOE B segm/ 0 05808000 000020 30:25 x000 00; "B 86?. 0 03? 0.020085 .0 590; 0.000 00 8:8: 00:38 3200 0089.0 8:825 00; um 2060a 8.. 0000088 :08 0020 E:E_c_E .0258 .6 00885: 8255:. 00; 8 5307.. 8 :060m a 205mm ZO_._.ow._.mo ¥.szm...z_ m>_._.<4wm FIGURE 9 466 Way/[”0]. \\ 500 SCAN NUMBER //////y/,V >._._mzm._.z_ w>_._.<4mm 4B0 5|O 490 470 FIGURE IO 2|? C >....wzw._.z_ m>_._. II area of the designate ion weight (mg) of internal standard added to the sample V : volume (ml) of sample extracted Figure 13. Formula for calculation of peak amount by MSSME T. The ratio of the area (or height) of the designate ion of the compound of interest to the area (or height) of the designate ion of the internal standard is combined with a correction factor (k) to obtain absolute concentration. The value of k is obtained experimentally; it reflects recovery from the chemical separation process, losses during GC analysis, percent ionization in the ion source of the mass spectrometer, and the relative intensity of the designate ion in the spectrum of the compound. If k is not yet known, a value of 1.00 is used; the results are then termed “relative concentration” rather than absolute concen- tration. Such relative concentrations can be used for comparison of the levels of a substance present in different samples, but are not useful for comparisons of the amount of one compound to the amount of another. 91 or height into actual concentrations. The quantitation factor is determined from the area of the designate ions of the reference substance and internal standard of a known amount of pure compound taken through the entire DEAE—Sephadex procedure. If it is not yet known, the quantitation factor is assumed to be 1.0, and the peak amount is referred to as the relative peak amount. It is also possible to avoid the use of a specific value for the creatinine standard in this formula by setting M equal to 1.00, in which case results are reported as relative amount per ml urine. In any case, the peak height and peak area results are reported separately. Printing of results. Two files of results are created and either printed immediately or stored on disk for later printing. The first of these is the “run” file, which contains information on all of the peaks of the designate ion of each of the compounds. The second is the “found” file, which contains only the best “+” match, if any, for each compound. It is this latter file which is used for later statistical analysis. The material printed in each file may include any or all of the complete dump of information shown in Figure 14. Selection of designate and confirming ions. The designate and confirming ions are those molecular or fragment ions which are expected to be the most differentiating for a Figure 14. 92 Typical MSSMET output. Depending upon the value of a MSSMET option (the “print” option), any or all of the information shown can be displayed for each compound positively identified by MSSME T. The first line of the output includes an identifying number for the compound and both its IUPAC and common names, if known. The second line contains the following information: an off—scale indicator, if the compound area is above a certain value; a sequence number, which indicates which peak of the designate ion within the window is identified; the match category, which is based on the combination of the value of the match coefficient and the deviation of the retention index from the expected (library) value (Figure 7); designate ion peak area; substance concentration (or relative concentration if k is unknown), expressed in exponential notation; observed retention time in minutes and seconds; deviation of the retention time from the value expected; observed retention index; deviation of the retention index from the value expected; and the scan numbers corresponding to the beginning, apex and end of the peak of the designate ion. The next line of the output for each peak contains data corresponding to those printed directly above each one, except that all values are computed from peak heights instead of peak areas. In the example shown, the ions peaking at retention index 1759 are found to match the library ion ratios for 3 substances, but only the entry for p-hydroxyphenylacetic acid positively matches both the observed ion ratios and the observed retention index. 08M MmM 0mm nvM vvm Ne mmma flaua + 0080M Mo ¢_mm30E mmmm.0 .mmmvwv mm + mmma.o .mmommmm mm + H * AQHZchzqoomrmv Umqyxomacolmlmzoo80033>8Im mad as mmma moms + MMHSM am 06 mmwfi monm I mNHQM do «a Mqu mono + «N QM do mmfiv.o .VVMN mm + wwmv.® .mvvm mm + N Umomzwm}xom8}1tz 000 mmmm.o .ONN00 mm + wva.m .movmm mm + a N0: 08M mmVM.Q .mMmmd mm + mvwv.® .mefim mm + 94 specific compound in a particular biological mixture. Hence, the best choices for these ions will vary with the particular choice of biological fluid, extraction method, and derivative. To facilitate the process of selecting these ions, a pair of methods has been tested. The first method was manual and involved intuitive selection of ions, which were typically intense and of high mass, with a bias toward ligh mass rather than intensity. The library of this set of ions was called MSSMETLIB; it evolved slowly as experience indicated which were poor choices. A second, and more recent, approach is the computerized selection of ion sets, utilizing two programs called MSSDSG and MSSCHS, whichwere developed for this purpose. MSSDSG (acronym for mass spectral system -designate ion selector) is designed to Select a designate ion and confirming ion set for each library Spe ctrum. It does this by comparing the library spectrum to the average of a preset number of Spectra (usually 16) taken from an actual urine sample. The key feature of this comparison, however, is that it compares the library spectrum to urine spectra centered at the retention index where the library compound would be expected to Occur (i.e., its nominal retention index). A ratio is computed of the intensity of each mass of the library spectrum to the corres— ponding intensity at the same mass of the averaged spectra, using 95 the formula shown in Figure 15. The value of q in this formula is a small real number, typically slightly greater than 1.00, to allow weighting of more intense masses. A value of 1.05 is typically used for q. Library values below a certain threshold intensity (typically 50) are automatically assigned an R-value of zero, and a minimum intensity value in the urine sample of 1.0 is always assumed to avoid division by zero. The 20 highest R-values for each library spectrum are then selected. If any apparent isotope clusters exist (usually masses within 3 of one another are defined as isotope clusters), only the mass with the highest ratio is retained. The resulting list is stored in a permanent file for analysis by MSSCHS. A sample MSSDSG output is shown in Figure 16. Construction of the MSSMET library from data supplied by MSSDSG is the task of MSSCHS. It may use up to 10 MSSDSG outputs, each representing a comparison of the library to a separate urine Sarnple, to arrive at an optimum choice of designate and confirming ions. To provide a direct comparison, the list of the twenty (or fewer) ions from each library spectrum is ranked from low to high (1= lowest R value, 2= second lowest, etc.) All R-values below a user- defined minimum (typically 5.0) are discarded. Sets of ranked ions fI‘om each of the comparisons provided by MSSDSG are then summed to produce a tabulation of the ions that are ranked best overall. II’Iasses below m/ e 80 are automatically discarded from consideration. *_ 96 Lq R - _r_n_ m s m where Rm : the ratio for the ion of mass r_n Lm . the normalized intensity of the ion of mass m Sm : the normalized intensity of the ion of mass _n_1_ in the summed sample spectra q = a factor used to weight intense ions more heavily Figure 15. Formula for calculation of ratio by MSSDSG. The ratio is calculated to measure the usefulness of a given ion in differentiating whether a particular compound is present in a sample of a biological fluid. The most discriminating ions will have the highest ratios. Typically, a value of q of 1.05 is used to weight the ratio to '- favor more intense ions. .88080 8808808 0 808 0800 0888808800 0880 000800000 80 000 0000. 080 000080 00 0880002 88 0008 080 0088080 8080 80808800 .000008 030008008 80080 5.83 000588 080 38 08:08.0 00008 000003 080 8083 0800 080 00 0000088 088. .0888800 208 80080228 0 0o 0808080 02-00 0 E 00080 808080008 08800 080 00 00800800 0800080 00 008088800 00 00 8083 088800080 88088: 8030 0 80 0880008080080 0008 08 00 08800 080 0080 0800 om 00 88 80 88080 0 0000800 008002 97 0:885 0000082 8038.08. .00 080088 98 mum 9mm mwm 0mm mmv mmm zcum .w .am .ava .maaa .m .«N .w .wv mom mzoa uzmzmmmzoo NmM 5mm awN maw 52¢ m. mmawi mam.a¢ .vaM new «mm.¢ .mM va vmm.a .ma 0mm amm.a .ma mam amm.¢ .mam mm upmzwmmmc mom mmuHozu no» 99 Finally, the 8 highest-ranking ions are selected and a library constructed in standard MSSMET library format. This library was identified as FINALLIB. FINALLIB was further modified as experience indicated shortcomings. Ultimately, two libraries were established based on it: a library of compounds known to occur in urine (BESTLIB) and a more complete 1ibrary including all compounds for which reliable Spectra had been obtained (PUBLIB). Entries in both libraries are identified in one of several ways: by chemical name, if spectra were obtained from commercially-available standards; by the prefix UNK if the spectra were detected as arising from apparent impurities in the comm ercially-available standards; and by the prefix U if detected as Spectra in urine samples. Tentative identification of some of the last class of “unknowns” has been made by reference to a variety 0f sources of published spectra; these compounds contain the SHSpected identity in parentheses after the identifying code name. PUBLIB has been published elsewhere (77G1); BESTLIB is included as Appendix B. MSSMET analysis of urine samples The final library for the MSSMET analysis of urine samples (BESTLIB) contains 155 compounds selected from a total library of 100 383 compounds upon the basis of a test run of the urine samples with MSSMET. Compounds have been eliminated from the library used for urine samples because they were not found in a sufficient number of urines or because they duplicated other entries. In addition, the library has been refined upon the basis of data from 4 of the urine samples to achieve an Optimum set of ion ratios (except for compounds 342 to 383, which were added later). Library retention indices are based upon the average of values from several urine samples where possible. All urines described in this thesis have been analyzed with BESTLIB. This library uses urinary metabolites as retention index standards, but also requires the co-injection of hydrocarbon standards with each urine sample. Nominal retention indices for the retention index standards are based upon their retention indices relative to hydrocarbon standards in a minimum of 12 urine samples. Statistics Statistical analysis of the data from the urine samples is performed on the PDP 11/40 utilizing two types of programs written specifically for this purpose. MSSTAT. This is a series of computer routines, written in FORTRAN IV, designed to provide parametric statistics on the urine Mk! 101 profile data. Standard statistics provided by this program include mean, standard deviation, standard error, coefficient of variation, product-moment correlation, t-test of product moment correlations and t-test of samples means. Six categories of data are capable of being processed: peak area match coefficient, peak height match coefficient, retention index, difference of retention index from nominal (library) value, relative peak area and relative peak height. Data are prepared for this program by RAMAST, which creates a “master array” from MSSMET “found” files and by RASE CD, which can perform simple data transformations (multiplication, division, logarithmic conversion) to create a “secondary array.” FRGENL. This program, also in FORTRAN IV, computes several parametric and non-parametric statistics on the urine data. It calculates means and standard deviations for the average of relative peak areas and relative peak heights; height match coefficients; and retention indices. It also uses peak amount data normalized to a subset of all of the compounds to compare different groups of urine samples using either the Wilcoxon statistic (non- parametric) or the “Student” t—test. Auxiliary routines include an outlier test and a logarithmic plot of peak amount data. Two pre- liminary programs, P'URXAC and TRNSPO, are required to create a data matrix from MSSMET “found” files and to take its transpose, respectively. This package of programs was written by Dr. B.E. 102 Blaisdell specifically for use at the MSU Mass Spectrometry Facility. Both MSSTAT and FRGENL use routine statistical techniques described in almost any standard text on statistics. Clinical report form. Another type of statistical output that can be generated from a MSSMET output is the clinical report form. This report, generated by MSSRPT, is a comparison of an individual MSSMET output with a file of means and standard deviations for a group of reference subjects. The relative amount of each substance calculated by MSSMET is plotted in terms of the number of standard deviations it is from its respective mean value. A correction factor can be applied to the data to normalize the data to the sum of relative peak areas. If k-factors are known, these are multiplied times the appropriate data. Means, standard deviations and k-factors are kept in files which may be edited using the system text editor, and are dated so that the clinical report form includes the date each file was established. CHAPTER FOUR: RESULTS The results may be arbitrarily divided into three categories: evaluation of MSSMET and mass spectrometer-data system performance, tests of the urine extraction procedure, and clinical studies. Each of these topics is discussed separately below. Evaluation of MSSMET and GC-MS-COM system MSSMET and the GC-MS-COM system were tested to ensure that they were performing as expected. Data collection reliability. Data collected by the LKB-PDP 8/ e system were examined using MSSOUT, the general-purpose mass spectral output program, to detect anomalies which might contribute to an unreliable data analysis by MSSMET. Several such effects were noted and appropriate changes made in the data collection routines on the PDP 8/ e. Most notable of these were problems associated with real-time ion peak detection. Extensive testing over a 10-month period finally revealed several problems. First, only one of the four intensity amplifiers was being tracked to determine the starting and ending points of the ion peak. For very intense peaks, especially if resolution was poor, this resulted in the peak at the 103 104 next higher mass being “missed” by the detection algorithm. Secondly, the peak-detection algorithm was too slow, so that relatively few points were collected during the sweep of the magnetic field across the ion. The third contributing factor was that the algorithm was too sensitive to small negative noise spikes in the beginning region of the peak; a peak was declared “ended” because of the noise spike, and the rest of the peak missed because of the delay involved in processing the spurious peak. These problems were solved by tracking the most sensitive unsaturated intensity channel, instead of simply the most sensitive channel, streamlining the peak-detection algorithm and installing a criterion that a peak must have a certain minimum width before it is considered ended, respectively. Data collection parameters. Also associated with the reliability of data collection were several operating parameters of the LKB-9000. Although the complexity of the interaction of these parameters with one another and with the data collection algorithm made a rigorous test of the influence of each factor impossible, several parameters were found to be especially important. The results suggested a set of minimum criteria without which reliable data collection should be considered impossible. This set of criteria, summarized in Figure 17, was implemented in all subsequent studies. Figure 17. Minimum criteria for LKB-9000 operating parameters. Whenever samples are to be analyzed on the LKB—9000, operating conditions are modified until all of the criteria listed have been met. 106 Figure 17. Minimum criteria for LKB-9000 operating parameters Noise levels 60 Hz and higher fre- quency noise on intens- ity from multiplier Random. noise spikes Hall effect Mass resolution Ion source focusing Hall effect vs. mass calibration Ion source leakage current Sensitivity Less than 0.4 mV (20 bits on data system) peak—to-peak Less than 0.2 mV above or below normal noise level Less than 40 uV (2 bits) 500 (10% valley definition) throughout mass range Should be optimized daily; no lens voltage should be at either extreme; peaks should be symmetrical Should be Optimized daily (more often if room temperature not constant); for masses 51-550 at scan speed 8 (4 seconds per scan cycle), calibration of each mass must be within 1/ 4 mass unit of value predicted by extrapolation from previous two calibration points. Day- to-day drift should be no more than 1 mass at highest mass, 1/4 mass at lowest mass. Should be stable. A 4 ul injection of test urine capillary should produce at least one peak at m/e 73 over 100,000 intensity. 107 Library spectra. Retention indices and complete spectra were needed for compounds expected to occur in the urine samples. Therefore, once the GC-MS system had been tested, reference spectra were made of almost 300 organic acids for which reasonably pure standards were then available. Many of these spectra are otherwise unavailable in the literature, but are too space-consuming to publish here; hence, the entire set is in the process of being added to the National Institutes of Health mass spectral library. These spectra form the basis of the more abbreviated MSSMET library spectra used in the following studies. Calculation of retention indices and match coefficients by MSSMET. Over 100 retention indices were calculated both manually and by MSSMET; no cases of disagreement were found, except where an error had been made in the manual calculations. Likewise, both manual and computer calculations of match coefficients were found to agree completely, although only a few of these were calculated manually. Mass chromatogram peak detection by MSSMET. One basic criterion was that MSSMET give peak heights and areas identical to those of MSSOUT, the general-purpose data display program. This was checked manually for a large number of peaks using identical integration parameters, and both programs were found to perform identically. 108 Sensitivity and linearity of system response. A serial dilution of several test compounds was performed, and the entire series analyzed by GC—MS and MSSMET. The results are displayed in Figure 18 for a typical compound. An expansion of the upper limit of this curve is shown in Figure 19. These serial dilution curves are usually linear from approximately 10 ng to 10 pg injected, but show marked non-linearity above 10 pg injected. Precision of retention indices. Two separate tests were made of the precision of retention index determination with pure compounds, each with equivalent results. The combined results for 156 determinations of retention indices at a wide variety of concentrations are shown in Figure 20. The standard deviation of the retention index determinations on pure compounds was 2.20 retention index units, or approximately I scan. Match coefficient reliability. Using the data for the linearity determination, plots were made of match coefficient versus concentration for pure compounds. One example is shown in Figure 21. These plots, which varied quantitatively from one compound to another, were qualitatively similar in every case. Specifically, the match coefficient appeared to be approximately constant above a certain minimum concentration (the limit of detection for that ion), but drop rapidly below that point. Reproducibility. Two separate capillaries of the same urine 109 .08008000 00880080 0800 088080 080 00 Sam 0 >880 808 00808000 080 00 0300008 0800 08088080880838 00 0808 0008m8000 080 00 00080 080 88080 BHSmmS 80, 0000080000 080 00080 0>80000m .83080 080 808 8000 800 00880080 00>880 308803 080 00 08080808 800880 08.0. .Fom 0 \88 00 0800003 000 {E 8.5.. 80-20 008 38 8.8.80 000 {8 "008800088 0803 0808 080800080 00880 00 0000008005 00000080 08 we 00 00000080 w8 8 88080 w88w808 08080080800800 00 000000 0803 0800 08088088088080.8808 00 03003800 0300888008880 080 00 0038800 .w8888000 030800808 .8. 008800088 00>880 0888803 03000008000 08 085.088 ‘ 0. ”rd—"V IIO 620 0230—20. m.E_>_xomo>I - 0 _ 50.0 10.0 VEHV HAHN-EH 111 .08 0.8.088 :8 00880088 80800 80808000 008800 080 00 8080808 0 800 8.3080 08 88080080800800 00x80 0 00 0800088 08 80883 08008000 00880088 0800 088080 080 00 0300808 800 0880 00 00808008 .0800 30880888080888 .80 03003800 0380888008080 080 00 m3 0 \88 00 08888080 8008880 .8 08080080800800 8m88 00 92.00 080 80 00808008 800880 0 3080 008 00 00080080 0008000880 0002 .w8888000 030000808 80, 008800088 08080080800800 898 00 0>880 w888803 0808000808080 .2 08:08.8 ||2 802 0 8.23024 w882880 00808088808880 00080088 0008000880 80800 00 08088080. .mSTOU 080 0088 00000880 0800 088008088 00 0808003800 8808088800 88880 080 00 00880880 w80880> 800 9882mm: 88 0088880000 0803 008080800000 80002 00000880 0888800 00 08808880 8088 0808080000 80008 00 008008080Q .80 088088 -.—_.-._——----- H6 _N mmDoE 820 000.0002. 0.200s? 000.0. 000. 00. 0. _ _ 0 _ o as: I 0.04 0300242 100 I00 -00 -00 . L filqlblldldllqllal o o 00. lNBIOIddBOO HOLVW 117 sample were opened and injected one week apart. The MSSMET data were then compared to obtain a measure of GC-MS reproducibil- ity. The results are shown in Figure 22. In addition, the sample was examined for the possibility that GC-MS conditions changed during the run, as shown in Figure 23. Data for both figures are peak areas relative to the quantitative internal standard; they have not been normalized to the sum of peak areas. In addition, the “unreliable” compounds listed in Appendix C, discovered by tests of 19 urines including one of the two tested here, have been removed from these plots. The median coefficient of variation between the two samples is 8%. Of the 14 compounds with a coefficient of variation greater than 35%, one is an artifact peak (probably from the pyridine solvent), 2 are substances just above their limits of detection, 3 have a retention index more than 6 retention index units from the library value (and 2 of these also have one of their match coefficients below 80), and 4 show evidence that the designate ion peak is poorly resolved. The problems with the remaining 4 substances are not explanable on the basis of the data contained in the MSSMET outputs. Quantitative precision. Using the data from the linearity series, it was possible to determine the precision of quantitation by measuring the deviation of the data from a first-order least-squares fit of the data. For 20 samples, the mean coefficient of variation of the relative peak areas was 4.9%. Precision on the same samples 118 0808080800 08008000 08808 ooomumvmq 080 80 08080 8003 080 00880080 0803 00088800 030 088. .0880 080 88 000880000 00 8080808888088 0088880880 0803 mm 8080 8000080 80800880> 00 008080800000 8008 0088088800 080 800 000G 003003800 83800880088880 080 00 000800880 080 088000088 008080880 80880 x000880mum>>>DD>>>D>>>>>> DD» DOD obtbflt ' 0 0 ’ ' . ' Tom. 0 b o 5 b b b o 100. b 5 :00. b 0 O O o o D 9N VEHV BALLV'IBU NVBW :IO 1N3083d 122 Table 6. Precision of isotope ratio determination* Compound m/ e Coefficient of Variation 3,4-dihydroxyphenyl- 389/390 4; 2.2% t 0.23% acetic-d 5 indoleacetic-d2 321/322 i 2.8% i 0.36% 5-hydroxyindole- 409/410 :I- 1.4% t 0.33% acetic-d2 * Based on 10 injections of 1.0 pg of each compound. Table 7. Accuracy of isotope ratio determ ination* # Isoto e ratio Per cent error Compound m / e Theory@ MSSME T SIM MSSME T SIM 3,4-DHPA-d5 389/390 2.92 3.25 2.81 11.3 -3.8 IAA-dz 321/322 3.54 3.25 3.43 -8.2 -3.1 5-HIAA-d2 409/410 2.72 2.56 2.56 -5.9 -5.9 * Based on 10 injections of 1.0 pg of each compound. # 3,4-DHPA-d is labeled 3,4-dihydroxyphenylacetic acid, 5 IAA-d2 is labeled indoleacetic acid, and 5-HIAA-d2 is labeled 5-hydroxyindoleacetic acid. @ Calculated based on average isotope abundances. \7 a - K o , \_ , x u l I . , I I ' ‘ a ' I ‘ c ' ‘I I . o I, I ‘ . I ‘- . , , \ 123 using a selected ion monitoring (SIM) system ranged from 2 to 4%, depending upon the compound tested. Quantitative accuracy. To obtain a measure of the accuracy of quantitation by MSSME T, as well as another comparison of quantitative precision, the same isotope ratios were measured by both MSSMET and SIM. These data, as summarized in Tables 6 and 7, indicate that the mean precision of isotope ratio determination is t 2.1%, compared to the :t 0.3% precision of SIM. Mean percent error using MSSMET to determine isotope ratios was 8.5%, compared to a 4.3% mean error on the same samples using SIM. Tests of the urine separation procedure Although the general DEAE-Sephadex procedure had been studied extensively by several groups prior to our work (71H2, 71H3, 72C2, 72C3, 72C4, 75T2), it did not perform well in our hands, so considerable testing was undertaken to improve the procedure and to document it in such a fashion that other laboratories would be able to replicate our results. The revised procedure was then tested to gauge recovery, reproducibility and sample stability. Recovery. Recovery data were computed by comparing the GC-MS results with a group of pure compounds to the results obtained with samples in which the same substances were added at 124 three levels of concentration to urine samples. The results from this study are summarized in Table 8. Recoveries for the derivatized hippuric acid are not given because of the difficulty in obtaining a reliable GC response with this substance in urine. Recoveries for the compounds ranged from 10% for ascorbic acid to 98% for vanilmandelic and Q-hydroxy-Q -methy1glutaric acids. Only ascorbic and citric acids had recoveries below 50%. Reproducibility. One of the principal measures of success of the separation procedure is the reproducibility of multiple separations of the same sample. Hence, conditions of the procedure were investigated and altered until a reproducible result could be obtained. Several tests of reproducibility were completed on the Varian 2100. GC profiles from a similar test on the LKB-9000 are illustrated in Figure 24. Chromatograms of m/ e '73 have been plotted to show only the trimethylsilyl derivatives, so that hydrocarbons do not influence the plotted intensities. Sample stability. Studies were undertaken to determine the stability of the unprocessed urines at various temperatures and to monitor the stability of silylated samples during storage in sealed glass capillaries. Examples of the former study are shown in Figure 25, and of the latter study in Figure 26. It should be noted that hippuric and uric acids (the two major peaks at the end of the run) varied with the GC column conditions, so that in some later 125 Table 8. Recoveries of organic acids using barium hydroxide- DEAE -Sephadex method. Compound name Mean recovery* Succinic 92% o-Hydroxybenzoic 67% Q-hydroxy- e-methylglutaric 98% Tropic 91% ak-Glycerophosphoric 87% Citric 16% Homovanillic 94% Vanilmandelic 98% Ascorbic 10% Indoleacetic 92% * Based upon analysis of duplicate samples at each of two different concentrations in urine. Figure 24. 126 Reproducibility of analytical procedure on urine samples. Two aliquots of the same urine sample were each separated on DEAE—Sephadex columns. The resulting organic acid fractions were lyophilized, derivatized and then analyzed on separate days with the LKB-9000. Hydrocarbon standards were coinjected with each sample; hence, m/ e 73 is plotted to show only the trimethylsilyl derivatives present in each mixture. RELATIVE INTENSITY I27 RELATIVE INTENSITY 13 15 20 25 "30 35 40 45 TIME (MINUTES) 13 Is 20 25 36 35 40 '43 TIME (MINUTES) FIGURE 24 a 128 .580 o? 0080008 0838 0 8 0 00 580.0 00 .80 o... 8 803 888808098 0.8008083 000308 08030000 .0208 0.3 02088 .3080 5 0.02002 .3 00330003 000000080 .0000. «0 000.800 30 05808 c 8030 08800 08.8 0 03080 m 8080008080 0:83 .8009 003 00800 05 085 05 00 0003000 08800 08.8 0 80.3 03 0. 8080808080 .00 003 0080> 0 00 0080800080800 000 0800008 N000am0mIm-Ho «7A °8 GAHqSHo-EWAH ovwowouvaa 4" amu ' . . 3 Vds-n - 380338 ~92 HLHdaaan'NVA '9 Vd9- °N Vd-Ho-v —9 VdH- - w" - - ovwoevavwoudoaim 2 ZNBB Ho 9 - 10190310. 'ZNsm-Hcgfimut;i °0 .Ln'ISHO-Dv OldIOV-BW-E -t omoaam oncnov DINOUHLAHB ’o 0 $2 olaums .Lauxoao 3 Huaaxxoao ””3 “9°03 3» !d 3 '— <1 :1: 3.. E 10830A19 ,9 3 103333 onvxo OIAOBAd ~ memo-9 . _‘_\J onoane OILOV‘I BIHO-D L 0 '03 +- °o h-Cl') ?— FIGURE 25 130 .88 \00 00 Ocean 00 000 80.3 00800000000 00000000800 08800 5430 00m 00 NH 0 0800 .8 000m 000005 0 00 0003000 0003 0000800 000. 0000000 000800 000 00 0000000000 00 000000 000 00 000 00 00 0000000000 00 000 000000 08000 00000 00 0000000003 .00000 “08800 00 000 00 000000000 000 0000 000000000 0000000 00 00000 00.8 000 00.80000 00 080 00.000 000. 0030000800 8000 00 000000 00000 803 0300080800008 008000 00000 000000000800 00000 00000 000 00 000000 0000020000 00003 0 00.0 00500000800 8000 00 0000000 0000.00 0080000 003 m 00000800 00003 .00000000 00 00000 8 000000 000 00 0808000. 000 00 00000000 003 .0 000000000 00000000800 8000 000 00 .000. .0007 00 0000020000 .00 0000000 080808 000000 0. 00 0000 00 00000000 0003 00000 0000000 000800 00 000800 0800 000 00 0000020000 030. 0000030000 00000 00800000 8 003003000 00000000008000 000 00 000000 000 00000 0000000 00088 00 0000800 00000000000 000800 000000 .00 000000000 .00 00:00.0 mm wmzoi 00L wmahdmwmcfih .000 .000 . .0w0 . .60 r .00.: . .00... . .0m. . .00. r .4. . .00. .00 .00 .00 m >m<._.:a 030. 30900, 00% 8090999000 00008 8000 009 05 0.00:3 00.80090 0003 00000 309 0 980 .0000: Mom :05 0000.00 00, 0008 809039000 08.08 00.00 0009 .8 809039000 00008 8090: 8009 05 .8590 005 003 80.8.0090 900.5: 05 8 .9308 0309009 0 809 080080 05 90 000 .0099800 0900 080.000 808.8 5 8 08.3 00869800 03 09 9.902002 .3 00059800 0003 0809039000 00008 3090: 0009 0809099000 5308 0008 90 000950505 .00 0.5000 I45 cor hm mm mmDGE ._.Zw_0_u_u_m_00 IO._.<_>_ Z_ 0m 5 mm mm mm mm ow 2. m-. 09. n2 Nmn Z_ ON mm SCanOdINOO :IO UBQanN 146 .00 8:08 E 00020 000 00809800 0800 05 90 8000 00.0. 0000905000 8808 0008 009. 0000099 0.00 080005000 0308 90 0093 50m .9. "52002 .3 0099800 0900 08000 808.8 0 8 0859 00859800 0.09 .009 00008000 0003 0809039000 80008 00.00 09009 00.0 8090: 0009 000000808 0308 80000308 00 000.8808 .00 0.808 I47 2: mm 0m ._.2m Gin—moo Io._.<_>_ 5 mm mm mm «on n 2 mm H 24.52 mm mm on NKDGE MN. on 0 NF 0N om we om Nb 0m AONSDOSEH . 148 .88 0800 08 90 009009 8080 8089 0030008 08 0009 88 080808000 08 8083 00. 080800 08 080 800089 00880050 80 0.85080 08» 80.05 90 880.0858 0 08 8008 89009 089. .0000 8000 8 8009 88 000808000 08 80 80.083 08 080950900 08 0005 0803 0099800 0885 8508 8 008509800 59 809 00880908 088: 0988 208509: 089. .08509800 8000 80 8009 88 080808000 08 80.8. 0808858 8000 08080 080 080.8000 08 08 00989 208508: 080 2858: 088 808 8 8.92me .3 00880908 0800 08 80 8809 .0808 0008 008 00808000 .5 008888008 .80 080089 I49 .m 0050.... m2<0w Z. I._.D_>> v__ WE. u Z_ o 2 cm om c0 om AONBROEIHJ 150 values ranging in individual urines from 0.935 to 0.999. Distribution fleak amounts. To check whether the peak amounts for a given compound were normally distributed within members of a group, the relative peak areas were plotted for several substances in the BCIU urines that had no values of zero (to avoid the problem of taking the logarithm of zero for the second plot, described below). The data are plotted in Figure 32, as the normalized relative peak areas, each expressed as a fraction of the mean value for that compound. Since the distribution of these values appears to be log-normal rather than normal a Second plot was constructed, using the same data converted to their logarithmic forms and then expressed as a fraction of the mean of the logarithmic values. This plot is included as Figure 33. Outlier test. A test for outliers (abnormally high or low values) was performed on the normalized relative peak amounts (listed in Appendix F) of both the neuroblastoma and BCIU groups. For a group of n subjects, the test compares the difference between values for subjects (n- 1) and (n) to the difference between the values for subjects (1) and (n). A similar test is made for outliers at the 1owest values. A table of outlier values significant at least at the 5% level is included as Table 11. One of the 5 neuroblastoma patients and three of the 9 BCIU subjects had a large number of substances identified by the program as outliers. -‘M --" ' ' .mm 08509.9 8 009099090 080 0900 0800 089. .0900 089 8 0005908 0803 00885 Bum m =0 8 9800089 08009 00 0099998009 9.92002 80983 0008080850 9980 08509800 9089 90 0080 89009 0390908 008908808 8008 08.9 09 09908 0 00 000008980 0803 08509800 8000 809 00080 8.9009 0890908 008908808 089. .00080 09009 0890908 9000809 mm 089 950— :0 90 850 089 09 0009908808 0803 00885 Bum a 8 00900 0980080 809 9.992.002 90— 00980908 00080 89009 08809009 151 0099800 0885 008080908 90 95080 0 8 0809008800800 90 8895098905 .00 0.889 Nm mane...— _._.<._mm_ OwN_._<_>_mOZ 243.2 “.0 22.—.049: o.m mé Nd mfi Tm o.m 9N NN mé 1F c... od Ndo l52 c or ON on ow om AONEIHDHEH 153 .0800 90 88880009 08 0809 09 08989 90 8098089 088 080>0 09 80080 8 .0005 008,. 00885 Bum o. 089 90 80 8 0800 90 8089089800800 0 8983 008080850 02 08509800 0800 08 809 00080 8009 0890908 008808808 089 90 088880009 088 90 8008 089 09 08808 0 00 000008980 808 080 89 0008 889880009 098 09 00980800 003 0080 8009 0390908 008908808 8000 “80895888080 2908808: 0 30:09 09 800990 808 090 mm 08:08.9 8 009099080 0900 08 00500008 .0099800 0885 008080908 90 95080 0 8 0803089800800 90 088880008 90 8099588905 .00 080089 l54 mm umber“. 0308 .83088 908 09 9800089 008090850 90 985080 089 8083. 08008090 9088098 90 985080 83059 0 8089 88 090808000 089 90 00809008 089 09 08509800 90 985080 830809 0 8089 88 090808000 089 90 00809008 089 08850008 98 00959800 080 0809009109 3999088079 08099089800800 09590080 09 00080 09009 0890908 9803800 09 0005 08 800 3 809009 8099008800 0390998050 089. .0008908 039 38 3 809009 8089008800 0389099980590 089 90 809908880900 .80 0.3088 202 002008 95. .3 00 809009 8099008800 0399099980590 089 90 8099088809009 .980 0809.9 o.o9 o6 0988000988 0. 9 9. .9 09900009008 00.0 «.0 098599999 m .N m .9 098990 mm .o 00 .o 0980890089080099UIJ0 9.0 .0 9m .0 8008089800990 0 .9 0 .9 0980050 - 00.0 998.90 090080 89980.? 0 .N m .9 09809590919890 8:» 1980808999.. 0 88. 80 9880805 80 v9 90 059099 08509800 uahff:tdttnv nhf.4ufa.I.Uf-Hud.h vaaJ Fun {-57:31 203 to a high degree Of accuracy for a period Of several years, at least. Hence, the practice I have followed in develOping statistical studies (q_.v_.) is to use a k-factor of 1.00 for all compounds throughout the analysis, and to apply the k-factor, where known, only when data are being displayed for publication or use by others. This allows me to keep a separate computer file of k—factors which can easily be updated without recalculating all previous results. In addition, every time this file is used, the computer printout contains the date it was last modified, so that there is never any question about which k- factor was used for a given calculation. Ion intensity variability One of the more serious problems Of any library search procedure is the variability Of ion ratios in spectra. This poses two difficulties: accuracy Of quantitative analysis may vary, and the spectra matching algorithm must be relatively insensitive to such changes if the compound is to be found. The fluctuations giving rise to these problems generally fall into two groups, random short term noise and long term drift. The ion current for most ions varies by about 10% from scan to scan as a result of the random noise and intensity fluctuations which are the inevitable concomitant Of a sampling technique (repetitive scanning) an) ‘ffll:'.’[f{1116 3“ 111:1”01' :1: Hi .1 9 . 3113.13591er 1 .t>.1:);'..~.r: .=:‘:I 03‘5“ . I ‘—---“ -8-I 9‘... 204 which takes a single value each four seconds (of the ion peak apex) rather than averaging values over an extended period of time (SIM, for example). Thus, short-term noise is important, particularly for low-intensity peaks, but does not seriously hamper collection of reasonably accurate data. Of considerably greater concern, however, are the long-term fluctuations in ion intensity ratios. The ratio of two ions has been Observed to vary as much as 50% over a long period of time; usually this is associated with changes in ion source temperature or other variables under operator control. Unless compensated or corrected, these long-term variations can result in considerable inaccuracy of quantitation and great difficulty in obtaining high match coefficients. There are several ways that long-term variability can be compensated. One is to keep a record of ion intensities for each compound found and constantly update the library values based on recent experience. A second is to run a control sample each day, and use values from it to correct for systematic fluctuations from previous values. A third is to assume that the deviations within a given run will be systematic, rather than random, and devise a “self-correction” procedure for these deviations. A fourth method is to use multiple designate ions, rather than one for each compound, to Obtain better estimates of true peak areas. Each method, of course, has its disadvantages. The first requires 205 extensive bookkeeping by the computer and is of relatively little value for sudden, marked shifts in intensity ratios (I have Observed at least one such sudden shift). The second method will only work for those compounds contained in the control sample. The third alternative, while attractive, is based on an as-yet-unproven assumption of systematic errors, and the fourth did not seem particularly promising when briefly tested on an early version Of MSSMET. None of these alternatives has been formally implemented in the current version Of MSSMET, although the library has been manually updated by inspecting several urine runs to achieve optimum intensity ratios. With this new set of averaged ratios, very few situations were encountered for which the ion intensity ratios varied by more than 10 to 15% from the mean. Hence, the entire set Of ratios from each run was accepted as containing only a small number of significant errors. However, in the long run it will be necessary to implement a vigorous correction procedure, since data will be collected over a much longer time, by several individuals, and perhaps even on several instruments. The procedure I would recommend is a combination of several of the above methods. Specifically, I would suggest running a daily control sample, which should be evaluated by MSSMET against previous control runs before proceeding. If a certain number Of compounds fall outside Of acceptable limits 206 (e.g., more than 20% of mean), then the operator should be warned to inspect the instrument and correct the source of variation if possible. However, if correction of the deviant behavior is not possible, then the control sample could serve tO provide corrective data for the MSSMET library. Comparison Of MSSMET to SIM The major technique used by mass spectrometrists for quantitative analysis is selected ion monitoring (also referred to as mass fragmentography by some practitioners). Hence, if MSSMET is to be considered a viable technique, it must yield comparable results to those of SIM on the same samples. As shown in Tables 6 and 7, and by comparison of Figures 18 and 35, MSSMET does indeed yield quantitatively comparable results with known standards. Unquestionably, MSSMET does have some disadvantages: it is not as precise and sensitivity is at least One order of magnitude (probably two orders of magnitude) better using SIM with isotope dilution techniques than with MSSMET. The real advantage of MSSMET, however, lies in its use for the analysis of a wide variety of sample components; whereas SIM is typically limited to One or at best 6 or 8 compounds per sample, MSSMET routinely quantitates over a hundred separate components in a single sample. MSSMET can be ”1- w—r— 9* 207 .80 0.80 000 .000 .000 0 \8 ”008509800 0899 809 99050080995890 00850008 0803 0809 850.9 8.91890 800 08809800 885900 99 o9 0 80 00399039800 9999098908989 089 00 00099080 0803 008509800 890m .009009 08099089800800 099800 90 00808 089980 089 809 0089099090 080 0900 08.9. .0900 09900099808980809809861080950009809 089 90 00900.98 09 9v 985080 0089 0 8909800 9089 008590898 8 9800089 0900 0990009980898080980106 90 0985080 0509803 90 98080850008 08 09 090199080 90 0999 0989 90 0998080 80. 092-00 98 090899080 0889099980590 809 0005 80990 080 0059088009 088099808 809 00900900 080 80995990 0909009 088099808 809 00900900 98 00850008 0850 089803 039909998050 .00 080088 208 QMPOM02_ 0 Z 090 0.8 t 0.0. 0.... mm mmDoE r .00. .0 09.0.. 40:06.0 10.0. a 3 .I V m A 100.. 3 V a 3 v '00.. 0.0. 209 used to quantitate compounds whose identity (chemical name) is unknown, and, because it is not a real-time analysis system, can even be used to quantitate compounds discovered months after the actual GC-MS data collection is complete. By comparison, SIM can only be used to quantitate compounds for which differentiating ions are known in advance, and for full sensitivity to be achieved, labeled isotopes of the compound must be available. Thus, SIM and MSSMET can be seen as complementary techniques. SIM is appropriate for rapid, high-precision, high- sensitivity studies of a very small number of well-characterized compounds. MSSMET and similar techniques are most suitable for the analysis of complex mixtures when the investigator is less certain of what he is looking for, or when he is interested in a large number Of components. Ease of Operation of MSSMET One Of the proposed features of MSSMET was that it should be highly automated, and hence easy to operate; therefore, it is important to ask how well this goal has been achieved. Based on my own experience, as well as that of 3 other individuals in the same laboratory who have used the program, MSSMET is as easy to collect data for and use as SIM, but, like SIM, requires some care _ _, : 2 10 in the planning stages. The primary difficulty in using MSSMET is the establishment of a working library containing all Of the compounds of interest plus retention time standards and internal (quantitative) standards. I approached this by running pure standards, 6 to 12 at a time, over a period Of close to two years. In retrOSpect, I would have been much better advised to spend a smaller amount of time looking at urine samples to Obtain spectra Of compounds actually present in the samples, rather than finding spectra of a much larger number of compounds, relatively few Of which occur routinely in urine. The other major impediment to the routine use Of MSSMET is the method used to locate retention index standards. This portion Of the program is fully automatic only if retention times of the standards are relatively close to those expected. Major deviations--caused by changes in carrier gas flow rate or variations in the actual starting temperature--may necessitate some operator intervention to achieve proper location of the retention time standards. There are two ways to make this process more automatic. One is to standardize the operating conditions Of the GC-MS so that carrier gas flow rates can be measured directly and by using a digital temperature gauge in the GC oven. The other is to use a different algorithm for the location Of the standards. In any case, this difficulty can be easily overcome by an Operator who has a minimum Of training with MSSMET. 211 Studies on urine samples Once MSSMET had been successfully tested on pure compounds and standard urine samples, a series of urine samples was collected and prepared to fully test the ability Of MSSMET and to check for preliminary indications, at least, that MSSMET would produce clinically useful results. Selection of subjects. The first step was to collect the urine samples. The major effort in this regard was the collection Of some 200 “BCIU” urines from a group of reasonably healthy adult subjects of both sexes. In no way was the group of urines collected from the BCIU group intended to be “normal.” As shown in Appendix D, the data Obtained from the questionnaires Of several subjects in this group indicate that this group is not a random sample of the U. S. population by age or sex or health status (or probably any other variable). Such a sample is not even available. Clearly, one does not go from house to house and collect samples of urine from random- ly selected individuals across the country. In fact, it is simply impossible from a practical standpoint to Obtain a completely random sample Of urine. However, what was chosen as a viable alternative was to select a group of individuals who had been carefully examined by physicians within the last 6 months, who were willing to fill out a rather lengthy questionnaire, and who were distributed at least wx't::.=_'._...a_ - .40. . .. M-i . ,- . . ‘9.- . lgm‘ ‘4‘: ' ~ » --~- ..- .. --—-—» - ’ -—_;;p—."W‘: 212 somewhat by age and sex (an earlier collection of urine from personnel within the Department of Biochemistry met almost none Of these criteria). In addition, as a test of the ability of MSMET to detect patients with a known abnormality, urines were collected from a group of patients undergoing therapy for neuroblastoma, a disease in which those affected, typically children under the age of three years, form a large tumor which is frequently Observed to secrete large quantities of one or more metabolites of tyrosine. High levels of vanilmandelic and homovanillic acids, in particular, as well as several non-acidic metabolites, have been detected in conjunction with this disease (64W2). In addition, several urines were collected from an age-matched group Of hospitalized patients to serve as control samples for the neuroblastoma patients. S_election Of M_SSMET library for urine samples. One Of the most time-consuming aspects Of this project has been the compilation of a suitable library of mass spectra of organic acids. As noted previously, this library was originally based solely on individual spectra of pure compounds, but was later expanded to include spectra from the urine samples themselves. This expansion has proven to be Of critical importance for two reasons: first, because almost half of the compounds routinely found in urine samples have been ones for which we currently have no commercial source of reference 213 compounds; and second, because it provided a valuable clue that led to a considerable improvement in the library. This clue was that match coefficients for library entries of spectra taken from urine samples were much higher than match coefficients for library entries taken from pure compounds. This suggested that the ion ratios taken from spectra of pure compounds were inappropriate for urine samples, a suggestion that was confirmed by inspection Of ion ratios for all compounds in the urine sample using a MSSMET “debug” option. Hence, all ion ratios in the library were updated based on the urine spectra, and a new library created. Success with this new library (BESTLIB) was far superior to that with the earlier libraries, and it has therefore been used in all urine analyses reported here. It should not be inferred, however, that spectra differ depending on the chemical environment in which they are taken; while this may be true to a very limited extent, it is much more likely that a shift in ion source conditions caused the shift in ion ratios, and that this shift occurred coincidentally at approximately the time the current set Of urine samples was run on the GC-MS. Regardless of the source of the shift in ratios, it is obviously important to continually check library entries against spectra of the urinary metabolites to maintain a set of valid ion ratios, as has already been noted. 0 .8 8 x .0 ... .. 9.. K . 8 . ..8 ... 9 8 .. 9. r9 .8 .. .8 . . r L .. . 214 Selection Of designate and confirming ions from urine samples. The problem of selection of designate and confirming ions for the 1ibrary to be used with the urine samples deserves further comment. McLafferty (74M1) has properly pointed out that some ions are more useful than others in differentiating a particular compound from others in the same spectra library. He has developed criteria for the selection of differentiating ions based on comparison of the thousands of spectra in a general-purpose spectral library (75P1). However, little attention has been given to the fact that ions differentiating of a given compound in one chemical environment are not necessarily differentiating in another. This is most clearly illustrated by Figures 36 and 37, where ion intensities at various regions in a GC—MS analysis Of the organic acids in human urine are plotted. Thus, for example, in Figure 36, where the sum Of all scans is plotted, any of the ions at m/e 205, 217, 220, 292, 333, 441 or 456 would appear to be poor choices for designate or confirming ions. Yet, as illustrated in Figure 37, any of these ions would be excellent choices in the first 100 scans, where they appear at low levels, if at all. In general, there is an increase in the number and intensity of occurrences Of higher mass ions at high scan numbers, but this effect is by no means uniform enough to be relied upon. The distribution Of given ions in fact varies markedly over much shorter regions than those plotted; ion composition may change over even ..9. 99990099 809 0809 088899800 080 090809000 09098908990 08900900 8 090 09 0005 08 800 0099908098 008850 00089. .858 0989-00 0 8 08000 990 809 008850 003 809 8000 90 99908098 089 1099800 0885 80858 0 90 80990089 0900 0980080 0899 90 09099080 092100 0899 08850 85000 09503 809 98080089 808990 0 9089 000890999 089 0850008 09 80080 89 215 0808000 02.00 099800 0885 90099.99 0 8 0809 90 809958989099 .00 08509.9 .1 1 2l6 0:2 000 000 Omm 00m Omm OON Om: tb‘bbDLDFbLbL—bbLLthhl—Lbbbbh LD’I1— PL3 00_ Lb—bLPt mm mmDOE Om... hibbipbtrlprbb mz0 0000199919 089 80 000899080 00900 0980080 8980885 90 0099800 .8. 8 008008090 809980908 089 90 8000 809 91920099 898 08509 0803 00899 8099809099 231 008008090 8008 809980908 90 00899 809980908 90 1999998098089 .90 0850989 232 10 0000.“. 900200000 05:8 20.820800 0000 0000 0000 0000. 009 000. com o J 9 9 _ 1 1 8 000. 0019930009808 019; - 00 0:00. .000. N 6.0030 m 0: 00050 M o: .1003 W 000890956000 8 3 218085190 0.91. 1009 N H1. 03089 W 00050005080030-00 \ .000. W 3 0:253:80; \\ X 0580000 \\ 10000 mm: c3052: \ 290000.005 \ IOONN 011000008909 1 .9 8 0000 I 8 r 8 mmnn=2 00 0003003000000 .00 000003 iii“ "L. - _ N0 mmnor... <>I ...o zo_._.<¢._.zmozoo m>F<4mm omN_._<_2moz 00¢ 0mm 00.00 0.00 000 00; 00. 0.0 0 0 <0 m% to N . 02 0 05200 ESE u u 00 W .v 00.200.00.002 n o co m .2200 0:002. u 3 0. 0 .000 o .2200 200< u a 8 E >00. . v H. .00» A 3 0 4 O 5 2 w .00? 3 N l w H. .000 m 0 .3 . A 000 W mo 00 . oo~ 0030300 .0000200 0 0000000 000 0 08.300 .0300 000 0.8 0080 30000300 00 300300 0.0000 003 00002 033 00300 003 Q 0000 m 000 .0 00000000 00003 30000 0000000300000 503 0300300 0:00 003 5 000000 0003000 033 003 5 0000.300 003 00000 0000300300000 0303 0000.500 000000000000 00 008000000 00800030 00 03 00000 0003 00330003000 030m .00 00000.0 000 00800000 00003 00 00000000 00000 003 00 0.32002 .5 0000008300 0003 00000 0330000035300093000030:00 000 000300 00 0003000000000 033080 0003000002 255 0030000 00000 00 0093 0 00 00000 0330000030 10300003000030-00 000 000300 00 0003003000000 .00 00000.02 256 w. m¢ manor. o_o< 050.036 ....0 20....dmkzwozoo w>_.r<.._um CNN—4453.02 w. .v. N. O. m 0 ¢ N O b p r p b b b r o N a w o a V m n noon 2 n 3 an 0 a 3 2.500 2.8:. u I o -000 W I. 0503030302 u 0 N .2200 0:002... u a H .2200 :03. u o ..oom W m >mx n N I. 8 room. w m N 0 -000. a o W H d H v 257 The overall objective of any of these data transformations is to increase inter-group variance while decreasing intra-group variance for each of the subject groups. Hence, it should be possible, given a sufficiently large group of urine samples, to test each type of transformation for its success in achieving this objective, Unfortunately, it may happen that transformations of univariate data will initially appear to meet this objective, but fail miserably when tested by bivariate or multivariate methods. Ultimately, it may prove more reasonable to separately normalize different classes of compounds (eg, citric acid cycle intermediates, sugars, amino acid metabolites, bacterial by-products, etc.). This normalization must be done carefully, however, so that it can be defended from a biochemical and physiological, as well as a statistical, point of View, It must also be done in such a way that important information is not lost; for example, if one class of compounds is abnormally elevated relative to others, this must not be obscured by the transformation process, CHAPTER SIX: EVALUATIONS AND RECOMMENDATIONS Overall, I believe that MSSMET and its associated extraction techniques and statistical programs accomplish the original goal of this project: to prove that a highly automated system could be designed for the quantitative and qualitative profiling of low molecular weight organic acid metabolites from urine. While by no means completely finished, MSSMET, even at its current state of development, has proven to be easy to use, precise, sensitive, and capable of processing an extremely large number of compounds. It has been successfully tested on a variety of human urine samples with results that suggest a great amount of medically-significant information can be gained from this type of approach. MSSMET has not yet been tested on the more subtle diseases, principally because of time and equipment limitations, but there is currently no reason to think that MSSMET will not be equally useful in examining these types of diseases. While current technology does not encourage the hope of using metabolic profiling by GC-MS in a routine clinical setting, it does suggest that MSSMET and similar programs may find considerable 258 259 use in detecting new biochemical relationships which can then be used as the basis for new, hospital-oriented test procedures utilizing other kinds of instrumentation. MSSMET would also be very useful in such specialized situations as drug overdose screening and some forensic applications. For the future, I can suggest several general improvements over current methodology which Ibelieve would make significant improvements to MSSMET or the general metabolic profiling process. Chemical separation procedure While satisfactory as a research method, the DEAE-Sephadex procedure unfortunately has few of the characteristics of an ideal clinical separations method. It is neither rapid, nor easily automated, nor tolerant of operator errors, nor pleasant—smelling, nor 100% effective. It takes approximately 48 hours to run and requires a great deal of manual manipulations, so that operator time per analysis is high. An upper limit of 10 to 15 samples per day per technician seems to be the maximum achievable. Samples require constant attention from the technician, so that the opportunities for forgetting a crucial step are high. The use of pyridine solutions requires that the entire procedure be performed in a hood. Recoveries of some substances (e.g., citric acid) are low. It is, however, reproducible, 260 which none of the liquid extraction procedures tested are, and this is its principal redeeming virtue. Despite its good overall reproducibility, the method has several disadvantages even in a research lab, where time requirements may be less critical. One is the need for lyophilization, which not only is the most time-consuming step, but also one of the least well- controlled. Chalmers and Watts have shown (7202) that the vacuum pressure and external temperature are critical variables when lyophilizing pure compounds. Although they may be less critical for urine samples than for pure compounds, a much more careful control of lyophilizing conditions than is possible with common laboratory equipment would appear to be desirable. An even more serious problem is the use of a barium hydroxide precipitation step prior to separating the urine sample on a DEAE- Sephadex column. While this precipitation eliminates a large fraction of the phosphoric and sulfuric acids present in typical urine samples, it also eliminates at least a portion of citric and other acids. In addition, it may affect recovery of other compounds as well. As mentioned earlier, (Chapter 5), this step in the procedure should be omitted if possible in future studies. In general, the DEAE—Sephadex procedure, while adequate, is considerably less than ideal. The best solution is probably develop- ment of some new batch separation method quite different from the 261 laborious column chromatography used here. Even if the DEAE— Sephadex procedure is retained, however, it can be improved by careful studies of all of the parameters involved. These studies should be completed using the entire MSSMET analysis system, since many of the preliminary tests on stability and reproducibility were performed with GC assays (before MSSMET was perfected) and hence may not be valid for small peaks not resolved on the GC traces. Quantitative precision. One of the most difficult, yet important, problems to be faced when evaluating a new procedure is to sort out which facets of the procedure are most in need of improvement, and which can be considered acceptable for the time being, at least. The former must be confronted immediately, while the latter may wait until some later date to be remedied. Influencing this decision is a variety of factors: ease of improvement, comparison to currently available methods, resources, external pressures (competition, funding sources), dependence of other portions of research upon progress in this area, and so on. An example of this kind of dilemma is the need for quantitative precision. MSSME T, which uses a repetitive-scanning l . . . I III Ilsll .I1 I. 0 _ .V . x w a _ _ . ,\ . k 0‘ ‘ V 00 . I 0 b u x 0 a“ n e . \ r _ a . v I\ . . \ <\ 0 . u , .v x x a v a. . . . u . .10 t a C 0 . a _ 0 V r0 . . , _ t. ( Q A 0 . n _ x t a . t ‘ \/ a o 262 based approach, has a quantitative precision of 5 to 10%. This is relatively poor compared to that of selected ion monitoring (0.5%— 3%), but comparable to that of many current clinical laboratory techniques. Does it therefore require improvement? For this particular case, a series of guidelines for answering the question may be established, based on statistical considerations. Assuming that the precision of the method is plus or minus 10% at all concentrations (this is probably not a valid assumption at low concentrations), and that typical biological inter-variation is approximately 100% (coefficient of variation), then how much do each of these terms contribute to the overall variability of the data? Assuming that the absolute mean of the group of individuals is 100, then (75w2, p88), the standard deviation, of is oaserved : Vi( (Md-hut )2 + (oat-cult: )2 O/obs = ‘V Q00)2 +C/o)z Jobs : 100.5- Hence, the method variability, which is 10% of the inter- individual variability, contributes only 0.5% more variability to the “true” variability under these circumstances. (It should be noted that this calculation assumes that the two sources of variance are independent, which is probably true). However, similar analysis for 263 an intra-individual comparison study where the variability of the individual is only plus or minus 10% (coefficient of variation), would suggest that a“: 14.14 that is, the overall results are 41% more variable because of the addition of method variability to the “true” variability of the individual being tested. Usually, then, the methods developed for MSSMET are adequate and will give normal ranges insignificantly different from those using much more precise techniques when measuring inter- individual variations. However, results from MSSMET for intra- individual studies will be significantly less precise than selected ion monitoring-based methods, and will require repetitive analyses of each sample on the GC—MS to achieve reasonably precise data. Whether a given degree of precision is unacceptable of course will depend upon the use to which the data is to be put. Other recommendations. Another major improvement to the current procedure, I believe, would be the development of a capillary GC—computer system for screening purposes. The cost of such a system should be much lower than a GC—MS—computer system, and the reliability of the instruments much higher. Such a system could be used to inspect large numbers of samples for the presence of interesting patterns or differences, which then could be examined in more detail utilizing a GC-MS—computer system and MSSMET. MSSMET itself should also be redesigned. Two of MSSMET’s 264 disadvantages are that it cannot find compounds which are not in its library, and it may mistakenly identify a single compound by two different library entries, since there is no provision in MSSMET which prevents the same data from being identified twice. A prototype program which would solve both of these problems has been designed by the author and preliminary tests completed; this program subtracts any positively identified spectrum from the mass spectral data file so that spectra left over at the end of the analysis are those NOT identified by the program; these can be added to the library for use in subsequent analyses. For example, this program was the means by which unknowns U50 to U91 were selected. This approach also prevents the same data from being identified twice, since it is removed when first identified. A third advantage is that it improves baseline detection accuracy by examining the whole run to choose baseline points, rather than using a “window.” The program needs further testing before routine use is possible, however. Another improvement needed is continued work on the MSSMET library, perhaps with an automated history-taking of data. A quantum jump in the quality of the library resulted from detailed inspection of MSSMET outputs utilizing the “debug” feature to examine ion ratios for all peaks found. Unknowns U50 to U91 would be considerably improved by further inspection of this type, since they are based on data from a single urine sample. Other compounds ~ . \ , . _ a C . I A . 0 , x \ . . v . \ . _ i . x x a 0 Nil ‘ r 'x 0 0 i _ _ . v _ x a V . 0 C . 4 t 0 x a c 4. . , x . l, _ t o v I I o \ . t . O 265 (for example, those listed in Appendix G, plus other compounds in PUBLIB) may also be re-added to the library as this process continues. Yet another improvement might well be achieved by designing a program to inspect raw GC-MS data for errors. The most common error on our system is a sudden drop in ion intensity for a single scan, sometimes associated with an abnormally high intensity in the next higher mass at the same scan. It should be easy to design a program to inspect raw data for such errors and correct them by interpolation. A better solution, of course, would be to correct the data collection algorithm so that such errors do not occur, but this particular type of error has resisted such corrective efforts in the past. Another recommendation related to overall quality of data collection is the design and implementation of a quality control program. While the hum an eye is reasonably good at recognizing abnormal results on large peaks, it tends to ignore fluctuations in small peaks. I recommend that a computer program be written to compare runs of the quality-control sample and to make a “go-no go” decision on purely objective grounds. I would similarly recommend that any gas chromatographs used in this program be interfaced to a computer and a computerized quality control program be designed for them as well. A history of past runs of the control sample should 266 be available to examine for trends in sensitivity, column degradation, and so on, similar to quality control procedures in clinical laboratories. Another project which should be undertaken by those interested in metabolic profiling is the purification of reference compounds. Either preparative GC or HPLC could be used advantageously to purify large numbers of compounds, since individual purification by more classical procedures would be very time—consuming. It might even be possible to purify compounds from urine samples in the same fashion. Pure compounds could then be used in recovery studies, to establish k-factors, and to obtain high-quality mass spectra. The establishment of k-factors, especially, would be useful for comparison to other quantitative methods, and would make clinical reports more meaningful to others outside this laboratory. While some k—factors have already been established, they are not yet sufficiently reliable to be published in this thesis, so k-factors are needed for all compounds in the library. Once many of the technical difficulties listed above have been solved, however, it should not be assumed that the task of the metabolic profilist will be an easy one. A great many philosophical and practical problems remain. For example, the problem of handling new urine samples needs to be confronted. Hilman (77H1) has suggested that a referral system be established, so that all 267 samples are screened by routine hospital screening methods and then sent to a clinician who will decide on the proper disposition of each. It is important, according to Hilman, to establish, in advance, the time period within which results are needed, and the likelihood of therapy being undertaken based on positive or negative results being presented to the referring physician. In essence, these suggestions are simply that we follow the guidelines of any well—run genetic screening center. In my opinion, it is well to establish a formal procedure of this type as soon as possible. Lack of such a system can easily result in chaotic sample processing, over-burdened staff, invasion of patient privacy, and unmet expectations on the part of both the referring physician and the group attempting to interpret the profile. The system must necessarily provide a series of procedures, including an exact protocol for the referring physician to follow in collecting the urine, selection of appropriate control samples, sample storage, transmission and secure storage of patient health and dietary data, provision of estimated processing times, computerized report forms and staff consultation on the specific meaning of results. It is especially important that a reliable means of communication with referring physicians be established, both to insure that urine collection protocols are followed and to insure that results are properly interpreted and followed up with appropriate treatment or further testing of the patient. 268 Another problem area which will have to be considered is a means of examining the effect that intake of specific dietary items by the patient will have on the metabolic profile of the subject. Hilman (77H1) has already noted several specific foods and food additives which can affect urinary organic acid content, but a detailed study of such effects needs to be made. It is especially important, there- fore, that referring physicians provide a completed diet and health questionnaire for each patient, even those with supposedly well- diagnosed diseases. Failure to keep a permanent, computerized record of such data can result in loss of a great deal of valuable information. One other suggestion about statistical studies is warranted. Some workers have cautioned against using metabolic profiling in a “shotgun” approach; this is, examining a wide variety of diseases until one is found which can be diagnosed by this means. I do not believe such caution is necessary. In fact, I believe it is in just such “shotgunning” where metabolic profiling is most promising, with a very high chance of positive results. However, I think that it is important to be sure that the person aiming the shotgun knows where he (or she) is aiming, and in the case of metabolic profiling, I think it is very important that the question be clearly defined in terms of what is possible. Thus, for example, it is not enough to ask, “can I detect disease X by metabolic profiling.” It is much better to ask 269 such questions as: “Can I find more than 13 compounds that differ significantly between the group with the disease and the appropriate control groups?” “Can I detect changes in the levels of more than n compounds when a single subject (or group of subjects) is stressed in some fashion (onset of disease, exercise, unusual diet, drug ingestion, etc.)?” “Can I detect changes in the relationships of more than n pairs of compounds between subject or treatment groups?” In each case, it should be possible to predict in advance the likelihood of a given number of significant differences occurring by chance; this number must be considerably exceeded to have confidence that real differences do indeed exist. Some differences or relationships may prove to be more interesting than others, because of known or suspected biochemical relationships; hence, the problem is not entirely one of statistics but requires significant interaction between statisticians and biochemists if progress toward under- standing test results is to be made. 2'70 Speculation on long-term prospects At the outset, the prospects for relating urinary metabolite concentrations to the diagnosis of specific disease states may appear to be minimal. After all, urine is the “dumping ground” of the body metabolism, so that what is observed in urine is a mixture of metabolic wastes from all of the body. In addition, there is the further complication that the process of urine formation in the kidney is itself quite complex, with wide differences in rates of diffusion, secretion and readsorption among various types of compounds. The major complicating factor, however, is the variation in excretion levels among individuals due to differences in both diet and enzyme activities. This inter—individual variability is the apparent bane of the metabolic profilist, making anything but longitudinal studies of single individuals seem almost pointless. However, there is hope. Looked at from a biochemical point of view, man is little more than an extremely complex set of chemical reactions, each related to one or more others, which in turn are related to yet others. In this context, most disease is the imbalance of one or more pathways, whether produced by dietary deficiency, inherited enzyme dysfunction, bacterial toxin production, viral infection, or almost any other causative factor or combination of factors. When these imbalances become sufficiently large, they 271 appear as gross physical symptoms that can be detected by the physician or patient. Very few diseases have the visible symptoms localized only at the site of the imbalance; effects are frequently observed at numerous locations in the body. From this it is possible to infer that diseases must be affecting a variety of biochemical pathways in a number of tissues or organs. Presumably, at least some of this imbalance will be reflected in the urine of the individual involved. While direct sampling of the affected tissues would be preferable, and sampling of serum constituents a good second choice, nonetheless, urine sampling may in the end be preferred because of the ease in collection of urine and the higher concentrations of many metabolites in urine compared to other body tissues. An extremely important point in this vein is that regardless of the tissue or fluid sampled, it will probably be relationships, not absolute levels, that are important. Undoubtedly, some diseases will be indistinguishable from one another because they have the same generalized biochemical effects; others will be undetectable because Q16 changes are too localized or at such low concentrations that more sensitive methods must be used. Yet others will be indistinguishable from dietary effects. However, for many diseases, it may be that the real value of metabolic profiling lies not in the “shotgun” effect, but rather in the potential for sampling some of the complex 272 inter-compound relationships that exist. Since all humans have essentially the same metabolic pathways, it may be expected that certain relationships will occur repeatedly in urine samples from “healthy” individuals, but be deranged in certain types of illnesses. Pauling (68P1) has argued for a concept he calls “ortho- molecular psychiatry,” of which the most famous (or infamous) example is his ascorbic acid therapy. He suggests that some mental diseases may be cured by obtaining the “optimum molecular concentrations” of substances normally present in the body. While f . I do not share his optimism about being able to use “orthomolecular therapy,” I nonetheless believe his approach of examining health from a biochemical vieWpoint to be very useful, and in the long run, very productive. I would further suggest, however, that there is no such thing as the optimum concentration, but rather, an optimum 1evel consistent with the current pattern of concentrations of biochemically related substances in the body. It becomes, then, the task of the metabolic profilist to discover those optimum relationships, and to measure how they have been altered in each Specific disease state. APPENDICES APPENDIX A Diet, health and drug questionnaire The following set of materials was provided to each subject where possible. The materials include a set of instructions, a medical history questionnaire, a diet questionnaire for the day preceeding the urine collection, a 72—hour drug ingestion history, a survey of whether the protocol was followed by the subject, and a consent form. Subjects were each assigned a unique number prior to the collection time. Completed questionnaires were coded into computer—readable format by a program specifically designed for this purpose. 273 274 000000030000 0000000 m000 000 00.000 .005 00m: :00A0 :0: 0:: 00:n0:: AAAo0nm an AA .02 A mewA A V. Amuo:uo:: 0000000 :0:00 0m: :0 m0uu0:mmA0 oxosm :0» 0: AAA 0mm0mA0 0m:3 AAAownm 0000A: .00» AA .02 A :0 m0> A AA wmuwmAu 0:: :0: :00000 a A: 000:0:0 chwn 0A0:0:::0 :0» 0:: AoA .A.000 .0mmumA: 0:00: .m000:~A0v AAAmA:: 0:A:0m0: 0mm0A: .mwx AA .02 A :0 m0> A AAwAmV0mmmmA0 A:m 0>m: 0A000:::o :0» on Am .AAA0A:: 00A:0000 0mu0A: .00» AA .02 A :0 m0> A VA pmcoAum:0:o :onme 0:: 0m: :00 0>mx Am A.000 .wumA 30A .0A:0AaoizoA .:~A:mu0m0>v 00A:0m00 AAumA:: .m0x AA .02 A :0 mm» A VA A00A: Aquwnm m :0 :0» 0:< AA .A.ouw .m:0~AAA:0:u:u .mmcoa:o: .mAAAa A0:u=ou :u:A:v mm::0 :oAunA:0m0:n mm AAuz mm .A:::xm :m:oo 0:: m:00AA:: wooA: .chEmuA> .:A:A:m~ mu :o:mv m0:AoAv0E :oAuaA:om0:0-:o: 0::A0:A .:00> ummA 0:0 :A:uA3 0AAmcoAmmooo :0 AA:mA:m0: :0xmu 0>:: :0» m0:AoAu0E 0:: “WA: A: mqu:00 eeAu mA:u m:A::0 00A: m :0 wu:mA0: mA:v 000A :0» 0A: 00A:0m 00A“ 00:: :0>o :0» 0:0: .m:003 .mnA ~0moA :0» 0A: :0:5 :0: .m0A AA : :0 mm» A VA ~00mA03 ::0A :A m0::o: m :m:u 0:05 000A 0A0:000: :oa 0>mx Am mu:mA02:0>o A m u:mA0: =Aa00A= ::o» :0 00::00 m— :A:qu A A0 .u:mA0::0::: A AA :0» 0:: .vAA:: 0:0: Am:0:00 :30» m:A:0uAm:0u Av mAmus m_:qu:0E< .:0000n A AA .0mm:0>~_:u :u:u 00AA0:0 :A :0:0o: A An :0 .000m 0:0 0:00: A .Am:0:0m :A An AO0A0 ::00A 000 :0» muOOA 0:» :00Amcoo :00 0A:03 A AVN z AVA xwm AN ow-AA A AA oA-A0 A AA oe-Am A A. om-AA A A. OA-AA A A. om-AN A A. o~-AA A A. 9A-: A AA 0:: AA .:00::Eou 0:: :A :oAumE:00:A ch000 :0: 0:: 0m0:0 .m:0:s:: AAmEm 0:0:0A .AA:0AAAA: 0:05 was: 0:» A0 wAmAAncm :00::Eoo 00:05 :0A0~E:oA:A uwuuAEo .coAumw:: A:0>0 :0xm:m :0» 00:0 AcuA> mA 00 .0AnAmmon 0: AAA:: mm chzoAAoA 0:0 0:0A0000 0m~0A0 zmmzaz .00: 0» ch0m 0:0:0: 0:An::oA000:a :0A00: 0:: 0::0 .0:0:u 0Ansmm 0:0 :qu 00 00 00:3 vAoA 0: AAA: :0A .:0A :qu 0:03 00 0Aasmm 0:0 ch:m .Am:m: mu umo:xm0:: 0mm .::03 cu AA m:A:: 00 :0:5020: o: 0::m 0: AAA: :0» 0:0:3 AA 000A: .0mAz:0:uo .A0AQme0 :0: .:mo ::00 :A 0uAmuaov 000A: :Aoo m :A 0A:Emm 0:0 0:00m .0:00 :o ::o: :0 :0: ::03 um 0AA::m :0: AAA: :00 AA .000A>0:: an: 0:: ou:A 0Aasmm 0:A:: 000A: .:mA: :00 AA :00»: :AA: :0:Amu:ou 0:: A0 00A mu:o 0:0 00:A: Ame :o> .:o AAwuwAasou :0::m:m mA AA :Amu:00 0:02 A0:A:: A0 :0cAmu:oo 0:0 :0 0:0 oAummA: 0:0 0:: .AAmEm 000 :o 00:0A 000 mEmwm :0:Amu:oo 0:: AA 0::03 00: o: .m0::oEm m:oA:m> 00::A:: 0A000: m:oA:m> .mAmxAmcm 0:0 :00 00:00: mA 0::osm AAmEm 0:0A m mm 00: cu m: :0: 0:A:: 0A00AA 000 m:A>m: 0:0:0 0:003 00: o: .uwAA00 0:0 ou:A AA000:A: em 00 0:A:: mmouxm 0:: 30AA< .0:A:: 0:0 :0 :0::A000: 0:0 youAAoo ou 000A>o:: :0:AmA:oo 0:» 0m: :0:A .00AA00 0:0 oacA xA000:A: om ou 0:A:: A0 :0A0:o: 0m:AA zoAAm .much:: 0: 0:00m umzoAAoA mm 0Aaamm 0:A:: 0:0 000AA00 .0mm::mw:: :Aunw 0:000: ch::oE 0:0 :A .0:Am::0AAmm:u 0:0 :0 mA:u 000: 0:: .u:mA: 0:» mo 0AuuAE 0:» :A wuacA:: 00 000: :0» AA .cm on 0:: :0» .000A 0:: 000Aasou .0:Am::0Aumw:0 0:0 :0 AmVE0uA 0:: A0 000: a 00:5 0mm0A0 .0EAA mA:0 :0000 005:0:00 mA ch:0000m .:0>030: 0:: :0::A umMAxmw:: 0:0 .0A0500 ::ox mumcou :0» 0:0A0: Am: 0:A .AAmAu:00A0:00 0AAaqu 0:0: 0: AAA: 0A Am:u :OA ucAE0: 03V 0:000A: m::: Ao:ouAm :0 mm::: 0::0 0A: :0A 00:» 20:: 00 m: :00 :0000: :u:E AAAm:00m mA 0A .0:A:: :0 00:00:00 0:0 000:0: on :30:: 0:: :A:A:ma 0:A A mm::0 0Aqum :0>0 wozAm .:0Au0: ::0A 00 u:m0:0:EA 00::0 0:: 0::AucoomA0 00: 00 :00 00:0 u:mu:0:EA mA 0A .:0>0=o: .0:A:: 000AA00 AAA: :0» 00A: 0:0 00 :oA:: Amxmu mv m::o: NA :0: A0:00Am :0 mm::0 :oAunA:om0::-:0: 0:0 0:00 00: on .0A0Amm00 AA 44 HmDUHh .m .w .A .w .m .0 .m .N .A .>:Au quA Az 0A AAQAAAAAAnA A .m_m:cA>Au:A 00m00mA: ucm AmE:o: :0 .00A: :0 muomAA0 0:0 .00>A000: :0» mAmA:0umE A0 00m um:AA 0:0 00000:: mA:A :A 000:0ucA ::ox 00:00:00m 0: .000A 0000 00A: 0:00: 0:0 c0 AAAA 0m0000 .000000 00: “A 5000 :0 00 .A=.00m0:nE0: 00 0000:000000 .mAA00z 0:0 =0000:A::00m .0m0m000: :00: nmcoquJLumzA .»0>0:m 0:0 00 0000 000A 0:0 :0 000A>000 00000 0:0 :A :0A00E0000A 00:0 000: .»03»:0 000 000 00000» :0» 0A:0:m .00>0:0: . . . .E.0 A 00000 :0000 00000 000 00000 0: m:A00:00 »0>0:m 0:0 00:0 000: 0000—0 A=0=A02 ch=0oz 00A0000= 0 000 0002 0:0 0m: >m>x=m >¢ 0000A00> .mw 000A00 0:0:A00 00 chAA0zm .mm =EmA00E:0:0= 0o mA0A0:00< .ww mm0c0E0: .Am 000000 :0000m .om m»mE:A0 :0» 00< .mm 00 »:m A00> :o» 00< .mm A0Am0zs0v »0oe0e A0 000: .AA 000050:mA: 0:0:0000 .mA chxA0z000Am .mn :oAmm0000o .00 0000A0002 .MA »0A0oA000:A 00 ch0000 .NA »0003 00 »0:00:00 .AA 0000000A 00 0000 .00 :0A0000c00:00 0000 .00 0A::o00 00:0A0050 .00 000cm:0>002 .Ao 00:0 00:00:: :00: 00>m .00 000:0:0A00:00 00 000A 00 m:A0:A00 .mw m:0AmA:>:00 00 »m00AA0m .00 mAAAAm»EoAA00 .m0 mAm»A0000 .No mAuAchcaz .A0 m520000005 .00 000A 0000A:0< .mm 0000mA0 :A0m .mm »m00AA0 0000 .Am 0:0A0z 00 0000 0:0000 .mm 0:0A030euc: .mm 0:0A0300>o .vm 0000A-Ao .mm 0:000000 .Nm »m00AA0 0000 .Am A00A00000A 0o .00>AA .:00eo0m .00 A0588 .00 0000000 0A:0500:0 .00 »mA0=000 .00 0A0000000::A .mm 0A0: 00 :Axm »00 .mm »AAm00 000A: 00 »0:00:00 .Am 00:0003 0000 00 000000000 000000000 .mm 0000 00 00:0: 0A00 .mm :0A0m:0:x0 00 00:0m .0m . 0:0A000 »m0w .mm 0A::00A A00000 .Nm 0A:=o00 0000»:A .00 000000000: .0m :0AmA> 0000 .mN 000:0000 .mm 000 ch:::x .AN m0000:0 000m 00m:0A000 .0:0:0000 .om A00>A00x0A m:A0A:000A :0A000A0m:00 .mN 00A00300 .AN 0332003020030.8 0A::000 00A00000 0000 .0A A00:000:0m .mAAA:0»mv 00000Ao A0000c0> .AA 0000:0Ao 00m:m .0A 0:A0: :A :AE::A0 0o 00m:m .mA 00000Ao »0:0Ax .0A 0~:0:A0:A .mA 00>00 0Ao:0»A .NA A00>00 0:0—:0::A wAmoAA0000m .AA A00>00 00A:0:0va mAmO0A0::0:02 0:0A000wcA .oA 05:0m< .Om 00>00 0A00E:0:: .m 0000 00 chAA0zm .00 00>00 00A000m .m 0050:: 0000: .00 0A00:0:0Ao .A 00::000 0000: .00 A0000:00 00:00 m:0A> .0m .0 0000: 00 m:A0::00 .00 x00AA0Ew .m 00:00000 000A: :00 .m0 x00:000A:u .0 00:00000 000A: :mA: .00 005:: .m :0000: 00 000:000:m .mv A»00 my 000000: :0E00o .N 000:0 :A :A00 .N0 0000002 .A 0K00 00 «0000 oz 00> 0000 00 «00:3 02 00» .0A:0A00A0 000: 0000 0:0 00 mAm»A0:0 M0005 000A: 0 ucA>00A :oA0m0:u :000 00zm:0 00 :A00000 0m 0>0: :0» :0A:z 050A0000 0o m0m00mA0 00 0005:z :00: 00 :00»: 00:0A0 .0000 0:0 :A 0EA0 »:0 0 00: 0>0: 00 :0: 00AA000:0 chzoAAoA 0:0 000AOE00 0000—0 {'l/ 05:00:03 002000 is: .N. 0.0.0008 .2. "00ch.0 276 000: 00:00 .00 :000E:0 .mo 000 00:02 .N0 000 0:000:05 :050: .00 000 0000030 .00 000 »000:u .00 000 00000 .00 00:00 .00 000:3000 .00 000000 .00 00030000 .00 00000 .00 000:0000 :0000 000:0000 00:00 00 00:00 00:00 .Nm 0000500003 .00 000000030000 .00 000:3:0 .00 00000000000 .00 0:00000 .00 000:0 0::00 .00 00::00 .00 05:00 .00 000:0 000000cw0 .m0 000000000 N0 00000 .00 00:0000 .00 000:0 00:000 .0m 00:000 .00 000:050: .Nm 000:0 00000 .mm 000000 .00 000:0 00:0000000 .0m 0030000000 .00 00000000 00:00 .mm 00000 .00 I! 000532 'w'dz - z 'W'stfiw'POL 'w’P OL‘V 000:0 »000::00o .om 00:00 »000::00o .0N 000000:u .wm 0:0000:00 .NN 00000000:_m .mm 00:0:00 .mN 000000>< .0N 0000000< .mm 00:000F00< .NN 000:0 0000< .—m 0000:: .ON 000:0000 00:00 a 000:00 :00» 00 »00P:o0 00 000: 00:00 .00 00000 000 00:00 .00000000 .0000 .0500:m .00 00030000 :000 00 :000 .00 000> .00 .mp .00 :0x0w:u .m0 000000 .0000::E0: “0000000n000 .00002500 .000000000000 .NF 0000:00 0000 .50: .0000 .:0000 .0— :000 00—:o0 .0002 000:0000 0005 00:00 00::0o> 000:0000 000 0:0 000m :00: 000 00 E0000 000 00000:0 00:00 0:000 0000>000 0000000:0 .0002 000:0000 E0000 0:0 .50000 0000:0 00020000:m 000:2 .000: ,— r—qumnorxooowo 0000 .50000 .0000:u .:__z >0>m=0 >m<0000 "€ 'w'd A 'm'd g-‘m'e m 'W'P Ol‘v 0.0.088 .04 00000000 00:00000 .00 . 000:0 000000 00 00000000 .00 000 000000 .0000 .00 _ _ :0000 .0000 .00 000000 .00 00000000 3000 000000 0000 000000 00 000030 00:00 .000 0000 .00 0 000x05 05000:03z .m0 0 00:03 .00000 00000>00 00—o:ou—< .wN— 0030000 .N0 000 .000 00002030 .00 0 000000 .000 00030 .0000 .00 0 0000000 000000 0000 .000 >.00000 .00 0000 0:000:00 000000 0000 .000 00300000300 .00 _ 00000 .000 0000000 .00 0 0000000 .00 000000 0:0 00003w 0030000 00000300 .00 00000000 .00 0 00000 .00 x03 00 0000 .00000 .Nm _ 0500 .00000 .00 w 03000000< .00 0000 00:00 .0 0000:0000: .000000000 00000 .0 :00000 .000000000 00000 .0 0000:0 0300 .000000000 00000 .m m 0 0 0 277 0:0:oo0 00 00000 .0000 . 0000000 .0000 . 000030 . 003: 00 00000 000030000 00000 00:00 .00 000030 030000 .00 0 0030000 .00 00300 000500 0 0000 :00: 000000 .00000 .00 . x>0c .00000 .00 0 0 0 0 00030000 00000000> 0 000000000> , S 0 0 0 0 000000000 00000 0 0000 .0000 0000003 .00 0000000000 00000 .000 000000000 0 000000000 .00 , 0000030 .000 000000000 .00 W 00000 000500 .000 000003 000 .00 0 030000 000500 .000 00003050: 00 00003000000 .00—00 .00 00000000 .000 00000 .00 . 00000000 00020 .000 0000 .00 0 000300 .000 00000000 .00 0000000 .000 0000000 .00 0 0300000300 .000 00000 .00 0:000 00030000 00000000> 0 000000000> 0.00000 00030000 00000 —'w:301 'M'e 0\—V 000E:z 'Lll'd ['E .w.d €_.Lu09 0L -w'9 Ol-t 'w'd [ — g 'w‘d g ___.___—_ m 278 “00:0000q :0:532 000: 000000 :00 0000 0:0:05 0 00 :0:00 0_ .:000:030 :0 0:0:00 :0: :0:00 00 0:300:000 .0:050:00x0 0:0 00 :000300:00 :003 0:050:00x0 00:0 0:0:00 00:000:0 000300: 0:0 00 0:05500 0 :0>00 0: 0003 0 .0:0000 00 0 00 .00:0 0:000:00:3 _ Am .03050:0:0 0000000500 :0050: 0003 :0:00 0:0 00:0 0:0 00:0000:00 000:00 :00: 00000:0 0: 0003 000300: 000 00:0 0:000:00:3 0 00 .0000c0: 000:0: 05 00 0: 0003 :00:3 000000: 0:0 003vo:0 00: 0003 :u:0000: 00:0 00:0 0:000:0nca 0 0m .00000:05_:00: 030:003 .0500 0:0 00 0:050:00x0 00:0 :0 :000000000:00 05 03:00:00000 00 00:0 50 0 00:0 0:000:0oc3 _ .00000000:00 00 0:00:00 0000:» 0:0 .0:050:00x0 00:0 :0 00000000:00 00 003 0:0 :0 000:000 :00: 00: 0>0: 0 AN .000500 00:0 0:000:00 mo 00000:0 0:0 :_ 0:0:0::_ 0:00: 0:0 00 0:030 000:» 50 0 0:0 .05 00 00:0000x0 :00: 00: :0:0000: 00:5 .:u:0000: 00000c0000 00 00000:30 :00 000: 0: 00 00 0:0:3 00:0 00:0 0:000:00:3 H 00 000:0 0:0 00:0500000 0:0200000 0:0 00 000 00:0 00:00 0 .0:00:00 05 0:0>00 :0 0:0:00V 50:» 0000500 0:0:3 00 003 0:0 :00 0:00:00 05 0>00 _ 2000 02mmzou 0.0.0008 .00. 00500.0 .00:00 00:0 50:» 00300: 005 00:0 0:000000_:30 0:0 00 0000 0 0:0_ 000 0030: A V02 0 v00> .0000000:000 0: ___3 00:05500 :30> .5:00 00:0 00 :00: 0:0 :0 00000:: 00:0 00 0:000_0: 0:0:0xc0 :0 00:05500 :o 0:000000030 0:0 000:: 000000 .0000 00:0 :0 0000 00:0>0: 0:0 :0 0000000 0>_0 000000 .=0:: 00:030c0 0:0: 0:0000000 0>0:0 0:0 00 0:0 00 00:0:3 :300 00 00300 0:0 00 000300: 0:0 0:000:0:00:0 :0 030003 0: 0:005 00:0 :00005:00:_ 0:0 00 000 003—0:0 300 000 .m _ 00000:00:: 0:0000 0:000: 0:0:3 0:0 0000000 300 000 .0 00:0::05 0:0 :_ 0:0:3 0:00000000 :o 0:00000:00:_ 0:0 300_00 300 000 .m 000:00::0000030 0000 0:0 :0_00: 0:0 000_0500 300 000 .N .0000 00:0 00 0000 00:0>0: 0:0 :0 000 30» 00:3 000:: 0000—0 .00: 0: 0:00000__00 0:0:3 0:0 0:000: 0:00: 0:0 .5.0 0 :0000 0000 300 000 .0 02 00> zo~0um::oo 02:0: 0O0 0000x0010 .5.0 .5.0 .5.0 .5.0 0:0:0 :30» 000000—00 300 0:000: 0:00: :0 0:0:0>0 0:0 0:0:00 000:0:3 300 000 0000500 00:: 0< .00 .003:0 :0_:3 0000000 00 00: :000:0 00303 0 .:0>0:o: .0>o:0 00 00 00 50:0 05000 0:05 :0 0:0 00530:00 0>0: 0 .mp 0:3ocx 00 00000000 :000000005 :00000:000:0 .__ 0:00:3 0000000: 003:: :0000_:u00:0-:o: :0:00 .00 0000:0>0: 0000:0000 :o _0:000< .0 :00:0: .0 om: .0 0:000_:0: .0 :_:00030 :0 :000:< :0 :_:000< .m 00>00000m .0 0:.00 3:080 00:00 .0 0__00 0000 .N 0005000> .— 02 00> 00:030—000 0:0 00 0:0 0500:00 :00 000 .:o_0u0__00 0:0:3 00 0500 0:0 0:00000:m 0:30: N0 0:0 :: >m>02m zo~0h\ .M 2k\ .m bM\ amm.mw.ooaatan H.Hm ammm oawmi amm~mm.aooa.am «.fim mmMN oavms «m.vmm.~maamdm.mmafitNON HiNoN mmamé mN‘vatmmN.mmN «.mmm oaam oqau§ oM«.mmm.oov~mmN n.nou mum mwmr mm.mm.ama.maw.wwimmm.aaaa.ama «‘HMH mama mnad* Nnm.MMfi.flwm.mm.ovnfmfimammma.oma H.0ma mama “add! a~.vmm.sm.waN.QNH.m&N vafig am :an .MH.“ “mm.“ .mma .MMN «.MMN mvammu.mMM.amN «.9mm wmmd oam«* fimwaimmmtmvmtmmmtnmdm.mmm.wvm.wvv aimmN om.amm.7m.wnm.wmmimmm H.NMM Mm.mmm.mma.«am.wmm..wm «.mwm mmia .am GU\ .vm Xxx .mm GM\ .mm.o :u\ .a mmx roeooav mo\ .0 :4\ .N @2\ 2E31 9.203 .8 EDGE mu imam.mw .nmm.oav ‘«~N.oam«.fima wwm .xqmioeofi.¢am.mum .oMH .eao.arama looa.fi.¢ou xwma «amu xewunnyxoanov "mama x2: wm . Auzhxo u~>=m>mc v: mam * ooa«.xem.oem .mum «mm iqmfi.axa .mmeiaooa.xqa.mm~ .mm .oeo.d.xom .omo.a.ama mama mma« u_zo4¢:4>x»m= mm . Au~>mhzm>xome>zlmv u_o=mp=m>xoza>x-m as x ogo«.mmmiosm .mamimmm xqu . soma‘m«~.oma .oma .aao.a.mmm .oaa.a.oma waa mafia xu~>mkzmrxomo>xluv oHozchnm>xoxa>xuv Mm . nmzsxo u_4>xoyngc ms~xo ushmocoxo he * amm .meuioooflimmm mosaimguiaxm .xxfi.oom .«ma logo.a.mv~ .omm.«.mem amma mmaa quOJc: mm * Ausnou>ncv u_pmu¢>x0¢s>z ma * ooa«.mx«‘Mfi« .mmfiioom .mva new .mmm.aoa«.dmm .oea.a.mm« loco.d.«mm mum“ add“ Auszhna>mnv o_oz¢pzm¢oxo-v mm a N: «mm * mmm rmfi~.aoa«.mo~.mm ixxa mam imam.moonismfi.oam immaiamm .mm .ooo.aimam .ooo.a.om« mmwa flea“ Anozuuynoc nohmpwzmaommxm.m.fi mm * AUHPQ¢4V osozcaomayxomo}:-~ m * mmm .xx«.oooa.mm« mm .mmmiaow .mm~.omofl.fimd.nm .mm .aaa,a.mm« .ama.a.flm« vmma mxaa n: mam . Aufimykszmsyxomayxucc u~ozcmouan>zhmzlm-yxomeyxnm m a wow lfim~.ooaa.mum.¢xm .mma ooofilmamimmv .mmdimmu .mma.xmx .mvfi.som .mmfi lass «.mam roea.«.mom mew“ «cod mm: vvm * on: Men * aaofi.mam.¢xw ~M3 mmm .wmm.aoa«.xmm .oos.s.mm~ .ooa.«ixom wvma omm «m: Men a a: man « amoaixvm.mmm .mo~.mmm .vom .«m aux iaao.«.xv~ .ou 22x vmma .am :hx m: mmm . .mm ::x .mauiooofi‘omu om“ .nxa .v >px .moe.d.oma .m hmx «mud mm.oam.MMfitmmuiomafi.om~ usncxo mm a H.om~ vmm .mam.NMH .«ma.ooeq.xaa coma .moa.a.xae u_aoap* mama ‘ .o aux mm: “mm * ragga zzx mmv tom«.NVfi .mm«.aamfi.mwa .mfi aux . oo.fiimm« . a mama ma xxx Anomuuuo m: xmm * .mm aux 3208 .8 85on NN« .mm~.mn .ona.wmw .nmfi.aaou.mma mm .mm«.ooouiomuimov .mM«.vwm .woa .soo.«.mmu .oao.«.om« omma mmva Aw xcmmuunzoutpnau4>xhuzlmv ow: NmM e ao—szou_zu u~4>xommcumz~o~m>mlm nm e vs“ .movtsaaa.mm~.vov .gmm.am« .mamiwmm .no~ mum .mmaiosou.vwa .ooo.w.~m~ .ooo.a.vmu wmma mwva Au—zom:~>mwv mu: mom 4 ofipwuc4>zwxm mm * mmm ~M-.mam .mm«.oam«.mva ooo«.aam.mam .wma tooa.a.mv« .aao,aiaom mmma mmva Ma xcmmuu_zooc~:404>zbmsumv mm: 0mm 5 mm: mvm } mom .mmm.emo«‘MmN.amM .oma mao«.mvm.mm tame .aam.a.mm~ .eoo,«.ena ovma mavH Aunnzzv u_o~om2ch:mrxoma>: mm * nunmncxv one—ouzmh3m1m_u mm . omd .mvmivw .«Nm_amm«.mom oom«.m¢m.omm .mwa.~mw .maa .oom.«.a~m tomo.«.mvu oama “av“ Ao_4c:cmh_uv u~oHouzchamurxhwznmu>xomo>xnm mm * u—mmzau Hm * moo«.mm~.ama .vamimvm .mma mooa‘fimm.anm .Nm~.mmm .m«m.mom .mma .oomia.mmw .aoo.a.nm~ mama mama Aummch=4u4>xku2~olmxmv Ono—QNchzwm451bwznoIM.m mm * Nuz|¥zn mm * 0a oosfi.ammimVfl .mmfiiamm .mma mm .~m~.aeofiixvm.oom .mxa oo .oam.a.flm~ .oeo.d.xem On amva me« Au_mcF:4oc uHo~ouzchzma xx ; . o~z_oo:m am * xxv .mmm.oooa.xvmixxa .mm~_wm~ .xau xmx imam.xnm comm rage «.xew .aoe.«.~m~ amva awn“ AnmmzomH szPucgozommxhv mm: mvm * nohzommzh>xcmo|vv a“: MQM * 06m .mma.@cmu.moa .900 HmeN .mao «.mhd mmVH NmMH Ao—zomm1h>xowov mm: wvm i umomzum Hm & mnN ”HNM.th .mfim‘aaow.mmm.nmm .mmd mmm ‘NmN‘mmH .omN .an.fi.HNM .ooo.«‘NmN nmvfi mmma mm: mvm i Au~zomxh>mm>xomolvv ma: Nom 1 vmm xHNm.mmm .wvwimaad.MMN‘mmm .flMN.mNm .mma can .Vdmioaofl.mmm.mm .Mwm.w1 rudm .OQQ,H.MMN .ooo.d.mmm nmva mVMA Au~zompuh>XDwOINV v“: wow I uamoxmmorm am I cum .«mnrooofi ~Smémm .xwa mud imam.awm imm~.eooa.mam.uum .mma .ooo.«.m«~ .ooo.«rm«~ vvvq . wan” Auuzcmhuh>xomoumv Md: mam e Aummwu>4u4>xhwtumv m: «cm § nmq .«mm.aoo«.maN‘«- .mmfl ox imu~.vx« imv~.aoou.wxa .ooo.«.m«~ . . ‘ ovva mag H fhu em: mvm . “an“ fimxonxz: N? * 2(33 Am:_xo u—acFDJOOFNXIcv mvw .nnmxasaa~mmm .aoa.«.mwm Gama mE—xo u—o~omzthzmmDXOIN mNH Q mmm ‘vmmtwnw .mvmivwm .mmN.@«m .va.am0H.«MN nuntchaJcomrmv van « xcwmqwzohucgozomcmc (Na i NMH .mmm.aooa.omN .aao.fitowm coma Au~momhv u~ozcmomm4>zmxmxml>xoza>1IM oma G Dumaxmmm.mm« .mm«.omd .wmfi.mmm .Mmd iaao.«.mmm Mama Aomamt~mv u~o~owz¢~mmx mad & emm .MNv.NNN .memxaoau.mmw.mvm .mHN .Qoa.«.NmN meA.. oumchmch MNH b vm .Nmm.ooo«.mmm.vwv .mw«.mmm .mhd .mao.a.nm~ vmwd 1m: mum § 900 .mwm.oom ~vmfi .oao.«.mmm vmna o—wmuc4}zwxm>xomoyrlo raw * mmmuxmvmxmsm .dMN.mmH .wom .Qam.fi.mvm mwwa mm: VHM § vm .mmw.oaowran~.mmm .MNd tomo,«.flNN nnwfi A0HDMDuINIJ>IHwt>xomo>Itmv Mm: mum * mm ammNiswod‘me ro@&.H.wmd Mnmfl o—4}xommcuumlm200~40mm>mum mad 1 omv .NmN.QOQH~nmN.Nav .MNN .0@o.n.nmm meH u—on:wm>xomo)1|: add i anu‘awm.MNM ‘mm«»VmN ‘qna .aao.a.dmm mmwd mm: vmm § .nmv.mvn .mmM.mmM .Mmemmm .mcm.ooo«.MMN .00Q.«.MMN omma mm: vmm i 3388 .E ambush mvm .mMMmem xmom.asaa.mom .omo.w.m0M ammu.mmm.mmma.mmm.aomd.mmfi.mmw .mma .aa&.«.MmM hvmd Au~th34O4>IhmzIml>xozokIImv UHohowzchzmm4>thIIMI>xomo>xlm odd 1 moodxmfim.wmm xvom.no¢ .«mH .mag «.mdm mmmd Hm: mmM e mama.mvm.&omuinvw.amm .MoN .Gam.fi.mvm mmma Aommch340}x0moyxlmv Q~0Howzthzwm>xomo>xum mad 4 Goad.«wm.oow .MVd NQSQAJWN mmwa man man * MNm .«wamecfitmmd.va .«hd .oao.«.mma NNma aw: Nnm * amo«.mwm.@m .mom‘om .mmd .aae.a.th «Nmu UHQNZwm1xoma>Ilo 30a r acourmm~.ovv .an .aa@.«.mw~ Au~m~ot4>IhUIIMv u~o~0mzcxm14>thzvm mad i Mm xwmv.m Kmawtwmm twam.ooofl.mm«.vvm .wmnxmmm ‘mNH.HMm .knfl y» [“l M M N OI "l .ooe,w.xafl mama «m: mum * emoflimx~.xwfi lmm«.mme .Nm«.mmm .mma .aoo.aimx~ emmfi nu~xnocv uHo_omzcxm: mm * em ‘mnmloooarmxa.~m imafl iaoa.«.mm~ «mud manmozczv UHFuucyxomo}:-~-4>zmza mm . was ‘mwwiaoad.~mm.mxw “mum.omm .mom .2; A .wa .u_zom mmma . fl «Inc x“: mam . Nam ~m~.ooea.xv~.mmv .maw.mmx .mm« .03 ..n .mvw . "mm“ szoho¢4-¢ «-ozomazhv mm: «mm . __...-.~—- ' coed.mflmwim .de .aoord‘mdm amma Auncnmmcv u~c~omzCzoz vua s mvw amNN.Q0v .mmfi.ooofl.mm isoa.d.wmm mwma mm: QNM i mmm .wnm.mmm ‘vwm.fivm .mvm_QQOH.va .Osa.a.vnm mmma mm: «mm s mam :aooa m.MiooNH. Mom .m’ '3 u~mh~u mm“ 4 emm imwm.oao«.mmfi.mam ixda.omm .mo« .ama.«.m~fl. weed mm: omm * new .mflm.aoafi.mx~.maa .mmN .aeo.a.xxm mmm« mm: mmm . NN .mwm.«mfi .mmm.ooma.mv~ M .93 A .mem on «man mm: mam * mm“ rm. m.mmv \mx~.m«m \xmm‘mMm .«mm.mm~ .omm .eao_«.mxm mvmfl cm: mmm 4 maoa.mxm.omm rmmN.mmm .mm .gao.fl.m~m New“ Aumkmzoocnmnov u~4>xom¢mowmhmzmmomml M.N.a| mac mmfi é xmmv mama mvm mvm .mmm .oam « WVM vafl mN: mam * Qfiv tmvvxama .nwm ~Nmm .mhM‘eNM«.mmm.ofiu .mmm.amaaxmmm.mm ‘NNN .oso.d.mmm vmwd Unmoxmmozmommu3401¢ mm“ * mam .voM‘mMV ‘NQN.QQad.mma .omo.d.wnd ommd mm: mum § aw im~1.mw .mvm.¢wm MMM.000« Nam wvm .mum .oaa a.Nmm mm: mum § 3.308 Am $5on @®&«.ameHaM «www.m«w .de .aoo.d.omm vomd vm: mmn * 9mm \mve.omm .mmm‘omm .mfim.ooe« immm.av¢«.m¢m‘mma ‘NNN .@©Q.a.mmm mmma QN¢OImmozmomuo}Julm vwa * mm .mhmxvma .Nom.mh ‘vm«.omo«_am« xomel«.~mm man“ 0 UHmNQIObmzlcv wz~xo uHouowzcxwzoXoIN mva * moomxmmm.mmm .nma .wom.d.mam wmmd Aummmmamv U~o~awz¢hoo «Va 5 mm meM_om ‘m_N mama amm.Nd« .vaioma .mmfl.vNM .mNH .ooo.a‘omm Emma mu_m~oc02~:clcv unomomzcwaOzHECIN mma * oomH.MMMxaavnxmmmxmoofl.nnw i00@,«.mhw mmmd u—zommmm amu * mum fwmm.omoa.mvmnmmm .mmN «moo.a.vmm Hmhd szFUCJQZOmHm mma * woma.mmmimvm .ndN.mMN .mm« ‘®00.H NMN mmba mm: mum * mmmfi.mmuxmmm .dmm.hmm .NmN .Do®.d.mmm wmmd UHFNQIJ)ZUI&>XDKQ)IIL mMH & wMN .mmm.mfim .MMM‘QGGH‘NmN .O®&.d‘NwN NVNH vND mam i 6mm .wmm.&mm .Hmm.mvv ‘NWN.OQOd.va .QQQ.H‘VWd mNNd Uuhm0¢4>zmlm)xomo>llz mud I mmm .mmwxowad‘nmwxmom .MNN.M®m ‘MMH.fimH .wNH .mao.«.nwm u~Oszm> VNNM - n xomo>zlm mNd Q x. x l . \ .. a wm PM¢ VFW MMr.hhM ham mmw .NmN~N?a .WNN.mVN .WNH .25 a .23 mmma mm: mum c 285 égqoov .5 EDGE ham .mvv.mmoa.mom.amo«.wmm mmn .vnm.«nm .mmuxmwm .mmq.ao&a.mwa .ooa.«.mom .an.«_mWH meow amma d xcmmlo~zoanh0¢4¢wla ham s Au~mpmmm>v an .mmm.mv .HmM.o@®«.mmw u~owzmm>xorhwtmoav.m wma ; DOGH.MQM.M«N .fidN .0QQ.H‘mMN .QQO.H.H#N VQQN mama AU~4m02¢24~z¢>v U~FMUEIA4>zwxm>xoxhu2|M|>xomo}:|7wl>xamo)zlw mHN 6 mzohumg u—mhwuoma «mu n no" ~nmm.mmv .Ham‘cwaa.vmw aha .mmm.mm .mmm.mv .«eM~on« towm.e@md.mmm tmaa.«i¢hu .oom.uinwm omm« ammn Nu: vwm w Au~4>mucmo>xmzovzp uhozcmommunJyzwxm>xomo>IIMv|MI}xomD>IIM amu a mooH.uMM.mvm ‘mam Gamd.vwm.mmm .mmm.¢mN .avm .eam.a.HMM .ooo.n.vmm vmma mmm« nonummmmv o~c~amz¢uuo mow l muHm~hzmcozoIv unhmoc4}zmxm>XD¢o>:Holw~N mmfi i an .nwm.amu .mnm.amv .QaM.omH .MmN.mm .HwNiocoHrmmH oom«.7mm.vm .mmm.0m .ovM.H«H .99“ .0m0.a.mmfl .Oo®.«.va Hde mmmH nu~hu¢44>zmra>xomo>zumv o—thmomman4>2m1m>xomo>xavv0MI>xom011IN mom é mm“ .nmmtoema‘mmm o~hwu¢4>zmrm>XQmo>IHo|1‘M Mwa i mwv .wwvxmnm .«mm.o¢@d.WSM .Qoa.d‘mmu .ooo.u.mmv Nmmu vmma an: mwM § m20humglv.dlonPQ¢4co mmfi i ONN .nmv.omm immv.amm .mmm coca.wmm.mvm ‘dfimngH .mmN .oem.u.mmv .ooc.u.wmm mmm« MNmfi quoQ=40 mom & AUNJJHZcfiotoIV MN“ ‘mnm.Mm .wTM.mmv .mom.oa®«.umw and ~mmvrmw« .mmv‘nnm .oam H.Nmm mnmu N xcmmlu~20m=huc4cova mad a @de.mom.wwm .«m« oMN .ooo.«xmom nmma n0~¢¢2200|ov u~2¢22~u>xozo}xlc VMH i mmm .Mnm.ovv .momxoomfl‘NmN.vmm .mmN mNm .omo.«.Nmm mmmd Nm: VNM o mm xamm.gcfi KHMM.OQQH.MuM.Q@« .mNN eaafi.mamiomm .mmm.aom .aoo.u.mum Nmmu mwzlxz: mm“ * ‘QMM.moofi.moM .mao.«.mwv ovm« N xcmmumzo»uz4020hucq¢o mmw § Ao~44_2c>v oanfi.mwv.amm woa«.mmmimh HuaqcxbxmwMuhu on“ .mmmioomdimvm‘mmm .nmw amm ~009.«‘mvm mmma on: NNM 5 Agnmhmuomnv murmuIJyzmxm>xoxhw:IMI> xcmo>zlv vma i HMNv.ooo«.MMM .@0&.«‘MMN «Hmd mN: «NM s .mhm.omafl.hmm «Goo.a.nmw UHomzmmrxo:»mznm:>xomo>xlo and.§ mama om: Nmm g .mVM‘moaH.«m« .aaa.«.«md mom« mm: «mm ! .mmaiavm ‘wmd .moaimxwum mama .«mw.mm .ova .ooe.u.me Nmmd u—qrxommmu_am2mNzumrv.H mud n .mmv.&ona.mnm .Qon.a.mnm «amu u_._}xommcuEEonma-M.~.T>xo~ayz-« m2 .. 286 om .vmm.am .mmv.omo«‘nvm.omm.uwmw .DQ®.H.NVN mVMN 0v: NMM § maaanmmM‘NNV .MNMxfinv .mQM.M«N .mmN.mAH .mNN .SQQ.A.mMM mmNN Au—szwuu N xcmmlum:czznorxoxhw:|M|>xomo>Ilv mmm s @aa«.wmv.maw .«vv‘mm .nmv.mm« ‘Nmm.nw inmm.fim .MmM .oas,«.mnv Aunmnv mzau3m>xomo>z_mhlm.m.N «mm e mmw .oqm.mowd.mmfi.mvm .«m«.NNh .aNH .coana.mm« OVNN Amazcuomav uHozumommla4>JON¢Q~E~IVvIM va l aooaxwmm.vmw ‘HMM.mm anom .0QO.H.me noamuucuv N xcmalomzczzhurxomo>xnone.M «mm * mam .nvm.wn .NmmxooQ«.QM« .Q@&.«.QMH VNNN mwzlxzn mvm é omm ~awném .vam.oo@fi.NON .9QQ.«‘NQN mmam u—wwom4}gooz~lm vvm Q no“ .MNM.DHM .mNM.omo«.NNM .ama.arnmm mafia mm: MNM § moou‘vmmndmm .mmd .0@m.a.?mN omaN um: mmM & mam .m«V.va .momxvmm .nmw.ooma‘mmm .ooo.«.m«v wNHN Ahuhouou2Cumo>xomo>Iv mm: mmm & 00“ xwmv.«m .mav.eood.nmw .oo0.a.mmv seam mm: mNM § ome«.mmM‘wNN .HwM.eMfi ‘nom .Dogifi.me mean Aummuucuv « xcwmlo_:czzHu>xomoyrnoIv.M WMN & «ma .wMN.ooo«.woN xoaa.d~wom Nsdw Au—mammHIV umhmuxcoatcmzwm NMN § @388 .E @9on oaaa.wmmioen AMmméom .mmm .oaeia.mmm wmmm Aunqzmwmv " xcmmnu~zczzuu>xo:»m:IMI>xomo>:|v mmm s Nwm .mmv.eo@«.vm¢x~m« .wmv xoamla.vmv mwaN mm: mmm u mNa .mNM‘Q@¢H.MfiM.@MM .mv« .QQQ.H.MHM MmON AUHH—zqcmv umozmuwomxmx Nan é man .mmm.mmn.mum.amm .Mmm‘ammd.umm ‘00&.H.wmm ‘ wmon. «NON x2: ”mm * vwd .mmm.ao®«.mm~ .Qea.d.mmm wmmw AUH>Da>m4>zmmeolmv NEHXO u_ozcmommln4>zw1m>xomo>I|¢vIMon0IN mNN ¢ mm .QHM.@ooa.mm« .uoolH.G«M ovum Au~mmmmuomo>xv uhozmmDMmla4>2mxm>xomarzmolv.lem MNN s mnw .mvvxeaonxmomtmmm .Nmm \QQQ,H‘NMN MMQN N Mcwmlo~zomnhum4cclz mNN & «mm «nomxmoca.nflmxdmm .vammeN .HMfi .oao.«\NHN NMON Au~ZDMDPQEJ¢DV «MD mNM & ouou.nom.mmm ~Nmm.wmm xfima .Qoo,«.amfi mmam Mn: mmm * wHM .moMxNDM .waxooaa.hfim.0&¢ .VQN ‘000.«.NHN NHQN AU—zomDFQCmev MM: WNM ! mVH xMNQ‘UQdeMMM .ooa.simmm adQN osomimvv.oaefi.vxmiaam .Hmm.aom .mvm.wwmomwmm* .oge.s.,xm moaw usmmouma mam ; 287 6388 .8 amps; NM ‘NmM‘mN .nnm ~009.w.wn «wvw nu~cnxnxmv omozmxw104>u>xomo>1~mhlm.v.M mum § can .N0¢xoh .Nmm.mm .vwm.oomaxommxmv .NQN u—hmocva4>40az~lmc>xomo>xlmv mum l moo«.onmxow« .dmu.om« .NNH .0Q@.«iodm movw N xcmmlontczz~o>xothtHakim.¢.M vmm * ve« .nhm.vm« ‘mam.owm .fimm.oooa.mmd .maa.«.mmd Nmmm NV: VMM i .mMM.oNN inM.oooaava\«n KVMmimmm .Mmd.vn .de .oaa.«.vm~ «NMN Au~mDLL—I>xomo>zltv Mme QCmocooEcnzmmiogiu: wwm _. 626: .mam.aa3.x«~‘mmm .93 .03 .« .Zm nmmu .3: mmm . APPENDIX C Compounds excluded from statistical calculations The compounds listed in Figure Cl were excluded when using the t-test to measure the significance of differences in mean concentrations between subject groups. These are compounds whose concentration values are suspected to be unreliable, based on internal evidence in the MSSMET data. 288 289 Figure C1. Compounds excluded from statistical calculations.* Number 33 34 35 36 65 66 6'7 '79 82 96 138 145 181 187 209 229 236 254 2'74 Name }( -hydroxybutyric unknown RA 183 methylmalonic malonic maleic phenylacetic nicotinic 3, 3-dimethy1g1utaric citramalic mandelic o( —aminoadipic ox -ketoadipic oxime isocitric lactone homogentisic sebacic p-hydroxyphenylpyruvic oxime caffeic-peak 1 urocanic 3,4,5-trimethoxycinnamic-peak 2 Reason** . x x x u C x 290 Figure C1. (Cont’d.) Number 277 297 299 320 332 344 345 348 354 356 358 363 368 373 377 Name shikimic unknown U5 (cresol) unknown U7 unknown U28 unknown U40 unknown U52 unknown U53 unknown U56 (deoxythreonic) unknown U62 unknown U64 unknown U66 unknown U71 unknown U74 unknown U76 (hydroxydecanedioic) unknown U81 unknown U85 Reason** 291 Figure C1. (Cont’d.) ** Compounds listed were excluded from consideration when comparing subject groups using the t-test. Compounds were selected by manual examination of MSSMET and statistical test data. 1 Found in less than 25% of urine samples. 2 Mean match coefficient too low. 3 Mean retention index too far from 1ibrary retention index. 4 Mean area only slightly above detection limit. APPENDDI D Summary of urine samples analyzed The data listed in Table D1 were collected from the questionnaires (Appendix A) completed by the BCIU subjects and from the attending physician (Dr. Krivit) for the juvenile and neuroblastoma subjects. 292 3 9 2 mamwofimm SEBomamummm Ummdmomu :08 N a. .... 269:6: ...; H B .... mmnwuwgn cmmqflmn .68 m B .... 03 59.3 5 300 693 mmxofim motoamwflo moxofim “5.3mm vmfismzoo mafia 85$. omumuwam mquofimm wfiZoomm 358800 m 9520 mm; ”we; cm; 34 mim mm; may; ové wmé wad FEKw Ev 65:3.»an omufiw 2 ca. N o«-Hm a} e oonfim 222mm ome In omuHN omuav 2 .m oonam Amumob macaw mmm Now go owfl omo 0m 22 ONH US $0 one 53 who a E233 bZomcbo rZobomo HZobomo oZobovo mZovoao mZmomoo $8 ammo gamete HZomoEV $205.30 .I .8955 395mm HonoH momNoH Nomdmo Hoaamo Nomamo Nomubo mommvo Hoomuo oomabo mommbo .3955 55 92:00 53.328 3388 $5.8 Ho fiafifism .5 033. 294 .osomflnoo “38833 3:04 5283.5 3888 £50 mmgsummm 602935qu E 5:820 .5383ch “£93m .mflzmmfiomumo «Eon—mmfiosflmm coanflm 3.5-950 68.80% :08 m B ...n 2855: ...; H .H. ... 3:68:80 h. u .m h. u 2 .m I .m h .. .m h. .. 2 m om; h m oo.m 2 z 3.0 2 z 35 h 38 mmfiv @595 655380 xom 31.3 m: 31.3 coo N} H E. Amnmomuw anon m m m .8 m < Ewflwm HmwoM Nomomo HfivovH HONOmo H002 woommo ammovm womamo HmvoM mooamo NZQbObO HOOHHH waommo woflmofl wabomo monoH mZobomo NOHMOH gang's .8985: 395mm a?“ 92:00 3.208 .5 033. m 3.3938 “oz I 3.335 mac Ho 5934 .H. 2.5 USN 5x85 295 .353anwa 3.3 mwsan .coHuomHHoo 33.3 8 329 3333 m. 3333.“ mwau mo omoc 3mg ... mom h Hohzoo 3333:. H. «8833832 2 Hum mH mHmth .m 332 2 ”3M :33; .3355 .nomommh £33“:di .m.< 9:83wa 333% m5: .3>HH £30.53 0395 H. . h H5 mm. HmmovH mONOmo :Homfiofisom £5333 3033382339»: .8538ch vowmaotmm H. .. 2 H H>HQ HovovH momomo 333%. HHS meN l. son 3 a m 398:5 398:: 338800 a mfiaHu—wmno snow .3 3% 23me 298mm :5 92:00 23393 HO 334.8 APPENDIX E Relative areas calculated by MSSMET Twenty—one urine samples were analyzed by MSSMET using BESTLIB. The relative areas reported by MSSMET are included as Table E 1. The compound number appears on the left-hand side of the table (refer to Appendix B for the names of the compounds). The run number of each sample is printed across the top of the table; the first six digits refer to the specific sample (see Appendix D), while the seventh digit denotes the type of subject (2 for BCIU, 3 for neuro- blastoma, 4 for juvenile control, 6 for Foy urine and 7 for replicate GC-MS analyses of the same sample). Note that one urine (031603) is entered twice in the table, once in class 4 and once in class '7. All values are designate ion areas relative to the area of the designate ion of the tropic acid internal standard, and have been normalized to include creatinine concentration for the BCIU and neuroblastoma groups and urine volume for the juvenile control and Foy urines. The table was produced from MSSMET outputs using the programs RAMAST and RASE CD. 2 96 297 Nm mNmH Na mama. mm mmnw. «m mwaM. «a mva «a mmad. am mmaa. aaadaadaaaaaaadaaaaaaéaaaaaaamadaaaaaaaaaaaaaaaaaa MNamfima Na mnMN. H ‘9 m v1 01 N ('1 '9 In N M 0. H a a a m m M 0 fl a ‘ fl a ' ' asedddasddagdeededadasaoaaedddeeaeosaeadesddaoeado am mdmw Nfimadfld r1 '3 U ‘7‘: m H O a m 9 D . . . a a aaaaaaaaaaaaaaaaaaadaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa Nmanmad aa mmNH am mmNn am mama aa mama am maaa «a mmww am mmaa Nm mmmm am mama Na mmva dm mmam fia mmav N«a«mafi aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaadaaaaaaa mm mva. «a m«wv‘ «a mNaa. «a mama. Ha mama ma mmmv. ma mama. fia maNm. «a mafia. Ha mmmv u a m a m aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa mammama «a mavn. am mmam am mmam. «m mama. «a mmnd. aa mnmm aa mmmw no mMMm‘ am mum? «m mvmv. Hm mmvN. am mmam. «m mama, mm mmma. mm mmlm. mm mmam. am mMM «a mmma. «a mmvu. «m mwman Hm maqm. «a mNMH. ma mmam. fla mMNH. ma maam. um mmnfl. am mama am mfimfl ma mmtm am mmma a0 maaa am mmma am mama «a mmmm Ha mmNH «a mwaa mm mmvm am mmaa mm mama am mama am mOma am mama aa mama am mmnN aa mmma Na mmmm ma mmam «a mmmm «a mmmm fia mamm Nmmmmma axncdatécia a'a a a’a a aWfi a'a a a'm m akfitncncratnuaa ao<§¢3otfic$o<§c$o<965o<§qio<§ciocéc>oc§¢$océcioc§c$o<§c5ocn o I o o mmmm vdomomo «o moan oo mooo “o mvod oo mooo No mmmm No mNOd do mmwm o0 mooo oo mooo No mum“ 0o mooo fio mmvm do mmvm No mvmd No mmod Ho mmwm 00 mooo oo mood. 0o mooo. oo mooo No mmod. No mafia. oo movm oo mooo. do mmmm. “o mvan. No mmva. do momv. «olmmmm. «o moMN. oo mmmw. mo mama. no mood. No mnmm oo mooo No mvom oo mooo No mmnm do moHM No mmma No mmva No mvna Ho mova No moov. mo mmmfi. .o do mmwn. oo mooo. No mmMfi. do mmdm. oo mooo. o¢9¢$oc§cio<§a$oc§¢$otpc>oqpcaoqnc>oc9¢ao<96>o<§c$o<§¢>o<§s$o¢§c$ocbciotbcaocsdao vvomamo mum vmm Mum .xl ’P «NM mmM mom nwm mom mwm *wm mmM «mm own an me "F «o moom.o fio mwoM o oo mooo.o Ho mmmm.o do mmom.o mo mooo.o oo mwmm.o oo mom .o «o mmoa.o No mmna.o do mow~.o oo momm.o fio moma.o mo mMfia.o o0 mooo.o mo mooo.o «o mama.o oo mmhv.o Ho moMM.o No mmma.o do vam.o w m7ma.o oo mooo.o Ho mva.o oo mooo.o fio mmMN.o fio mnmd.o mo mooo.o oo mmma.o do mmfim.o do mMmm.o w moaa.o do mmam.o do mmmd.o fio mmmm.o oo mooo o No mnmm.o N mmvm.o No m«m«.o oo mooo.o oo mooo.o oo mooo.o do mwmm.o Ho mmmm.o Ho mmam.o Go mamm.o do mdm«.® Ho mmmw.o No mMaN.o No mmmm.o vmomdmo No mmmm oo mooo «o mmmv Ho mmmM oo mooo oo mooo No mmoa oo mmvm om mooo «o mmmm No mmam No mNmm fio momm «o mhfim No mmmu oo mooo No mmfim oo mooo do mmmm oo mooo ao mmmfi. No mafia. fio mmmm do mmmm aa mmmm. «o moan. oa mooo mo mama. oo mooa ao mmm oo mooo oo mooo oa mooa mo mmom. oo mooo oa m¢mm mo moan mo mwvm oa moao oo mooo o mmma. «o mmdm mama mmow mama o mom“ mohm. o mhma. mama o mamm. mmoflmoa «océotbciotscao¢§c$o¢§¢$o1§c$oe§e$o¢§c5o<§c$o<§cio<§¢$ocn¢$ocaesotficaotseiocéchicao No mwam. do mmmm. oo mooo No mvoN oo mooo do mhmm oo mooo do mom“ Go mooo. mo mamo. No mmom. oo mooo oo mooo 9 mhma. a0 momm. oo mooo No mmmu oo mmmm .o mmmm No mnmm No momv No mMmM oo momo. go mum oo mooo. No mnwm. No mfiow. fio mmmfi. «o mmam. fio mmvm. Go mooo. mo mnov No mvmw «o mHNH No m¢mv oo momo o mamm. mo mmmm. mmmm oo maav. No mmam do mmam No mamm- No mnVH. .o movm no mwmh. No mmvm mom? mo mmom. mMNm. mmaumoa otnaioencéotscichocéocsciotficiotncao¢scaocna$oc5o1§¢$otficiocsGiotocaotficéocficbsaoo do mvma do mMVfi oo mooo. do mNmM. mo mooo oo mmmw oo mooo doummmm. om mmma ND mmdfl fio mmmm oo mmoN No mmmm oo mooo so mooo oo mooo. «o mmmm oo mmhd do mnMN do mmom do mmmm GD mooo. oo mmmv. oo mmvm 9o mooo do mvnm. do mmoa aonmmmm doommmm «o mM«u o9 mooo «o momm. oo mmom oo mvvm. fio mMVT oo mooo mo mmmv. No moat No mmmfi. «olmovm 9o mNmn. oo mooo. fio mMov. «o mmda. do mmmd. oo mmae. fio mmom «o mmNN. No mnma- No mmwm o o oc9c>ocbcaocsc:ocposocsc$oo n MMomuoa apéoov Ho mum o mum oo mwmm o vnM oo mooo.o mum do mmmd.o m n 00 mooo.o um oo mfiNm.o mom fiolmom¢.o mom woummmd.o mom oo moa~.o mom Ho mamm.o mmM do th«.o vmm oo ma¢m.o n M No mNma.o amm No mNMfi.o omm oo mva.O m M oo momm.o mmM «o mmmu.o mmm oo mmnm.o mmm Ho mnwm.o vmm do mwmv.o M m «o mmhv.o NWM fio mmdm.o «mm oo mMmN.o omm oo mvmm.o mvn dotmmmm.o wvm oo mmvo.o mom oo m«N¢.o wvm oo mmm«.o mvm «olmmmm.o vvm do mmoa.o Mvm do modm.o Nvm do moun.o HM oo mafim.o MMM «olmvam.o mom fio mMmM.o mum oo mooo.o mum No mmmN.o mNM No mfimd.o mum a9 mmfim.o VNM dolmd«M.o NNM oo mMNm.o ANM oo mooo.o oNM fio mmov.o mam oo mn«M.o mam 0o mmmm.o mam 9o muom.o vHM do mmmw.o Mam ao mmmN.o oHM do mmhm o mom No mvmm.o mom MNomHmo .HH Human. 3()4 aa mavm.a aa maam_a mum aa mvam.a aa mmam.a vam aa maaa.a aa maaa.a mum aa maaa a aa mmam.a mum aa maam.a aa maaa.a an” aa maaa.a aa maaa.a amm aa mawm.a aa mama.a ama aa mamm.a aa mmnm.a mam Na mmva.a ma mmaa.a ma mamm.a aa mmam.a aa mama.a mam aa mama a aa mmaa.a me aa maaa.a aa maaa a ma mmaa.a aa mama a aa maam.a mam ma mmaa.a ma mnma.a amp aa mama a aa mmam.a «a maam.a aa mamm.a aa mama.a amu aa mpam a aa mama a van aa meam.a aa manv.a aa maaa.a aa maaa.a aa mmam.a mum aa naa.a aa mama a mam aa maMM.a Na muav.a ma maam.a aa mmem a aa mmam.a aha ma maaa a aa mama a ama aa mama.a aa maaa.a Na mmam.a aa maaa.a aa mamm.a mam Na mama a a mmaa a am” aa mamm.a aa mnma.a aa maaa.a aa mama.a aa mvmn.a mam aa maaa a aa maaa.a mam vmamama mmaamaa mmaamaa mmaamaa maamaaa aa maaa.a aa maaa.a amm aa mama a aa mama a may aa maam.a aa mana.a mmm aa maam.a aa m mm.a vmm ma maam.a Na mama.a amu va mava.a ma mmua.a «a mamm.a Na mmam a ma mmm¢.a Mam aa maaa.a aa mamm.a mam Na mmaa.a ma mavM.a aa m~a~.a aa maam.a «a mama.a mam ma mana.a ma mama.a awn aa maaa.a Na mmaa.a ma maaa.a aa mmam.a aa maam.a aaa aa maaa.a aa maaa.a awn aa mmma.a ma maaa.a aa maaa.a aa maaa.a aa mamm.a mum aa maaa.a aa mmam.a men m maav.a ma mm¢1.a ma mvaa.a aa mamm.a ma mamm.a mam aa maaa.a aa maaa a mam aa maav.a aa mmvm.a aa maaa a aa maaa.a aa maam.a ham aa mamm.a aa mmam.a mam oo mooo .o oo mooo .o oo mooa .o ao mmmm ,_o «o mmom .o mum an. mama .o 2. mmma .o mom mvoommo emomono vmomomo vaomowo vvowamo oo mooo.o oo mooo.o mvm aa mama.a aa mama.a vam aa mama a aa maam.a m m aa mama.a aa mama.a a m a mama.a ma maaa.a vmm ma mama.a aa mama.a Ma maam.a ma maaa.a ma mamm.a mam aa mmaa.a aa mmaa.a Mam ma maaa.a aa mavm.a aa maaa.a aa mamm.a aa maaa.a Nam aa maaa.a aa mama.a mam aa mmvv.a aa maaa.a aa mama.a aa maam.a aa mmwa.a aaa aa mamm.a aa mamm.a mam aa maaa.a aa maaa.a aa maam.a aa mamw.a aa muaa.a ma aa maaa.a aa maaa.a mam aa mmaa a aa maaa.a ma mamm.a ma mma~.a aa maaa.a mam a. mmnm.a Na mamm.a mam aa maaa a aa maaa.a aa maaa.a aa mmma.a aa maaa.a mam ma mwm¢.a ma mavm.a mam aa mamm.a aa mma~.a aa mama.a aa mavm.a aa mavm.a mum ma mama.a ma maaa.a cam maaaama maaaaaa mmaamaa aaaamaa «aaaama aa maaa.a aa maaa.a «an . aa maaa.a aa maaa.a amm aa maaa.a aa maaa.a amm aa maam.a aa mvam.a «am no maa~.a Na m~vm.a Mam ma mamm.a ma mamm.a ma mmam.a ~a mmam.a Na mavm.a Mam aa maav.a aa manm.a aam H9 wnmd .9 fl& wmmfi .0 Nwm No mwmn .9 fio men .9 No mdafi .0 no whom .9 fiw wmwm .0 NmM do MQ1N .0 fio undN .Q mfiM aa mvmv.a aa mmmv.a aam aa mmNa.a aa mmaa.a aa mavn.a aa mm-.a aa maaa.a aam aa mama.a aa mumm.a «am aa maaa a aa mamm.a mum aa mma~.a aa maaa.a aa mmmm.a aa mama.a aa maaa.a mam aa mama.a aa maaa.a mam aa manv.a aa mmam.a mam ma mnnm.a ma maaa.a ma mmma_a «a mama.a Na mama.a alm aa mamm.a aa mmav.a aaa 00 mooo .o #0 mMMd .0 mm 0o mooo .9 Do mooo .0 90 moao .o On. moon. .0 00 mama .9 hum mm. mHMN .o No mMHN .0 mom aa mmma.a aa mama.a mam aa mamm.a aa maam.a aa mmm~.a aa mmam.a aa mama.a mam Na maam.a ma mamm a man naamaaa umamama amaamma aaaamm maamam amammna «mammma muamama “mamaaa A388 .3 mam; APPENDIX F Normalized relative areas Table F1 shows the same data as Table E1, except that the values have been normalized to a partial sum of peak areas. This partial sum was calculated for each urine by summing all of the relative areas listed in Appendix E except for those of the compounds listed in Appendix G. The resulting sum was then multiplied by a factor of 0.001 (to achieve a value between 0.1 and 3), resulting in a “corrected sum.” Each relative area for the urine was then divided by the corrected sum for that urine; the resulting values are referred to as normalized relative peak areas. 305 3()6 ma mmmN‘ Na mama. aa mNam, aa wava. aa mamN. a0 maaa‘ aa mmmm. aa mamM. aa man‘ Na maaa‘ Na mmaN. aa mnmn. aa mmma. Na mama, aa maNm. aa mva. Ma mMNM. Ha mama. Na mmma, Na mvma aa mmmm aa mnmm Na mama aalmaaa aa mmMa aa mvmn aa mWNN aa mnvv aa mama aa mVMa aa maaa aa mmmm Na mama aa mmmv aa maaa Ma mmMN aa mema aa mmam aa mmva aa mwmv aa maaa aa maaa aa mmaa aa mmmm aa mmwm Na mmaN aa mavN Na mnaa aa mmmm aa mvah MNaaama ciacéciaia a s-9 6 615 cisia dcficiatncSQ'atbt§t§cia'cia‘a'arpaacva arena 0 OIDGDG>Q(D¢)E)O a «Q mavn.a aa mamm a Na mmma a No mMNH.a Na mmMN a N mmaN a aa mava 0 aa mat? a aa mmva a aa mava a aa maau a aa mmam 9 aa maaa,a Na mNau 0 aa mmww 0 aa maaa a as maaa a aa maaa a ma mWaa.a Na mmwN a Na mama a «a mha~.a ca maaa a Ga maaa a aa mham.a aa mnmv a aa mamm.a aa muvv.a aa menu a Na man a so maaa.a aa mfiva a Na mvna a a9 mmm~.a aa maaa a aa mvaM a aa maaa a an maaa 0 aa maaa a Na mama a Na mama a Na moan 0 aa mvma,a aa maau 0 aa mwav 0 aa mama.a a0 mama a 06 mmam a Na mamN‘a aa maaa a Ha mmaa a aa ma~N.a aa moaa a 90 maaa 0 aa maNN a aa mmmM a aa mVNW a aa maaa.a aa maaa a Na mnaa a aa maaa.a as mama a aa mmmv a aa mama a as mama 6 aa m1m~ a Na mmhm a aa maaa a Na maaa a aa maaa.a aa maaa 0 aa maaa a aa maMN.a aa mmmv 0 aa mem a aa maaa a so maaa a ma maam a aa maaa.a as maaa a as mmnw a aa maMN a aa mnmm a aa maaa a aa mwhm.a aa maaa a aa mmvm a aa maMm.a as maaa 0 aa mmam a aa maaa.a aa maaa a aa maaa a as maam.a aa mean a aa maam a Na mama a «a mNNm a Na mmam 9 aa mVaN a aa mnvm 0 aa mmMN a aa maaa.a aa maaa 0 aa mama a Na maMm a Ma mama a Ma mama a aa maaa.a aa maaa a aa maaa a 09 maaa.a aa maaa a aa mmaw a aa maaa.a aa mavv a aa mmaa a aa maaa.a aa maaa a ma maaa a aa maaa.a aa mmav a aa mmma a aa maaa.a aa maaa 0 aa maaa a aa mwvm.a aa maaa a aa maaa a «a mumm.a aa mmwv 0 aa maaw 9 aa maaa.a aa maaa a aa maaa a Ma mama.a Ma mvma a Ma mmaa a aa maaa.a aa maaa a aa maaa a Na maaN.a N mmmm a Na mmMM a Na mnan.a Na manN a Na mmaa 0 aa mavn.a we mea a aa mmam.a maaaflaa Nmaamaa Naaamaa 93.3 2533 3538qu Na mVNa o mma ma mNMN.a mma aa mmav,a vva aa mamm.a ava aa mmwa.a Ma Da maaa.a ama ma mmMa.a mma aa mnaM.a mma Na mMMa.a mNa Na mnaa a mma aa mmea.a mua Na mwva a vna aa maaa a Mma Na mvmm.a am aa maum.a aaa aa mmma.a vaa Ma maNa‘a Naa aa mmma a aaa aa vav.a aaa aa mamm.a haa aa mmav.a maa aa mmmN,a vaa aa maMN.a am aa mmmm a mm mm mNaa.a mm aa mmhm.a Na aalmnmm.a mm aa mNmm a nu aa mamn.a hm aa maav.a aa aa mama.a ma aa mva.a am No mvam.a aa aa maaM.a m aa mawv a an ma mamv.o am «a mama.a Nv aa mmuN.a am aa mmma,a mm aa mama.a vm aa maaa.a Mm aa maaa.a NM aa mava.a am aa mnaa a mm as mnvn a N Ma mmva.a aa aa maaa.a ma Na mama.a aa Na mamm.a a aa mvah a m Naamama .fimmdm<fi adddadoaaadadasddaadeadaadeaadaoadasaasagaaaaaedda aa maaa Na mMMN a0 mMaM aa mmmw Na mkva aa mnam Na mwaN aa mama Na mMNa aa m1¢w aa mML aa mmam ma mama Na mmam as maaa aa mama aa mMMM aa mnh aa mNnM Na mwma am mavm aa maaa aa mmam aa maaa aa mmav aa mamv aa maaN aa mm¢N aa maaa aa maaa aa maaa aa maaa aa mMNm aa maNM aa mmmm Na mava aD maaa aa maaa aa maaa aa maaa aa maaa aa maaa aa maaa aa mMan aa maaa Na mmam aa maaa Na mVaN Na mkmm ma mmNN. NNamNma aa maaa Na mvem aa mnmm aa mmam aa mmaa aa mmmv aa mama fia mama Na mmaN Na mnwm Na mvma uwawmno déaoadddddéssdadaadaadddsdsoasaasadaadgaaasasaadda aa maaa Na mvna aa mama «a mmma aa mawm aa mum Na mmNN aa maaa ao maam Na mama aa mama aa mama aa mama Ma mmma aa maaa aa mvma aa maaa aa man aa mmmv Na mmva aa mmMN aa mmam aa maaa as maaa aa maam aa mmaa aa mmmm am mama oa mmhv aa maaa aa maaa aa maaa aa mvmm «a muaa as mvma ma mama aa maaa aa maaa aa mamm aa maaa aa mmnm aa maaa aa mnma aa mMMm aa maaa aa mam aa maam Na mnva as mamn Na mama «Hoamma addsadddaasaasaadaseaassaagdcddddaadsadaaaodeaoaoa a0 maaa Na mvmm aa maaa aa mmaa Na mwaa aa mkma Na mwhm aa maaa aa mNaM aa mmmm aa maaa «a mmna aa maaa ma mama as mama aa vaa an maaa aa mama aa mama aa maaa aa mmmm aa maaa aa maaa aa mama aa mmam aa maaa aa mama aa maaa aa maaa aa maaa aa maaa aa maaa aa maaa aa mama aa mvnN Ma mNmN ea maaa aa maaa 0a maaa aa maaa aa maaa aa maaa aa maaa aa mamm aa maaa aa maaa aa maaa Na mNam Na mmma ma maam wwamwmo aaéadaaddddeadsedgadaaaaasaddaaaaaoaaaaaasaadgaase aa maaa‘ Na mmmm, Na mwaa. aa mmmw, Na mnva. aa maaa. ma mavv. aa maaa aa mama aa maaa aa maaa aa maaa aa maaa aa maaa Na mama ea maaa Na mama aa mama Na vam «a mmNm ~o mmmm. flmonwmo adsdéadasaaéossaaeddaaasasddasasaaddaasaoaageoooao 307 oo wooo.o mo mvo«.o oo oooo.o oo mooo o No u~m«.o mod oo uooo.o no m¢m~.o oo mooo.o mo unnH.o mo uaoa.o mna «o mono.o mo ummm.o mo mmoa.o mo mmom.o no uo~m.o Nod mo mmmm.o mo momn.o No moo~.o mo mmmo.o mo m~o¢.o «ma mo momu.o «o unom.o “o unno.o “o mooo.o «o umom.o ecu oo mno«.o «o mvoo.o «o ooom.o mo unoa.o do unoo.o we“ we ummm.o «o mwom.o oo mooo.o «o whom.o «o movv.o woo «o momv.o «o ooom.o “o mmva.o «o moom.o do uoo«.o do“ No wooa.o «o wvom.o «o momm.o oo mooo.o «o moma.o and «o mnoa.o do moon.o do moon.o «o momm.o «o monm.o oma oo mooo.o do menu.o oo mooo.o oo uooo.o «o uuun.o oma oo mooo.o oo uooo.o oo mooo.o oo mooo.o oo mooo.o oMa mo umoa.o mo uoam.o «o monn.o mo m~o«.o No moa«.o mma mo mom~.o No mmn~.o mo mafia.o mo umm«.o mo uooa.o omH «o unau.o oo mooo.o oo mooo.o ao u«m¢.o oo mooo.o nma do moov.o ao mmo«.o mo mvoo.o no momm.o «o u~m~.o mm“ «o mna~.o «o umo«.o «o mm~«.o «o mnom.o oo m«oo.o on“ «o mmm~.o do unvo.o «o mmam.o «o uoom.o Ho mooo.o mm «o mvom.o «o unm«.o «o m«o«.o «o unoo.o No momm.o “ma do uoom.o No unmm.o oo momm.o no mama o No uwom.o moo oo mooo.o No monm.o do mvo«.o «o unnn.o. ~o mm¢«.o nod «o mmvo.o mo momm.o ma mava.o mo moow.o «o moon.o mma oo mooo.o oo wooo.o «o umo¢.o mo unna.o «o mmo«.o cue as momm.o «o mmom.o mo mona.o «o moao.o mo uaoo.o «a oo mooo.o oo mooo.o «o wona.o oo uooo.o oo mooo.o mud oo muoo.o mo mmnfl.o oo mooo.o No maow.o mo m~a~.o "ma «o mwom.o No monv.o mo uo~«.o no mm«~.o No mono.o oma mo mooa.o mo mmov.o mo mmm«.o mo mumm.o no uomm.o ow“ «o mnm«.o «o uno~.o oo mooo.o ao womm.o «o wooa.o odd «o momm.o «o mama.o oo mooo.o «o mome.o oo mmoo.o oaa oo mno1.o oo mowm.o om whom.o «o mmv«.o oo mvoo.o cad «o mmaa.o «o u«~«.o oo mmov.o oo moon o oo uuom.o .«a mo mmoa.o no mono o Mo mmo«.o «o menu.o mo momm.o «do No uumn.o mo monm.o no uao¢.o mo mmvm.o mo muao.o mad «o mno«.o oo mooo.o oo mooo.o oo wooo.o oo mooo.o «ad oo mooo.o do mmvo.o do umo~.o «o mvom.o do wmnv.o «ad mo m«o«.o do umon.o «o woon.o ~o mao«.o «o moan o odd do mauo.o ~o moaa o «o mooo.o No mo~«.o go mwom.o oaa oo mooo.o ~o mon~.o mo m~e«.o «o uav«.o «o onuo.o No“ No mova.o mo ummm.o mo umma.o mo unma.o no mooa.o Nod do mmoa.o oo m~w~.o oo momm.o «o omom.o do uoom.o noa . «o manm.o am m~o~.o do mmm«.o do mov~.o “o mama.o no" «o uoma.o oo mov~.o oo uem~.o oo uvoo.o «o ma~«.o vow oo mooo o oo moom.o No uneu.o «o mnam.o oo mooo.o coo «o uomm.o “o mmmn.o “o moov.o oo mvofi.o oo mmow.o oo «o mamv.o “o mnmm.o «o movo o «o movm o Ho umov.o oo oo mooo.o oo mooo.o oo mooo.o oo mooo.o oo mom~.o mo oo mooo.o oo uooo.o oo mooo.o oo mooo.o oo uooo.o oo oo uooo.o mo moao.o «o mm«~.o «o mooo.o «o uvm~.o oo «o mnmo.o «o unno.o «o mo««.o no oooa.o «o mooo.o oo oo mooo.o oo mooo.o oo meoM.o oo wnoo o oo mooo.o mo oo unwo.o oo mooo o oo monm.o «o mood o oo wmoo.o oo oo unvn.o oo uooo.o oo mooo.o oo uooo.o oo m~n~.o on oo mooo.o oo mooo.o oo mooo.o oo mooo.o oo mooo.o on «o u«m«.o ~o woma.o ao mmvm.o «o uanv.o «o mov«.o on «o moom.o ao m-M.o «o momm.o «o momm.o ao mmom.o an oo mooo.o oo mana.o «o va~.o oo mooo.o «o unofi.o no oo umo«.o om wooo.o oo mooo.o oo mooo.o «o wqau.o no oo muon.o oo u«o«.o oo mmov.o oo mooo.o oo moon.o mo oo mooo.o oo mooo.o oo mom«.o oo mooo.o oo m~m~.o oo oo wooo.o oo mooo.o oo mo«~.o oo wooo.o oo mooo.o no oo wooo.o «o mooa.o oo momv.o oo momm.o oo unma.o no «0 mo~n.o mo uaou.o «o momo.o do mnov.o «o uouo.o «m «o ummm.o «o mono.o «o ovov.o “o ummo o «o wfioo.o «o No uooa.o «o mmvu.o no umnn.o mo momm.o «o unna.o om «o mvm«.o No mova.o mo umflm.o No mmma.o No mmam.o oo «o mov«.o «o “no".o oo uooo.o «o unnn.o «o mumm.o on do mnnm.o «o momm.o Ho mm~a.o «o munu.o «o mona.o on oo mmnm.o «o umoo o oo wooo.o ao umom o «o ummo.o an «o mmm~.o oo mooo.o «o u«o~.o do mooa.o «o m«oa.o an no mmnm.o mo un««.o mo ma«v.o Mo mmom.o mo mow~.o on mo uonm.o no uaoo.o no memo.o no mofim.o mo mmo~.o on do mono.o «o umo«.o no mo~«.o ao moom.o ao umov.o uv do oomo.o «o mmaa.o «o umoa.o do umvv.o oo umnm.o we oo wooo.o oo mooo.o oo mooo.o oo mooo.o oo uooo.o om oo mooo.o oo mooo.o oo uooo.o oo mooo.o oo uooo.o om oo mooo.o oo uaom.o oo mooo.o oo mooo.o oo wooo.o no oo mooo o oo mooo.o oo whom.o oo uooo.o oo uaon.o no oo mooo.o oo m~v~.o oo mooo.o oo mooo.o oo mwum.o on «o umm«.o oo mooo.o oo mamv.o oo uvom.o oo unoo.o M «o mvo~.o oo monm.o oo uooo.o oo mooo.o oo mooo.o Mm oo uooo.o o9 uooo.o oo mooo.o oo oooo.o oo mooo.o MM oo mooo.o oo mooo.o oo mooo.o oo mooo.o oo mooo.o om oo mooo.o oo mooo.o oo mooo.o oo uooo.o oo uooo.o no mo uooo.o oo mooo.o «o mm«m.o oo mooo.o oo mooo.o ow oo mooo.o oo mooo.o oo mooo.o oo unom.o oo wooo.o on «o movo.o «o mmum.o «o ummv.o “o mv-.o «o moom.o on No unou.o «o umvv.o do momm.o «o mnmv.o «o moam.o oo mo mom~.o .uo mmm~.o «o mmvm.o «o moon.o «o moom.o Mw mo u~o«.o No mmo«.o «o mooo.o «o umoo.o «o mono.o mu do mov~.o oo mooo.o oo mooo.o oo mooo.o No momo.o ou oo mooo.o «o mooa.o do mono.o «o u~o«.o ao mo-.o oa «o mwo«.o oo mnom.o «o mn«a.o «o uaom o oo wooo.o “a do mmnu.o «o mmm~.o ao mma«.o no uom«.o «o mom«.o mo «o ”Mun.o «o uoan.o mo weo«.o No mmov.o no umam.o o« No momo.o mo um««.o No unoa.o «o mmnm.o «o wovn.o oa wo u«n«.o oo momm.o mo omm~.o No ono~.o mo umv~.o o no mmum.o mo mona.o No uvm«.o ~o mnma.o «o mmnw.o o «o movm.o «o muon.o ~o m+n«.o No moov.o «o uomm.o m «o mnmu.o «o mamm.o do mamm.o «o mvom.o do moon.o m mvoommo emouomo omomono odouomo vvooamo vmooomo mmoamou moo«Moa «Moomoa moooaoo égooov .5. SEE. 308 Ne mNaN Na mavm Na mmma aa mama a9 maam aa mmav Na mamn aa mmmm aa mmma aa maaa aa mmmv Ma mmuv aa maaa aa maaa aa mvwv Na mavm aa mhmm aa maNM aa mNNm aa mwma Na maaM aa mama Va mvaa aa mmvm aa mama aa maaa Na mman aa mth aa maaa Na mmmm Na mnma aa maaa aa mavw Na mMNN aa maaa Na maNM Na maav aa mmam aa maaa aa maaa aa mekm aa mufiv aa mmmm aa mmaa aa maaa aa man? «a mmvv aa maaM aa maaa Na mmma. NNaaNma oooooooooooooooooooooooooooooooooooooooooooooooooo aa maaa, NNamNma oooooooooooooooooooooooooooooooooooooooooooooooooo Na mmmN Na maaw Na mva aa maam aa maaa aa mnav aa mmaa aa mhmM aa mmma aa mmmM aa maaa ma mMmm aa mamm aa mmaa aa mama no vaa Ha mMMV aa maaa aa mmad aa man No mmam Na mmaa aa maaa aa mam aa mama aa maaa Na mwav Na mama aa mmmw Na mama ma mmmm aa maav aa maaa Na maMa Na mNNa Na mmmN Na mavm aa maMN aa mmNm aa maaa aa mmwv aa maaa Na mmaa Na muaa aa maaa aa maaa ca maaa aa mamm Na mNaa Na mmaa. Naamuha oooooooooooooooooooooooooooooooooooooooooooooooooo Q Q m 0 O 0 '4 a m 7‘. v4 . » H oooooooooooooooooooooooooooooooooooooooooooooooooo NwamNna Nmanmna oooooooooooooooooooooooooooooooooooooooooooooooooo a a m v P1 . m .. oooooooooooooooooooooooooooooooooooooooooooooooooo mwammma A388 aa maaa,a mma Na mmMm.a Una Na mmaa.a Vva do moov.o a.“ aa mmma a mma aa maaa.a ama Na mamw.a wma aa mmav.a mma «o mmmo o oma aa mmwa a nma aa mNVm.a wma aa mNnm.a vma aa maam a mca Ma mama.a awa aa maam a aaa aa mhaa.a vaa Na mam~.a Naa aa maaa.a aaa aa maww.a aaa Na mava.a maa aa mahN.a maa aa maaa.a vaa aa mamv.a mm aa maaa.a mm aa mmmw.a aa aa mmNm.a N oo uooo.o on aa mmam.a Ln aa mmaa.a mm aa maaa.a am am maaa‘a ma aa mMMM,a aw Na mvaa.a am Ho umnm o on aa mMNN.a am Ma man .a an aa mama.a NV aa maaa.a am aa maaa.a mm aa mmma.a VM aa maaa‘a M 90 maaa a m oo wooo.o on Na mnaa.a aw Na mNaa a mu aa maaa.a aa aa mmmN.a ma Na mmuv.a aa Na maaw a a No “no“.o m “mooflmo Arm mama? :-"——“»_'-=:='—"E%.__ w? ""' ’ ' 309 vmamama oooooooooooooooooooooooooooooooooooooooooooooooooo N D W D W V. a a a a m m m a H O m a . . densiaWicia'd<§c$a<§qiat§cia<§aia'a'auaciaududa'a1§e$a<§q$¢ia o‘seysso‘o o'ots‘seso'oto MMaamaa rd 9 W In a I) nwaaMaa oa<§c>a<§c$a<9¢$o .a .a N mama. Na maaa aa maaa Na mmnm aa mmam Ma mamm Ma mNaa aa maaa ma mva 0a maaa aa maaa. aa maao Ma mvam aa maaa Ga mMam aa mnma N mama Na who? ua mmmm acficiacécicHSGSo<§c$acnciatéciacficiacfisidcficiatficiacpciachcid:§cidcficiatnciacficao Na mmwa. Naaaaaa aa maaM. a0 mama. Na mvom. aa mmav. aa memm aa maaa aa maaa aa maaa Na‘mvnm aa moaa aa maaa Na mmma aa maaa aa maaa aa maaa. aa maaa aa mmmm aa mama Na mama. Ha mmmm aa mama aa mwam aa maaa aa maaa aa mama Na mmMM aa mama aa maaa aa moaa 0a maaa Na mNmM Na mNNa Na mmNM. Na mmaN Na maaa ma mmma Ma mama ma mnNM aa mmah Na mnam aa mama ma mamN ma mamm Na mama No mmmu Na wmmm aa mMNm Ma mmma. Ma muaM Na mmnm Nmaamaa 6<§c>audd«5656<§c$d<§c5d<§c50c§c$d<§c$d(fichficiacéciocsciatnciatfia$a<9¢$dcjc$ocncpa Na mwaa. aa maaa. Na mMNN «a mmmw aa mnmv aa maaa aa mawm aa maaa Na mvvv Na mmaN. aa mvma Na mmmm aa maaa aa mmma aa maaa aa maam Na mvvm a0 maaa N mNaN. aa mmmm aa maaa Na maaa aa mMaa aa maaa Na mmma Na mmmm ao mmvm 0a maaa 0a maaa aa maaa aa maaa aa mvmm aa mama aa mmmm aa mavm ma mmha Ma mama Na mvvm aa mmmm N mvma 0a maaa Na mmfim ma mamN Na mvvw ma mmaN Na mvna aa mvvv ma mvma Ma mvem Na maam. acficidciacfiaia(ficidcbcjdtficdokscdatficdacchd<§c$d<§aadt§c$a<§c$a(ficaatficaatfichficaa b Naaamaa fla mMNm. 0a maaa. Na mn¢a. aa mawv aa mvmm aa mmav aa mmma aa mmMm Na mNNm ma mmma aa maao Na mwaa Na mmmm Na mama aa mmmn aa mmaM N mmaM aa mmww. Na mmmw. N mvmv. Na mmmm Na mmma aa mmwm aa vam. as mmam. aa mamm. aa maaa 0a maaa, aa mmwm. Na mvma Na mva Na mmaN aa maaa ma mmmm. aa maaa ma mmmm Ma mmma Na mmvm Na mmNa aa mama aa maao aa maaa Ma mvam aa mum? aa mnaa aa mva Na mMNN ma mmma Ma mNNN. Na mmmM. acaciaadatficiaacncaatnaaocscaacncaaao<9¢>acncidtncidtficid<§c$6<§c56<§u$dt§cid«scidcficadcncid acficiacéuaa NNaaNma ma mmma Ma mama Na mnvm N mama. aa mmmm aa maaa. ma mmmm. Ma mvav. Na mmaN. Na mavw. Na mmwN. aa mNMa. Ma mava. ma mNmN. Na maav. v ac9caa<9¢>atncaotna>acfiaidcficiacficiandciar§ciat§c$a<§c$a<§aiacficiatsciacficid<6q$a<§ NNamNma a0 maam.a aa maaa.a aa mmmM.a VNM aa maaa.a Qa maaa.a aa maaa.a MNM N0 mvmm.a Na mamm.a aa maaa.a mum aa mmvm.a Na mNMN.a Na mmwa.a ahM aa mth.a aa mmwv.a aa maa@.a mam aa maaa.a aa maaa.a aa maaa.a mmM a0 maaa.a aa muaw.a aa maaM.a an aa maaa.a aa maaa.a aa maaa.a mmM Na mvmm.a 0a maaa.a Na mmnv.a mmM Na mmma.a Na mmma.a Na mmnm.a vmm aa mmma.a aa maaa.a aa maoa.a mam Na man .a ma maaa.a aa maaa.a amm aa maaa.a Na mnmm.a Na mmmm.a awm ma mama.a ma mama a Na mmaa.a WM aa moaa.a Na mmva.a Na mmma.a mwm so maaa.a aa maaa.a aa maaa a NMM aa maaa.a 0a maaa.a aa maaa.a mm Na mNMN.o Na mnam.a Na mNmN.a VWM Na mnmm.a Na mMaM.a Na mwnm.a Mmm aa maaa.o a0 mmmm.a Na mama.a umM aa maaa.a Na mmva a aa maaa.a amM Na mama.a mo maaa.a aa maaa.a amM aa maaa.0 aa mMWa.a aa mmma.a mvm a0 maaa.a 0a moaa.a ca maaa.a m¢m Na mmaa.a Na mmma.a Na mNmN.a mvm Na mmmw.a Na mmav.a Na mmmN.a mvm aa maaa.a aa maaa.a aa maaa.a mvm aa maao.a Da maaa a aa mama a vvm aa maaa.a Ga maaa.a aa maaa.a MVM aa vaN.a Da maaa a aa maaa.a Nvm aa maDa.a aa mama.o ca maaa.a ¢MM am mmmm.a aa maaa.a aa maaa a MMM N maaa.a aa maaa.a aa maaa.a NMM Na mwav.a Na maam.a Na mmmN.a mNM a0 mmmN.a aa mmaM.a aa maav.a wNM ma mmam.a Na mmmv.a mm mMMa.a mNM Ma maaa.a ma mmma a ma mMWN.a MNM Na mmvm.a aa mama.a Na mumm.a eNM NO mmaa a aa mmmm.a Na mMav.a NNM N vam.a Na mmmN.a Na mmaa.a aWM aa.maaa.a aa maaa.a aa maaa.a aNM Na mumm.a ma mwaM.a Na muna.a mam ma mmMN.a Ma men.a ma mmmm-a waM Na mme.a Na maam.a Na mmam.a mam Na mmmm-a Na mNmm.a Na mana.a vaM Na maNN.a Na mnnc.a Na mmaN.a MaM aa mmm~.a aa maa¢.a Na mmaa.a aaM Na mmmm.a Ma mmva.a Na mamm.a mam Ma mMma-a Ma mmmm-a M0 mmmm.a mam Na mmmM.a Na mamm.0 Na mmmm.a maM Naomwma Nwammna Nmammha 3.208 .E 59:. 312 aa mmNm aa maaa Na maam aa maaa. aa mwaa mao maaa aa mmmv aa maaa Na mmmm Na mvam a0 mnnm aa mMNm aa maaa aa maaa aa maav aa mmaa aa maaa Na mMma aa mNaa aa mNNa aa maaa aa maaa am mnmm aa maaa Na mmaa Na mnmN aa maaa aa maaa aa mNMv Na mmma Na mama aa mamv aa mmmh Na mama aa maaa ma mmma Ma mmnt ma mNmN aa maaa Na mvmm aa maaa Na mama. a0 maaa Na mama Na maaa Na mnva aa mmmN va mama Ma mmmm. Ma mmmN. a aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa avaamma aa mnav aa maaa. aa mnam. aa mmma aa mMNn aa van aa mNma aa maaa Na mMMM Na mama aa maav Na mmMN aa mama Na mMan aa maaa Na mmaa a0 maaa Na mama Na mmNa aa mmvv aa mmam aa mev aa mamm aa maaa Na mmNa Na mNVN aa maaa aa maaa aa mmam aa mvvn aa maaa Na mama aa mamM aa mama aa maaa ma mVNa Ma mama Na mMmm aa maaa Na mNaN aa mamv. Ma mnaa Na maaa Na mNVN Na mmaN Na mNaa aa mMaM Na maaN aa maaa. Na mmav. o aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa vnawana Na maaa. aa maaa Na maaa. Na mmmw aa maaa aa maaa aa maaa aa maaa Na mmNm Na mMVM. aa mawm. Na mmam. Na mmma. aa maaa 0a maaa aa mmva. aa maaa. aa maaa. Na mamm. Na mmma aa maaa aa maaa. Na mMam. aa maaa Na mama Na mama aa maaa aa mama aa mhnm N maNN Na maav aa maaa aa mamv. Na mamN aa maaa ma mmaN ma mmaN aa maaa aa mnmv. aa mmmm. aa maaa. Na mmaM aa maaa Na mmaa aa maaa Na mmva Na mMNa Ma meat ma mNMM «a mm¢v. vmaNana acdaiaciatficia(Dana1565a<§u$a<§c$a<§a$at§ciacseaacicunaoacduiatpcaacficiacficiacnqia aa mama. aa maaa. Na mMNa Na mnva. aa maaa. aa maaa aa mamM 0a maaa Na maaa Na maae aa man Na mMNm N mva aa maaa aa maaa. aa maaa. 0a maaa. aa mNNm Na maaa. Na mamm. Na mvah. aa mva. Na mmmN aa maaa Na mMaa. aa mNMN aa maaa. aa maaa aa maan aa maaa. Na maaa aa maaa aa maaw Na mNmN aa maaa ma maMM Ma mmam Na mmaN. aa maaa aa mNNN. aa maaa. aa maaa aa mNmn. aa mmav. aa mama. aa mamv. Na mNNN ma mnaa Ma mMNN. Na mamN. aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa vaawama Na maaa. aa maaa. Na mmWN. Na mmaa. aa maaa aa maaa aa mmma Qa maaa Na mnmv Na mmma aa mamn aa maaa aa maaa Na mva aa maaa Na mMaa aa maaa Na maMM Na mama aa mmav aa maaa aa mmaM aa maaa aa maaa Na mmma. Na mmNN. aa mNnv aa maaa. aa mamn. N mmma. Na mamN- aa mama. aa maaa. aa mmmv aa mama. ma mVNM. ma mmam. Na mvmn. aa maaa Na mNaM. aa maaa. ma mnaa aa maaa. Na maaM Na mmnm. Na maMM. aa mmmN Na maaa ma mmaN. Na mmvm vvamama acaciacbcaacnciatnciacpcaanficaacpc>ac9ciatsciacéciacfiaiacéciacficia<§a$a<§¢$acnaaacn Na mama. aa maaa. Na mama Na maaa aa maaa aa mvvm aa mmaN aa mmav Na mmam No mmaM aa mmmm a0 mach Na mmvv aa maaa aa maaa aa mnam aa mama N mmma Na mwwh N mmmN. N mwmv ma memo mo mama aa maaa aa mmNm as mmmv aa maaa mo wmmm «a wmmm Na mNVM «a mmmv am momm aa mmam Na mmaa aa maaa Ma mamN ma mmam we mmnm ca maaa aa maaa aa maaa No menu «o mava ao mVnm «a mumw aa mNan N mama Na mama Ma mVNN. Na maam. vmawamo atpaiacdcia<§c5a<§a$atfiaiacsciacéc$a<§cfia<§c$acfiGSanficia<§¢$at§¢ia<§¢$a<§c$a<§c$a<9 Na maaa. 0a maaa. N0 mmaN. aa maaa aa mhev aa maaa aa maaa aa maaa Na mwmm a9 maaa. aa mavm Na mmah N0 mvnv. aa maaa aa maaa Na maaa aa mmmv Na mmMN Na mMav Na mmva aa maaa aa mvmw aa maaa aa maaa aa maaa Na mmva a0 maaa aa maaa Na mvma Na mvmv Na mmvv aa mavm aa mavv Na mmvN aa maaa ma mmmN M mmaN Ma mmaa aa maaa Na mama aa maaa Na mmam aa mnvm N0 mvwa aa mmvn Na maaN Na mmma M0 mmwa M0 mvav N0 mmmm. mmaamaa oa<§a>a<§c>acsc$a49¢:aa<§c>aatsc>atsc$a<§cia<§a5a<§c$a<96>a<9 Mam awn amM mm mmm mmm mam va vvm mwm NvM VMM MHM NMM mNM NM wNM WNM ¢NM NNM aNM aNM mam aaM mam vaM mam aaM mam aaM mam .ULcoU - Am mama; 313 N6 mvnn. a6 maah. a6 m6ma 66 m666 N6 muma 66 m666 a6 mvvm N6 mmNa NNOMNN6 66666666 N6 mNna. a6 mM6N a6 mnma a6 mawN N6 mMMa a6 mMNn a6 maMM N6 m6aa NM66aM6 - . a . o . I M6M N6M a6M aNM mun NNM mum nNM w v6 mth. N6 maaM. 66 m666. a6 mmnv. N6 mmNN a6 mamm 66 m666. a6 maaa. 6'66MM6 60666666 N6 mNma. a6 mnwn. a6 mnma. a6 mamN. N6 mmma. a6 mMNw. a6 maNM. N6 maaa. VM6maM6 OQOOOOOO 6cbci6cbc$6ts M6 mava N6 mNMM. N6 mMMa 66 m666. a6 mMVM. 66 m666 N6 mmaN. N6 moNN MN66aa6 N6 m6vm N6 mVNN. a6 mun? a6 maMh N6 mMNa 66 m666 a6 mhmm. N6 m66M. NN66NN6 06666666 M6 mhna.6 N6 mvam. N6 mMNN a0 mwNa N6 mmmm a6 mamm 66 m666 N6 mamv. VW6N6m6 6666666 N6 mhmm.6 66 m666-6 a6 mmma. a6 mmmm. N6 mNmm. 66 m666. a6 mwnM. a6 mnam. MM6aM6a 099009 N6 mvmv. a6 mama. 66 m666 66 m666 66 m666 66 m666 a6 m666 a6 mvwn. Na66aaa 66666666 M6 mmaa. N6 mNMN. a6 mamm. 66 m666 N6 mnmm 66 m666 a6 mN6a N6 mvma. NN66NN6 66666666 M6 ma6a.6 a6 m6NN.6 N6 mNNN.6 66 m666.6 N6 maav.6 66 m666.6 66 m666.6 N6 mmmv.6 VM6N6m6 N6 mMNm. N6 mamm. N6 ma6N 66 m666 N6 mhvv N6 vaa 66 m666. N6 vaa. MN6aM6a 66666666 M6 ma6m. 66 m666 N6 mmMN N6 mava. M6 mava 66 m666. a6 mamw N6 mamN. N66aM6a 66666666 M6 m66N. N6 mwmm. a6 mamN N6 mvaa N6 maam 66 m666 a6 m666 a6 mmmm Na6mNN6 6c§c$6<§a$6<9 M6 mmma. N6 m6Na. a6 m6va 66 m666 N6 maaa 66 m666 N6 ma6a N6 mmma. va6N666 aa6c6¢$6e6¢>6 N6 mmav.6 N6 mmmv.6 N6 mNVN.6 a6 mNav.6 N6 mMNm.6 66 m666 6 a6 mmam.6 a6 m6va.6 MM66N6a 0 66666666 M6 momv N6 mNaN. a6 mnmm N6 maNa N6 mmmm a6 mmNN a6 mask. a6 mmwv. NaoaM6a M6 mm6N. N6 mmvm. a6 mmaa a6 mme N6 mwmm 66 m666 a6 m66N 66 m666 Nm6nNN6 66666666 M6 mwa. N6 mmNN. N6 mama N6 mnma N6 mmNm N6 mNaa a6 m6MN a6 mM66. vv6maM6 6:6656c665616 N6 mmva. N6 mnvm. N6 mava a6 mmmm N6 mv6M a6 mN6M a6 mN6n N6 maaa. MN6aaa6 66666666 N6 mamn. 66 m666. a6 m6MM. a6 mvam 66 m666 66 m666. a6 mvvm. a6 mm6v. Na6aaa6 M6 mmom. N6 m6aM. 66 m666 66 m666 N6 m6a6 66 m666 N6 mnNa N6 m6na Nn6mNN6 66666666 M6M N6M aaM amM 6mm mum mum mum M6M NNM awm aNM 6mm NNM mum WNM Mam Nam a6M aNM 6mm NNM mum mmm mmm N6M a6M aNM 6mm NNM mum mum N6 mNNa. 66 m606 N6 mNna N6 mmoa 66 m666 a6 mMaM a6 mw6N a6 maNM N6 mwvm N6 mmnN a6 mMNM N6 m66v N6 mvnv 66 m666 66 m666 a6 mmwh a6 mNaa N6 maNa N6 mmnm N6 m66M N6 mmmc 66 m666 N6 vaa 66 m666 a6 mNNa a6 man 66 m666 66 mvmm a6 mmam N6 mamM N6 mmmv N6 ma6M a6 mMav N6 m6VN 66 m666 M6 maaN Mo mnua N6 mama 66 m666 66 m666 66 m666 N6 mmNN N6 mava a6 mnmm a6 mvaN a6 m666 N6 mvaa N6 mNnm Me mNaN N6 maaM. NN6MNM6 0 66666666666666666666666666666666666666666666666666 N6 m6Na. 66 m666. N6 mwma. N6 maaa 66 m666 a6 mvvm a6 maaN a6 mmav N6 mmam N6 mN6M a6 mme a6 mmvm N6 mmvv 66 m666 66 m666 a6 mnan a6 m66a N6 mMMa N6 mnnm N6 mamN N6 m666 66 m666 N6 mmma 66 m666 a6 maNa a6 mNav 66 m666 66 mman a6 m666 N6 mnvm N6 mmmv a6 m666 a6 m666 N6 mnaa 66 m666 M6 mamN M6 mmaN N6 mmnm 66 m666 66 m666 66 m666 N6 m6MN N6 mava a6 mvnm a6 mnnN a6 mNaN. N6 m6aa. N6 m6M6. M6 mvNN.6 N6 m66M.6 NM66aM6 666666666666666666666666666666666666666666666666 VNM MNM NNM aNM awn 66M umM 66M 66M va Mwm amM 66M 66M 66M mnM 66M vmm Mum NMM awM 6mm avM 66M nvM 6VM mvm VVM MVM NVM vMM MMM NMM aNM 6NM mNM mNm vNM NNM aNM 6NM aaM 6am 6am vaM mam 6am 66M 66M 66M 3658 .i- wags APPENDIX G @mpounds omitted when calculating statistical normalization factor In performing statistical comparisons, the following compounds were omitted when calculating a normalization factor. (See the Methods section for details). Compounds listed in G1 are referred to as the “35 compounds omitted” in the text; those in Table G2 are referred to as the “12 compounds omitted.” Figure 61. 35 Compounds omitted. k Compound number Compound name 18 6 -hydroxybutyric 50 phosphoric 1 12 py roglutam ic 120 tropic 152 oL-glycerophosphoric 168 citiric 184 homovanillic 197 hexuronic-peak 2 198 unknown NE-5 215 hexuronic-peak 3 2 1 6 vanilm andelic 225 hexuronic-peak 4 314 Figure G1 (Cont’d.) 315 @mpound number 237 257 266 290 297 302 303 306 308 309 318 319 324 325 326 360 361 365 378 383 Compound name hippuric uric m -hydroxyhippuric m -hydroxypheny1hydracrylic unknown U5 (cresol) unknown U10 (4-deoxyerythronic) unknown U11 (4-deoxythreonic) unknown U14 (2-deoxytetronic) unknown U16 (erythronic) unknown U17 (threonic) unknown U26 unknown U27 unknown U32 unknown U33 (hexuronic-peak 3) unknown U34 (hexuronic-peak 4) unknown U68 unknown U69 unknown U73 unknown U86 unknown U91 Figure G2. 12 Compounds omitted. @mpound number 50 120 168 237 257 266 308 309 Compound name phosphoric tropic citric hippuric uric m -hydroxyhippuric U16 (erythronic) U17 (threonic) Figure G2. (Cont’d.) Compound number 318 319 325 g 316 Compound name U26 U27 U33 (hexuronic -peak 3) APPENDIX H Clincal report form This form is produced from a single MSSMET output by a computer program called MSSRPT. The relative peak areas calcu- lated by MSSMET are converted to concentrations in mg/ ml using a file of correction factors, when these factors are known; otherwise, data are reported as relative concentrations. Each concentration is compared to a table of mean values and standard deviations (taken from a table similar to that shown in Appendix J). In the example shown, this is a table of normalized relative peak areas of a group of juvenile control samples. The correction factor listed on the report form is used to convert relative concentrations to normalized relative concentrations. Each datum is plotted as the number of standard deviations it is from the mean. (E.g., if the mean 4.: the standard deviation is 5 a. 2, a value of 9 would be plotted in the “+2” column of the table.) In addition, the minimum detectable value is estimated by assuming it is zero, and plotting a “greater than” symbol at the corresponding location (e.g., in the above example, a value of zero would correspond to -2.5 standard deviation units). 317 £318 URINE PROFILE HNRLYSIS MSU/NIH HESS SPECTROHETRV FHCILITV CURRENT DRTE: 25-JUL-77 DRTE 0F MSSMET HNBLVSIS: NHHE 0F OPERRTOR: S.CHTES ID NUHBER 0F URINE SHNPLE:090?6N5 RLL VRLUES CORRECTED BV FRCTOR 0F 0.690 SRHPLE DESCRIPTION: NEUROBLHSTOHR URINE FROM OR. KRIVIT 0?-flPR-?7 ML URINE EXTRRCTED: 1.000 H6 CREHTININE/HL URINE: 0.180 UG INTERNHL STHNDRRD: 50.000 9915 FILE OF REFERENCE 991955 1951 90917199: 25—991-77 9915 FILE or 9-9991095 1951 90917199: 14-999-77 99999994494999m49mu9*9999m99999399999999999999999949499999999949299999999999999499)» 99999999 99799 c 99995 IN 51999999 999191199 99115 -4 -2 9 +2 94 999» 109 L0“ 909991 HIGH 9999 HI id'lifitflfltiulflkfififllflktfittfltkflfittittktltfii‘tttlflnlfi****#*****fitllfll‘itt’lfiflhlflhhkikfllfiflflhmkflt 293 91 19.599» > 9 5 9-99999991599919919 2.547 >9 9 199119 22.499» 9 294 92 5.399» > 9 19 91999119 17.999» 9 17 91»ox»119 OXIHE 1.549» > 9 19 9-99999999919919 9.199» > 9 295 94 9.979» > 9 297 95 (999591) 139.999» > 9 23 OXBLIC 13.999» > 9 299 95 1.219» >9 343 951 7.499» 9 344 952 2.179» > 9 299 97 1.779» > 9 29 91999991 5.459» 9 35 9919v19919919 9.445» > 9 34 999 99193 (9199991197? 9.555» 42 999-9991 2.739» > 9 391 99 <2-9919»191vc9919> 29.999» 9 59 9995799919 1219.999» 9 392 919 (999999991999919) 93.399» 9 51 9992919 5.443 > 9 393 911 <999xv19999919> 27.599» - 9 59 59991919 71.979 9 59 9999099 992 1.799» 9 9 51 9999919 7.179» 9 55 991919 9.729» > 9 345 953 1.959» > 9 55 999991999119 9.252» > 9 345 954 29.199» > 9 395 913 <999xv15199919> 29.199» > 9 395 914 93.399» 9 347 955 34.799» 9 349 957 9.379» > 9 77 91919919 5.795 9 9 92 9119999119 1.199» 9 95 99119 35.339 9 399 915 <9991999919> 523.999» 9 372 999 (3-HE-GLUTRCONIC) 29.499» 9 FIGURE H1. Clinical report form . 2 -_ 3“-"':T'.'.1'.:'."‘."*"%-=—- I319 999999999999999999999999999999999999999*999999999999999999999999999999 COMPOUND UG/HG C RHNGE 1N STflNDflRD DEVIRTION UNITS -4 -2 0 +2 +4 ” 'VERV LON LOH NORMRL HIGH VERV HI 9999999999999999999999999999999999999999999999999999999999»9999999999» 351 959 39.399» > 9 399 917 (19959919) 299.999» > 9 99 991919 17.599 >9 195 3—951991991919 1.959» > 9 194 0-99999999592019 21.399» > 9 352 959 49.999» > 9 319 919 43.599» > 9 197 9-999909991919919 22.999» > 9 353 951 55.799» > 9 119 999999995199191919919 21.129 9 9 313 921 24.999» > 9 354 952 35.599» > 9 374 992 , 9.999» > 9 111 9-99999999592919 9.521 e 112 999991919919 129.429 > 9 375 993 29.599» > 9 314 922 7.949» > 9 114 0-9999999995991995119 9 791» > 9 129 199919 (19159991 519.) 279.999» > 9 124 99999901991995-9599 1 23.999» > 9 125 9-951991919919 99195 29.499» > 9 377 995 29.599» >. 9 127 9-99990999592919 31.999» > 9 129 9-9999999995991995119 7.993 > 9 315 924 34.999» 9 9 135 9-9999999995991995119 17.399» > 9 379 995 54.999» > 9 135 9199991991995 15.199» > 9 139 99999919 9.229» > 9 141 5995919 2.999» > 9 144 9—91995909905999919 4.499» > 9 355 954 9.939» >9 357 955 23.299» > 9 152 9-91995909905999919 31.549 > 9 319 925 14.799» 9 9 319 927 99.199» > 9 359 957 5.999» > 9 359 959 197.999» > 9 159 911919 514.999 > 9 329 929 3.129» > 9 174 9251919 4.599» > 9 172 159599199119 3.919» > 9 177 99911119 29.929 9 9 391 999 29.299» > 9 392 999 197.999» > 9 321 929 21.399» > 9 194 909999911119 411.999» > 9 195 991991099-1.4-1991995 9.379» > 9 193 3.4-91-99~9959.995119 4.299 > 9 197 999995911519 4.299 > ' 9 299 9-0999599199999999119 243.999» > 9 195 95991919 7.559» > 9 322 939 9.411» > 9 199 9919919991991995-99 2 19.999» > 9 199 9999999 955 44.799» > 9 FIGURE H1. (Cont’d.) 320 *********$*4¢*4¢*4flk**********4¢***********83*“!#4“:$40.“?********$************ 99999999 99799 c 99995 19 51999999 959191199 99115 "'* -4 -2 9 +2 +4 9599 199 199 999991 9199 9599 91 *************4¢******td‘ifltlklktlktiflhlfllflhki*Ill’ill'**************************#*## 324 U32 62.3000 > * 194 O-COUNHRIC 14.883 > * 197 GHLRCTURONIC-PEHK 2 62.7000 ) * 200 GLUCONIC 24.0009 5 * 207 P-OH-PHENVLLHCTIC 6.3208 >* 364 U72 20.3000 > * 216 UHNILMRNDELIC 4200.000 > * 215 HEXURONIC 84.6000 ) * 383 U91 83.1000 > * 325 U33 292.0000 > * 365 U73 88.1008 ) * 326 U34 371.0004 ) * 225 GRLRCTURONIC-PERK 2 86.500# } * 212 PHLNITIC . 9.530“ > 4 367 U75 1.5803 > * 237 HIPPURIC 504.000 >* i 236 314-DIHVDR0XVCINNHHIC- 1.3108 0 369 U77 6.570“ > * 329 U37 46.4000 ) * 244 INDOLERCETIC 16.324 > * 246 UNKNOWN NE8 5.7004 > * 251 314-DIHVDR0XVCINNHHIC- 6.832 0 258 FERULIC-PERK 2 1.620“ > * 332 U40 1.2100 >* 333 U41 69.4000 > * 266 H-HYDROXVHIPPURIC 82.8008 > * 334 U42 40.7000 > * 274 314.5-TRIMETH0XVCINN 1.680“ 3 * 272 5-HVDROXVINDOLERCETIC 1.442 > * 29 LEVULINIC 0.000! 3* 67 NICOTINIC 0.0004 >* 79 313-DINETHYLGLUTRRIC 0.000 >* 96 HRNDELIC 0.0000 >* 118 PIHELIC 0.0008 >* 123 THRTHRIC 0.0004 >* 155 CIS-HCONITIC '0.000# >* 176 ISOCITRIC 0.0000 >* 209 SEBHCIC 0.000“ >* 218 HSCORBIC 0.000 >* 223 HYDROCRFFEIC 0.0004 >* 235 FERULIC-PEHK 1 0.000% >* 254 UROCHNIC 0.0004 >* 257 URIC 0.0000 >* 328 U36 0.000# 3* 342 U50 0.000% >* 350 U58 (3-"E-GLUTRCONIC) 0.0000 >* 361 U69 0.0000 >* 363 U71 0.000# )* 366 U74 0.000“ >*, 368 U76 0.0004 >* 371 U79 0.0004 >* 376 U84 0.000“ >* 379 U87 0.0000 >* 9*9999999999999999999999999999999999999»999999999999999999999999999999 KEV: > SHONS HINIHUN VHLUE DETECTED IN SYSTEM. 8 IS COMPOUND FOR HHICH REFERENCE VHLUE HRS NOT BEEN ESTRBLISHED. 4 IS COMPOUND FOR NHICH K-FRCTOR HHS NOT BEEN ESTHBLISHED. HENCE VHLUES HRE REPORTED HS RELRTIVE HMOUNTS / HG CREHTININE. FIGURE H1. (Cont’d.) APPENDIX I Complete MSSMET “found” file A complete MSSMET “found” file is shown in Table 11. The entries have been explained in the caption for Figure 14. This “found” file was produced using the library illustrated in Appendix B to analyze the organic acids in one sample from the group of subjects with neuroblastom a° 321 322 0] IV] N III-N .- . M H N NNN NWfi and med de dVfi (‘J l"~- Nova NmMfi mde NMMH mwmfi mama nmwd mVNH mwmfi «o mumm.o .owoma mm + 50 mamm.m .mmwm m + vouo + «a «a «o mama.o .mmmmm mm 4 N med mMH mm« o VMNA memo + «0 ma co maaw.@ .mMfimu mm + m ommczau «w m: wmm do mmv«.o .omom mm + «a umflm.m .mmmfia Mm + mane + mv «N do mmva.o .mmnmm mm + « vvd hma HMfl w MMNH mama 9 mm mu «m mum«.o .nawwm 4m + a muzuxz: mm uchxo mm «o wmmw.@ .cvmmm mm + «o mmwd.o .wmmm 4m + vane + mmufiw do uvvm.o .mmmmmd mm + m mvfi vma mm“ o nmwa mama + mv ma «0 mvmd.o .mwanm mm + A unzmuoam mm mm: «mm mm wmm«.o .Mmmnm mm + No momm.m .Nmmmmm mm + none + ma «N No mmv«.o .mmmmmw mm + w wva mma mwa m Anna ma 0 + Mm ma Na ummm.a .ovm«ov« mm + a «a: mom m: nmu so mmmN.o .Ndmu mm + do mmma.o .mvmoa mm + comm + mm em 00 mmmm.o .vwmm mm + m mm“ mma flma « mama ma.a + mo ma a0 mwma.o..e«mmm mm + u onomzmm an en mmm No mmma.o .mwmmm am + «a momw.o .mfiomm mm + vane + nmnom N9 uvma.o .ommomm mm + w de mad 994 m mmfifi mono + mm ed «a mmmm.o .Nwmmma mm + a ma: mom mo~>m»:m>xoma}:1mv umozchnm>xomo>xnm ma No wmmm.o .oovvnm mm + om mamm.m .mmav mm + mono + no on no wvmm.s .ommaoma mm + H 4N“ mad wag m wmfia Nona + mm va as uddm.a .mmmwa mm + m ouaozmmozm an Amzmxo o~4>xorqov mzuxo uuhmucoxo «a mo mmfim.o .amfim mm + we mwom.o .vomvm vm .+ memo + «N ma as m«o«.o .vvmom «m + N mm wk mm N amaa mafia + «a Na «0 mamm.o .vamnua vm + a m: «cm moaqooygov o~wwoc>xomoyx 0a 00 mmam.9 .nvua mm + an mmmM.o .ovam 4m + Nona + mo ma on wafim.g .smu¢ mm + a mm mm mm m Mada mono + «muad we mmwv.o .mmvm mm + « «mxouxz: me N: ¢mm so wmva.o .mmm mm + as mmmm.o .«mmma mm + meme 1 mm mu so mmv«.o .mvmm mm + N mm «m mm a “mad ammo + mm o« as mnmm.o .mvmmv mm + m Aononqrxquov mmacm x2: em manhucnv uaozmmomm>xomo>xnm m 00 wmoM.o .vmma vm + do mmvw.o .Moova mm + ammo + anm« as wmwm.m .ommm mm + N mm «m m4 a whoa mama + mauau do mmmw.o .«mmmm mm + w uuzogczq>thx mm Aomm>hsmom~>xoaoyxucv oHozmmommnyrhmzumuyxoma>11~ m as mmmm.o .vmmfi mm + «a mama.@ .mvmm am + mmma vane + do we so mmmm.o .ovam mm + N mm 4d N Na: mmm mama 1 menu «a mom«.o .vmmmm mm + « Ao~rmhamyxoma}:uov onozmkam>xomoyznv mm on: mvm amlmomw.o .owv mm + a01mmma.o .mm mm + Mans + mm ma om mmo«.o .vmfiw mm + N ma ma w w: mwm game 1 mmnn flonmwmm.o -Nmaa vm + m . 0H204mz mm a: me Houwmmm.o .vmv mm + as moao.o .mammma mm + mane 1 Mmuma «oawmmw.o .mmmu «m + m mum 0mm 9mm a send comm + mm «M 0o mooo.@ .mmmmmv mm + a Aumz~43>m4v oHozchzwmoxonv mm ammomp am wmda m .am4w mm + .m.~ a: moo on mono + mg ma «a Mama.o .mamvu mm + a Agomuquov Jonuhwzcmommum.m.a mm ~>~ux.mzmmomo 4: @.m mm mm ma momamoudzo go mvNN.e .mmma mm + mzmznwcmmu o: om«.m News + mmuma we mNMN.o .mmmv mm + a 4: coo.« mm: vvm Nomamo «ya so wmmm.a .Nnam mm + Hmu mHJFm m «x0 {ozsou maznomzoo9 nuimmcuuo mono + mm ma ea mmvm.o .tmoma mm + « Hm: mvm 2: 22:8: emzmmz 39388 AHWAQxamm¢o1m1wzoozoam>m1~ NHH 3: mom as mmuv.m .VVMN MM + as mNmM a .MMQDN Mm + HmM me «mm o mmw H 09. a + MM. am so mmov .a mvvo + N 69M “mm mmN m quH memo + nmth «9 MHNv a nmm¢m mm + N UHOszmyxomomxl 1: «NH am: own «a uNmN a .QNNvH mm + H9 mmvN.9 .NNQVH vm + vwm va mvm Mu wwwH mama 1 mN an «a memm.o .mavmr mm + H vQM mmm mmm d H¢MH Nona + hm .N «a uMmm o .mmmmm mm + N mm: vmm ag—szv uhonomzmh3m>xomo>x cm «a mmvm.a .mmmMH mm + a9 wmwN .9 .ode vm + mmm MmM vvm M mme name + HN on Ho mvmv.a .wmem mm + M uwN NNN mmN H1 mend oaho 1 9M. .mN ea mmwm .o .vwmm mm H mm: vmm AUMJNEchHUV u~o~owzcr3m4>Ihwt1 N1 >>omo>11 N+ mm «a mmmN.a .owomH mm + Ha1ummw.o .er Nm 1 mwM Mmm mvm M Mmm a mg. 9 + «N QM Hm MNNN.0 .evmvm mm + N awn mum VAN H1 mmvfl a&.m 1 ma mN fim1mmmm & vva Mm + N Hm: MHM AU_M¢k:404>IPm:~aIM.MV uHQHQmZNHzmm4>Isz~o1M.m mu He mwvw.e .mmmvN am + «a mHmH.® .me Mm + 90M IVM me mvmd me. o + hm mN Ho Nmnv.9 .womHm Na + N mum sum MmN m1 vqu @610 1 mvva do meH,e .onva Na + N AuHm:»:404>:hm21m1}x0¢o>zumv uHoHGuzmkzmm431km:1M15xomo>x1m oHH nu~xchnnov uwaHemzmhzmm rm Na NMaH.o .mmtmm mm + He m1Mm.m.aNMHM mm + 00 mvm NvM mNM w mmmu no. a + 91. .MN N9 uan .9 .vmmmm mm + H MAN mwm MwN NI mm¢n Heme I «v VN Hm won.© .mHmva mm + N on Hw: mmn mm: mvm 00 «a whom.a .¢NwwN mm + 99 mmHm.m .NmHm mm + omM Nvm nMM v mmmH name + NM mN Ha wmmm o .vawHH mm + N mmN NmN NwN N1 mmVH Nana 1 NH em 00 onm.m .HmmmH mm 1 N AQ~NCFDJU)xomo>z1¢u uuonomzchzmm>xom 0:11 N 99H mm: mwm Ho wma¢ @ .mmmMN mm 1 am mamm.a .maNNM mm + m1m mmm mmm H wNwH Hams + MH mm as mmM¢.e .mmnmm Hm + H flwN wmw NmN N1 mmVH 9mn9 1 MW mm as mmmw.g .NmmmNH mm a N NH: QHM mm: mvm Ho wmam.o .wvwmm mm + No mmMN. .mhmMMH em 4 va WMM «MM v wNwd none + MHHmN Ho wmmm.m .VNmmvH mm + m va nmm mVN 0 mmeH song + anMN Na mmVN. .vanmv mm + H mm: Nmm 4H: mom on mmmH a .NHoH mm 0 do MQVN a .HmmMH mm + mMM mMM OMM N MNmH mafia 1 me mN 65 uHmH.m .vvmm mm + H MmN wvm MvN v1 ov¢H mama 1 «N MN He mmMN.a .mmmmv mm + N u~0~2mm>xomo>z1o ¢®H MH: man Hm mmmH a .mmsHH mm + H9 vam.o .mmmMH mm + Nvm MMM NNM H med Nana + H@ mm «a mHmH @ .vnmmm vm + M Mmm mvN McN o 014“ oaum + aN MN Ho umvm.o .mmamv mm o N Ho~m_o¢4>zkuz1mv u~o_ouzcxu:4>zhuz1m maH em: w¢m «a mmNm.o .mHmmN om + so vaN.o .mmmH mm + mMM MNM mHM M mme wonm + Hmth «m unflm.m .MmmMOH Nm + 9mm mam mMN m1 NMVH oHuo 1 HQ MN om mva.o .mmmm m1 + H Au~m_ocv Ufiouowzcxu: ww AQHzHFOUHZV o~4yxom¢cuwzuonm>m1nw Hm1mmNM.e ,va Mm + Hm1wNwo o .vn mm + mNM QNM NNM m Hana wfl. o + mm mm H91 wwrN .9 .Nmn th HvN mMN HI mmwd Howe 1 mm NN Halmwmm o .vme mm + H nuHJuozc:.. UHHuom>xomoyle14>zmxm1wa omkmo¢4>zwxm mm No urnm a .HmwgmH mm + co mev.m .wwvm mm a QMM mHM «HM m Nwmd mums 9 mm mN Na wmom.o .awaaNaH mm 4 N mmN HVN mmm m1 hNVH oHHo 1 Mm NN so uvvn.0 .NnmoH mm + H NH: mam mm: mwM 30.988 .: mama. Mama mono + MA mm «o AQHJJHZ¢>V 324 Aumhnzoomlmmov mvmd vouo + No mama.o MNmfl Gena + mvumm No ude.o nuHJ4~2¢>o201v do mama mama + MMHNM «0 Na mama comm + mm NM No no mama mmha + «N mm «a mm 00 NNV¢.0 Mmmd Nana + mmumm oo ummm.o AUHJCIPINwmth No mamm.a mmdd mMmH. mNQN. mmHfi. mvmn. mmfiv. mvmm. uMNa. .mMNVN mm + .Nmmmmm mm + N umhm0¢4>2w1m>XOIPmZIMI3xomoyxtv va . .mmmm mm + .mdmom vm + v mm: HNM .ommhm mm + .deacm mm + a 5m: Nmm .mooww mm + .wmwom om + M ma: Hmm .mHmn mm + .mmmvm Mm + N omoszm5xothzcm1>xomo>x1v mm .mmvw mm + .mmmad mm + N Q~4>xommcunomzuNZMm1v.d Nma .oooom mm + Mmmd mono + mm mm Na wvmm.o .mmmmwd vm + M nouakmuommv H 0~4>xommco—MszcmommIM~N‘alrxomoyzla mNH do mde.o coma Dana + mo NM “0 mmNN.o AuHCJuNCV No mmmM.o Mm“ mono + mmumm N0 mmfim. «a wmdm.o Mona + mm mM Na mch.o HS wmmm.0 home + NH mM Na mmafi.o No woad.a mmmd vona + «0 mm No mwad.o as wmvm.o mmwd Nona + Mmhmm do womm.o No uNNN.Q mva vmha + qumm «0 NVMN.Q Ho wmmm.@ vaa Gong I MMHmM Na uHo«.o .mmmdm am + .mmfimmN 9m + .tmmwm mm + .memoN mm + .mmmm mm + .Novmv mm + N ononowzczoz vLH .NNomoN mm + .mfimMNv mm + N mm: “mm Quahmo muH .ommwn vm + .ommNMN mm + M .mnvmd mm + .ONQNMH mm + N .fiNnNH mm + .Movmm mm + N .mmovm «m .NhNMON QHeromm¢QHMMuzwmom¢1M.N~«1mHo WWH H m + + N do wONN.Q .mnmaa Na + No moHN.o NM DM «0 MNNN.Q .vmnwv mm + N mm: mHM .vmwmdfl om f vmmfi None + «N mm No mmMN.o .mMNva on + N Unmoxmmoxmommo>401c NmH M M f4 99 (0 M M (”’1 O) (”l '0 t") dth Nona + mv «M unackDJOOHNMICV p. D"! P '1 v T'- meha Name + m? NM (‘1 P" omma ammo + mwuam F’I M M Ho + «a «M M M ”9 + «a «M No mum“. No mvma. No mmma. No mama. «Q thM. No mham. No «Q nunmmmnmv mo mme. «o womH. so wmmw. No mNMN. Ho mvmw. “a mwad. «a Name. do MNQN. do vad.o HQ mmmM.a do meM.o do mmmm.o .vamm mm + .dmmmhw mm + .mvwmdd mm + .QNNMNM mm + N .Hmvmd mm + hmflm mm + N Unmoxmmoxmommo>qolm vvd wm«N.o .V¢0Nd wm + wva.o .OGNNW mm + M UNOHQMZCHQO «VH .mumv mm + .MNNQN mm + v umzomcmm mMH -emon Na + .NmmmN vm + N szhQCJozomnm DMH .mmNm «m + .thMN «m + N mm: mum .dmmmu mm + .momov mm + N lomumqyzmImyxomorxlm mMH Ho mmmd.o .Mmmm mm .«mmwm Nm onhmoc4>zmzm>xamo>xlz mNd .VNNGN mm + .wmmmw mm + N + + N omomzwm>xomo>11m NNH .ovmmm mm «o wnmm.o .mvdeH mm + + uzux° oHoHomzchzuaoxo1m mm« «o mNNN.o -0vmma om + «a wmvm.o .mmmom om + N No mnMN.a do wmv«.o HO wedN.@ Aumamzumv om mfivv. 00 womv. «o meN. do womm. 09 wmwu. ma unva. .vmm0¢ mm + + a xcmmlwzohUCJOZOmcmc vNH .thNMH mm + No meN.o .mNmmm¢ mm + Qmoz¢mommJ>zwzmlmI>xomo>zlm ONH .mtmm mm + .NmMNv mm + N H omouowzchmw: NHN .mmVN vm + .wmom mm + U OHMCFGEF MNN .NmNNva mm N vw: mum .ovw mm + .mamN mm + « u-moc4>2mzm>xomo>110 VHN 1.588 .2 35.9 325 b. N '0 0mm «mm mNm ADM):¢}mq)zm1m10Imv mfim mmw Ao_m=v wz~mnm>xoao>xnmhIm.w- mmm so ummm.o .mnom mm + dam mum a mmwm «one + MM me «o mm««.o .mmmmm mm + u AommumcUV m xcmmnouzczz~o>xomo>xmoIv.m «mm «a umnm.o .oommv um + mnm mwm NI mmmm wane I ma mv no mamm.o .vmmnwu mm + m . mmzuyz: mv~ do wmn~.o .mmmuv em + mmm mmm a mmam smug + gm cc «0 unmn.o .muvmma vm + m ompmo¢4>JQoz_Im vvm as mmfim.o .mmmna mm + «mm «Va 9 mmum flame + mm mv «a mmnm o .mavmu mm + « um: mmm do w~ov.o .mmmmm mm + mcm ¢vm m- mvam Nana I mm M? ac ummv o .mammm mm . a ~53 mwm as u«¢~.o .vwmma mm + m¢m «mm «I mmum mono I mm me «o unn~.o Adm¢nn mm + « awo~o~omz¢uuo>xomo>1v mm: mmm so umm«.o .mmm mm + mmm mum o oflam none + anme 00 umum.o .mmW¢ an I m mm: mum on moo~.o .«m«« om + mm» «mm m- mofim mono I mm we we mmv«.o .mamm um + m AuHmuumov a ycumuo~smzz_o>xomo>:~oIv.m mmu mo ummn.o .amvNV¢ mm + vmm mum m wean mono + nwnmv mo mnm«.o .mnmonnw mm + u . Ao~¢amm~zv ohhmucoo~zc~zum nmm «oImnmm.o .mmm mm + “mm man «I mmom some I ma me «oum~«m.o .m¢m« vm I a Auudnmumv a xcumuuuzczz.oyxozhuz-muyxoao>zIe mmw «a mmm«.o .mmooa mm + mmm mum « mmow moms + no ~¢ «o wmwa.o .oo«¢m am + H m~= ~wm so manm.a .wmom mm + mam mmm o mmom «one + mo «1 oo moan.o .vm~o« mm + m Ao~h~24¢¢v onozcomocxmz man so uvmn.o .mmmv om + «um man a mwom meme + mm «c as movm.o .mmmma mm I m «now ya: «mm «ouwnmv.o .mvm mm + man v«m‘ «I «new «fine I «N «a «oumwmm.o .«moa om + a “am How vI do mew.o nmmm Gena + «Nuwv N0 mofi«.a .MMmmv mm .nmmdNN wm «v: + + N MMM No moNn.o .Hmmomd vm + mmNN Name I «cum? ND mknm.o .«Dmemd em + N uzuxo uuozcmoumln4>mem>xomo)IIvVIMon0IN mNN «Q mmmd.o mMoN vona I Mnuov do mmmN.o .ovmofi mm + .vmomm mm + N vm: mmM 00 mmmv.o .NMNN mm + ham «an mom MI NMQN Nana I mmHov 00 mumm.o .vaNfi mm + v nonmmumoomoyxv UHozmmommlaa)zm:m}xomo>:~oIv.MVIM MNN do mmdm.® .meww mm + mam mom mam «I NMQN «mum I mvuow No madd.o .mmmdmN mm + N N xcwmIonomnhucchIa mNN No mefim.@ .ommmmN em + «N mam van 0 NmoN mono + mv a? No mmmm.o .NMONWMH mm I N munzomahomumov vm: mum No mmdd.o .«finmw mm + mam mam vow m NMQN mono + mvumv N9 mama.@ .mONNaM mm + N Mk: mmm No mnmN.Q .Qmwmmfi mm + ¢ww Now um? «I fiaom mama I mfi a? No wveN.O .QMNMNV wN + « no~zoxzhom4cov Mm: MNM No wme.o .mHNde mm + mam Nam mm? ml mGON mono I ma av No mmhv.® .mmmmmm mm + N am: Mwm mo mmmv.0 .mmnN mm + mgm flow mm? o mNGN Gama + mama? Do wmmn.o .«mwva mm + a Qummocmc mau «a mmmm.@ .wmwNM mm + man Nam Mm? v dfloN mono + manmv do wvmm.o .mmawmd mm + a d McmmIonomDhUCJCOI: MAN No maam.9 .wmmvmd am + vaw mmv Mm? NI NQON Mane I dauov N0 mmmw.0 .mwommm mm + N Aoaqmozct4mzc}v QHhmocIAJ>zu:m}uoth:IMI}xomo>1I¢vIrxomorle mHN . No NNQd.Q .mmmmm mm + dam mm? mm? NI wmmd name I mmumm No MNNd.o .mdmmvN mm + a mu: vmm «a mmmm.o .wmvoM mm + mm? mm? Mm? m mmmfl 99pm + mNumm «a mmmm.@ .vmfimmd mm + N mUHPU¢44>zmzm)xomo>IImv umozmmommIm4>zmxm>xomo>zlvvIMI>xomo>IIN kmN do mavq.m .vnmh mm + mmv mmv mm? NI Hmmd «cum I mm mm «o womN.m .mommm mm o N am: mmm do mucw.a .mNmmM mm + vmv mm? mm? « hmma Mama + W? mm fia mvvm.@ .vamomd Mm + N xcmmlomzomakQCJCOI: “ma “0 mMNd.o .mfimm mm + va nmv H~¢ NI mwmd Mono I hmumm do mmod.a .mumoN om + N Aommctaoolov o~=¢zz~o>xcuo>xlo vmd Ho whvm.o .MNmmM mm + cmv mhv mmv N nmmd caua + mv NM HQ mvvm.@ .VQNONH mm + N NM: vNM «a MNNn.o .vwva mm + Wm? mm? mhv m mwmfi oano + we mm «a mem.@ .mmmmod vm + M waIxZD mm“ Na mmnM.a .mnmooN mm + va wmv owe a Nmmd mo”o + «summ N9 umvv.m .omwvom mm + N Au~grmocao}IoIOIzv omozcmommlA4>zwzm>xomo>IIMvIMI>xom0}:IM 0mm @388 .3 Human. 326 vvm va mmm v muVN mane (‘d 4- mmm hmm vaw ml mwMN mafia I «alwmmm.9 .vwv Hm mmnmv fiavadm.m .Nvmu mm xxmmlumthZHurxokazmmhIm‘V.M VLN No waN.9 .mmmwwa mm mmnwv No womm.m ,mmmmam mm « AU~¢31m~I>XDmQ>IIEV Umhmucoonzcmzmm>xomo}zlz wwN 3288 .: wage ’1 I -—-I— .. APPENDIX J T-test of log10 of normalized data and tabulation of compound means, standard deviations, standard errors and coefficients of variation MSSMET “found” files from 9 BCIU, 5 neuroblastoma and 5 juvenile subjects were analyzed using MSSTAT. The normalized relative peak areas listed in Appendix F were all converted to their logarithms (base 10) before further calculations were performed. Compound names correspond to the numbers listed in Appendix B. As an example, the first entry in the table refers to compound 6 ( at-hydroxybutyric). The first line of numbers in the entry corre- sponds to the mean, standard deviation, standard error and coef- ficient of variation of the compound for the 9 BCIU urines (group 2); each value is expressed in exponential notation. The second, third and fourth lines of numbers give corresponding results for groups 3 (neuroblastoma), 4 (juvenile control) and 10 (sum of groups 2, 3 and 4). The t-test section of the entry compares each pair of subject groups. Thus , for compound 6, the value of t for a comparison of 327 328 groups 2 and 3 is 4.37, which is significant at the 0.1% level, as indicated by the four asterisks (l = 10%, 2 = 5%, 3 . 1%, 4 = 0.1% level of significance). These values are computed using standard techniques found in any elementary text on statistics. ll“:- 329 an n n m NNDJ£> wu34c> mu m mw34c> mu m n a mwba¢> ma n n a mw24¢> OZ .02 02 oz 02 muz¢u~u~2a~m co wooo.o so umnu.o co ueoe.o ea uwnn.o Go wooo.o 00 muav.o ea mooo.o co unmn.o at) umou cam Arm 9 . loo uUZCQ—m—zonm co mooo.o 90 quN.o so meoo.o co umon.o so woao.o 99 mmmv o co mooc,o co uvnN.o ¢¢> umou mum chm mUZCUnuuzuum oo mooo.o co wOOQ.D oa woco 0 so wcoo.0 co wooo.o 0° wooa o 09 mooc.o 00 uoao.o ac) umou «cu arm .9; uuzcuuunzanm No monN.o «quNNv,o No wwsu.o «Ovaom.o No u~«u.o «quMaN o No maca.o «QIwmmn.o cc) uwou new chm 9009 0.... muzcu—u—zo—m .0 waoo.o .0 momm a N0 wmnM.O .0 unaw a Na un«a.o uqumwv.o so mac. 0 o. unnn.o 18> uwou cam ohm n? a an." Q«.e h «0 mac” 0. mmou. «0 made no maaa >wo ohm m~.. NO.NI «v.NI » do um0« do uvun co munm OD m«on >wo ohm 00.0 90.0 a... b .0 maaa. oo woeo. co mooo 0o wooo >ua ohm NN.«I mo.o 9“." h 00 umau 0° uNON «OIunnm o. unbu >uo Ohm MM.«I an,wI on wl h «0 won“. 00 wmuv. co uvau, do anew. >uo ohm doéo> oooa> 9966) 6600) 00.9) v v v a n o mmooo x mmooo _mooo “o moon I oa on «a man“ I 1 mm «o mood I M on a. ovoa I ~ on zoo: mmcoo oz oono v n v o n ~ mmcoo x mmooo .omuo_ «o um'«.I o" .m «o woaa.I V vn o. mono I n .n «o u~o~.I o «M zoo: mmcoo oz coco e n v a n u mmooo x mmcoo omooo «o mono I o. on «o mono I v on «o mono I n NM no wom~.I o NM zoo: «mono oz oano v M v ~ M m mmcoo x mmcoo ”smock oo moon o on oo oo moon o v on go moon 9 m co co ammo o o oo zoo: mmoao oz coco v n v N n ~ mucno x moons Homooo oo moon I on no no mnaon v m oo onuo.o n n~ «o oooa.I o no zoo: mmcao oz coco O. .00. uoz¢o_o_zo.m ma no uoov.o oo oaoo.o o oo mooo.o oo ooo~.. o no ovum o oo uo.~.o a no u~o«.o o. o~o0.o moon¢> .ox a¢> aooo «so arm 9. O... moxoono.zo_m ma oo uooo o oo ommc.o n no uooo o o. oman . n ~o mono.o aoIovm» o m oo uooo.o oo ovmn.o «mono; .oz x¢> uuoo «mo arm 0. ’0! oozoo_u.:oom a. No oom~.o «quoooIo n no on?“ o oo ounan o ~o ouoo o aoIowom o a on om¢a.o aoImoa~,o moon¢> ,oz m¢> umoo mam opm . 0Q oo:=o_o_=o~m m» «o mn~m.o aqumno o n no wovo o «quoao o n ~o o’m«.o «oIoooo o o ~o who“ o no unoa o mw34¢> 02 11> uwou cam arm I. .909. oozoo.o_:oom ma No unom.o aoIoooo o n uo moon.o oo mood o n «o moom.o oo oaoa.o m «o moan o aoImvuo.o muono> oz o¢> ouoo «um ohm nm N v m «v.M v N Na.“ M N h > mmcau x mmCJQ .hmw—h no unun.o co uNmm o on mu «0 mama o «o uwmn I 7 on 09 mawv O 00 unmn o M an no m«n«.o no man" 9 N o« >uo ohm zxu: mmCJU oz omru ma." 1 m vm.~I v N om.vl M N h r mmcdo x mmmou “pmwhh «a me«.o co mama I 9a mu «0 mafia O 99 uhqm I v mu so uvm«.e 90 udwN o m Na «0 u90«.o do moan I N ma >uo chm zzwt mmcqu o: omzu an NI v n QN e v N M0.n M N r r mmcgu 3 mumgu hmubp Go unen.o HG mama a on 9‘ aa Noun 0 no man“ 0 v ca 09 uNON.O «@ maaa o n o” 00 uM4~.Q «@ umvn a N o“ >mo arm :cw: mwcou oz omtu on NI q M on o 1 N MN.~ M N h > mm¢go x mmcou .pmmhp 9. unmN.o do wnuu a a“ w 09 maaa o «9 menu a v a 90 mac“ 0 «a mea o M a o. moan e do Nana a N m >uo Ohm 26w: mmcdu oz cmzu 9m NI v n D« oI 1 N NM v M N b > mm¢4u x mmcau hmmp» 09 ummn o «a maaa 0 9“ w 09 mvflv.e «a un«« 9 v w 09 uwflN 0 Go uNon a n m 90 wmvu o «0 man“ a N o >wo ohm zcu: mncau oz omtw zm>35u7.o¢:w2um.anumnm Amamtu NHItzmv ¢~¢o auNudfiamoz no 904 20 hmwrah mood? S38 coo. 6975 .2. Human. 330 ma m mw34c> d >mnh¢h wusqz mu m ww:4¢> MI. muDJz> an n m m mm34¢> :2 02 oz 0: wu=¢u~u_:a~m No Mann & HaIunwn.0 81> umau mew ohm wu2¢u_u—:a_m at} muau mum Ohm wuzcu—u_za~m 1¢> mmou «kw ohm wuzco_u_z¢—m x¢> uuDu mxw chm wuzcu~m_=o—m no mean a Do uONn.» ma w~mm.9 «almemm a N9 umnn a «OINMMw o 90 maaa.a co ueNn a 11> uwau “um ohm 1w 9 v M an NI v N TN HI M m . h > mmtau x mmyau hmmhh 96 mnvN a 69 warm 0 0« ma 00 w«-.o a. mwmm @ 9 ma «cummaw.5 so Nana.G H mm as NHNN 9 so NVhw.o N mm Mmcgu 62 euro m N N mcau .hmwhh cu mm e um n nu N «N mmGJu D: amzu M N N h > mmcgu x mmcau >muhh dc wmwn.n we uvoa I cu ~w no unan.a so unnfl I 1 mm vs uwn« e as uvmm.I n «m «a mmmw c we mama I H mm mmcgu 0: onto >Ua Ohm 28w: NH on v n «a BI v N NVIQI M u hmuhh ww ww as mnnw Q «a uvnn.l m mu :0 uNwm a «o unmu I N wm >uc chm zxw: mmxdu oz omzu nN.a v n ew.II v N vm.NI M N p > mmtau % mmcnu nmuhh «a wmn« a Halmwvv I a« «w a& u&c~ a ma uwmn a 1 «m as uo««.o ac maaa & n «w 4% uwn«.e um wamm I N «u >mo o—m =cu: mmcqu oz omtu an mm34x> ma n n m mu34¢> mu n n M. ww=4¢> «H m n a mw34¢> m— n m m mu34=> o: 5: oz 0: oz uuzcu—m—:c~m ND mmam.fi no Mama 9 No mmnn e «qusnm.o *5 was“ a aaIuwNn.o Mo uumd a a: ”Man 0 ¢¢> muou mum chm muzcu—u—zu_m MO mmWN.a as mmo« 9 Mt} luau 1am 0pm wu:mu~m.:a—w 0Q maaa.o as unmm o ax) uwou mam chm 2 so wu:¢U—m—zc_m Na mmnu a daIummm a 3%) mwou a¢> quu Nw,wl 9 we «I v «u «I M w > mMCAQ JNH r x mmZJU .hmu~_ am no Nova 0 «a mvau 9 an a o I m m N n a a a u m N a o n v >uc arm n o v «0,, 1 9+.en M >mc ohm 61 a: v ?« 9| 0 mm o n h > mmcq >ma ohm o~.~ v nn 0 v M? NI n h > mmEJu 3.9.88 on N am mmcau o: cmlu tau hmmuh an an m an N am mmcgu a: aaxu ¢4u .nmupp on u v Nv M Nv Ah. mqmaoqm. 331 H >mnnm aw mw24¢> mu mu34¢> «a n n o mw34¢> mu m mw34C> o: a: 02 .6: l. oozzo_u_zo_m No own" o aoIooom.o No mom" o oo mom" o mo unma.o oo wqqa o mz> umoo zmu.oom mozzo_u_zo_m no uooo.o oo uaow o oz) uuou zoo ohm O. mozzo_n~zo~m oo mooo o «ququ.o no mooo o oo oven o oo uooo o «quaom o no ovoN o “quooo o E>n3o :3 Em ...? ...! oozzooaozozm M0 umNN 0 so wnw1,a No WNVN @ aw unnw a NO mwhn o no mmmw o 99 maaa a an unmn o m¢> mmou amu arm a... .90. muzcu~n~zo_m Go mama a .0 ummn 0 0a moon 0 Be mace o uc> umOu amu ohm av." M7 GI um NI h > an unmN a v M o N n m mmzoo z mmzzo omooo .o un«~,o oz and an moaw o . ans «o oNHm.o n zoo so ono~.o n oma zzwz mmzoo oz azzo v m I ~ n N «mono z mmono .ommoo oo oozn I oa was oz omna.I 7 ma“ .o o~ma I m moo go moon I m oaa :zuz mmzzo oz oozo . M z N n w mmzzo z mmzoo .omoo_ «qunan I on 7‘. «quamm I 7 .«a so maaa I n .oa «o-u~«m.o ~ vaa zzuz muzoo oz nzzo v n v u m ~ muzzu _ mmzoo .ommzo oo moan o oz ~aa a. mom. o . Non Ho ooaN o M Na“ oz moon I ~ nan zzoz umzou oz oazu v n v n n ~ macho x mmzno ,omooo so ooov I o“ ««a «o mono I I «so so onom o n ago .o-mmna.o z.o N zzw: mmcau oz emzu ma Hm24¢> mu m mm34¢> mu n m m mwa41> oz .02 .0: dz on. too! wuz¢u~l—20_m Na wvmm o «axwmmm o xt> Luau 1mm chm o » muzxu—u_:o_m no wwvm o No umnm a No unanc «almwnn «o wnwm a «aIumnn e as mowers as ummw o «t: mwau mum arm uu:¢u—u~zu_m no mnmN.O as uvnd 9 at) mmzu xmu ohm uUz¢u.m_:u~m 00 maaa D as wmmN.o .0 mac: 9 as uwuv.a 0o w996.$ Go wmmn a :5 None N ca ummn 9 mc> muau amu arm wuzco—k~2a—m ma unnh 9 Na wmwu o «0 mean a «stnom o No wnmn a a0 mama D on wane 9 QQ Mann 0 at) mon «my ohm h 3' >wa Ohm v n v N M N mchu x mmtJu hwmhh no umvu a Ga ca“ 96 uvma a v sun 60 mmnm & M Qua 9° unvw c N e«« 21w: Mmcqu cz nmxu v n 1 N n N mmcdu x mm¢4u .hmwpp on wwNv 9 ON KG— «0 Wang 0 O ne« «0 uuNu a M Nau 60 what I N ha# zxu: mmcdu Oz omzu Q n 7 N M N mmzau X mmxou .hmmpb 90 wvmm a a« man no mama a v nmu no mum" O N now 96 mama o N wad :zm: mm: 0 a: oazu 9 M v m H N mmCJQ x mmmJu hmuhb so wnwn I ma 7&« 09 mth I v va« on mean I M v0« «0 wmeu I N van 2¢u: mmcau oz ctra v M v N n N mmcau x mmtdu .umwhh ca umNN 0 ea mm ea man“ a v mm ac uwvnIo N an as uNmn I N no zcw: mmcgu :2 Gang 3.208 1:. mam/a. 332 an n n m Mung!) an n n m mu34¢> an n m m mw34¢> a« fl n m mu34¢> m“ n n m mm=4¢> a: a: a: c: Wuz¢u~u—2q_m ma wond.e as thn.o ND wwnNIG «alumna a NO mamN a «OIuwmm a Mo WWWM.O as mcvn a uc> mucu «um abm wuztunu—=a~m no “no“ .0 ca mama a v. munn.c a6 wmmn a No Numb a 90 mm?“ a No Wuqm 9 «@IwNnm D ¢¢> mwou mam arm muzcu_u_zu_m me uqu o 09 wwMN e Na mun" e fiGIWNQm a Me mvm~ a so NORM e u¢> muOu mum arm uu..¢u~:zd_m NO mama o «eIunvn o No uwnN 0 so unvu a NO wqu a as umaa a No wmmq a «aIummn N m¢> mmou «mm a—m o 0.. uUzCu-u—z9~m ca maaa 9 09 uamn a an was. a we Wham o Na wnnn.a 9e uNn« a we w¢¢° 9 9a quv.o kt> mmau amu ohm «a uno« a «a umnn 0 so wand 9 we NVNN 0 >mo chm vr OI m1 Q aw n h > no umMN o >mn chm 7 n v N M N mmcau x mmCJu ,hmwhh 9c mmnn a 9d vvn am wnMw o v 7'“ 69 umvm o M 77« a0 wmuN e N vva zzu: mmcau Oz aazu v M v N M N mmzqu x mmtdu ”—mupn Ga mmmn e a« «VA uquuna.a ? uvw as when 0 n «v« 00 uan o «v— zcw: M34u oz omxo 7 Mu v N n N mmCJQ m¢4u .pmmph be UNa' Ga «Ma «elunom a 7 and on mmmw a M amu ea unmn.o mm— zzwz mquu 0: onto 7 M v N M N mm¢4u x mmCJU humhh no unN« @ an de «9 unNa a v and «a mn««.e M wnu "O u«nfl a and 2cm: MMMCJU 0: onto v v N M N mmxgu : mquu ”hmuph .6 Nona I a“ nnu as wM«« I 7 was Do mflnn 9 M mmw do vaa. I N nm« zcm: mmzau Oz oazu nu mu34¢> Mu w m m mw:J¢> 335 mu mm24¢> mw34t> 0: oz 0: .oz .cz uu:¢u—u.zo—m NO m?1w a «OINWMN a ma u@7«.a no unv«.s ND maNM a «elmnmm e No waun.6 co wfinn a «a? mucu amu ohm on. wu=£u—m_zu—m Nu Mann 6 «Olwmaw a N9 uwnN e «quvmh u¢> umau saw arm .- v.9 wuzxu—u—za-m Me weww.a 9m UNQM o ¢c> muou uxu ohm uu:¢U—u_=o~m no MNMN 0 M9 wave 9 OD wmmw a N6 un~u a no wovm e ¢c> mung amu cum I do 0.}. Nazca—m_za—m 09 wage 0 go wavm 5 ea moan a 09 mmmn o ¢¢> Luau «mm 9pm av N 7 M Wm N v N VN »Q n N > wmcau Y mmcqu .hmwhw an» u~.9M .o as wann a an med 0o umNn .0 on mvnN a v «Na an uewm o co uMNm a m an“ Q0 wm¢m Q on mwa e N mum >u¢ a—m xcuz mmzAu c: omzu No.“ 1 M ea c 9 N mu MI M N » mmcau X mm¢4u ~mwhh so wmnn a «a aqua a an haw co NBM1 Q as uNmm a 1 NNN ea unmN a no uwv« a n nu“ co whnN a «a Mama 6 N sud >wa chm :zw: mmcdu oz amxu Mn « v M «m NI v N Hm MI M N p > mmcuu x mmcdu .mwh» «a wvmu o co wnsN & Na Us on mmwn a 05 manw 0 9 «Nu co mwNM O «0 menu 0 M Nu «m mea 9 ea unmn I N Na >uo chm zcwz mmczu a: euro MN « v M an 1 N Nv ~I m N h > mmxau x mm¢4u hmwhh ea Whmw a we qu 0‘ wmwu a q vNu cc whom a M 7N“ on wMaM @ N qua >wo chm =¢wz uMCJu oz QLIu ea N 1 M in NI v N nv nI n N h > mm¢49 x mm¢4u —mw—— «a umvw & «a mnnu I n« mNa «a menu a do umM— I v MN» «a mafia & co uawm a m MN“ 06 moon 9 as quN I MN“ >wo ohm =¢uz mmcau oz omtu A6288 .2. "Hugh. 333 nubdc ma n m m mw34¢> MIDU‘OI meJI a. 0‘81W'h mu34c> an n n m mwagc> q >4hv‘n~n u‘ o: .02 92 Oz 02 wuz¢0~m~2o_m we waoo.o go unum 9 7o m«mu.@ 90 wan a co wooo.c 0a whom a 09 wage a so mmn7 0 at) «moo mam ohm 0'0. 9.. «99¢ muz¢u~u_zo_m Na mMnm.o so umuq 0 N9 wsn«.¢ «QIwNnm.a No maau.o so wwm«,0 Nm when 0 «ovmmmfl.o at} muou «mu ohm .... uu:¢u~m_zo~m Mo umum.o no unwm o 90 uogo.o co unnn.o no wmmN.o co mmun 0 n9 umwn a co mama a at) muou «mu o—m wuzcunuuzo_m no wna7.o ca noun o 75 umaN.o co wnmn.e Me unam.o co u7mw o no wmwm 9 am wwMM a at) umou «mm o—m wuz¢u_u_zo_m so memo.o oa momm o no ueoo.o co unmm 9 Mo wo~N.O 90 um~u o oo waoo.o go wwnn 0 ac) uwcu amu ohm m7 OI wN OI DM 0 h «a uvMM. «@ wu7a «a Ufinq. «o wm~«. >wa Ohm M7.7 MnINI mm uI — > as uNM7 0 co wnmq.o so ummM a co wwda 9 }mo chm Mm d Mn.N wMIa n no unnd «a w7mu no mama. «a u7na >mo ohm an O 67 o no @I p > «a maaa a «o wmnu a o O @0603)- 06°01'1- «o wmma do maaa >wc arm MoIo 7w 0: Nw,«l — no wand do w~m« do uNwM. do NHWN >mo chm @006)- 7 M 7 N M N mmcqu x mmtdu hmuh— cm ummu I an flaw «aImmNm a 7 mm" on uan I M mad as wm««.I N haw amu: mmcdu :2 auto 7 M 7 N M N mmcqu x mm¢4u .hmmpp no man“ 0 0« 7o« «0 wNNq a 7 7nd «0 wch a M 7m« «a wmo«.@ N 7w“ zzmt mmcqo oz onto 7 M 7 N M N mmcdu x mmcqu ,bmuhh $9 whmu Q on Ma” «9 unqu,u 7 mm" 90 ummn a M Man as w7mm o N Mod zcut mMCJQ oz amxu 7 M 7 N M N mm¢au x mm¢4u pmup~ ac N7WN.0 on mm" «a N7Nm o 7 mm“ a: man o M um“ 00 moon 0 N mu zcw: mmcau a: euro 7 M 7 N M N mmtau x mmcau hmu_p 0e wean I o« mud 00 momN I 7 m~« as mmom O M mun ac m77m I N m « zcuz mmcgu oz Gaza Muzzm mNDJt me4I 0a m m VII. > 02 mM n m m > 02 an m m > 02 m« w n m x» nv._ mu 0 m m .> o: uuzcw—u_zo_m ND wumm o «Oqumm o No wnnm.o o0 wand 0 no man? 0 00 wMMn a No m7nw G 00 uuafi 6 mt} uwou mam ohm umzxu—m-za_m co maaa 9 co wmma a mo wooo a 90 mnmv 0 NO mam» O uon¢mm a 70 umdu o no umOm o at) Luau amu ohm 9 wuzru_u~zw—m mo wmm« a co uno7 9 mm NWON.Q Go uvmn 0 Ne wmm«_o 00 wnhn a M0 maMm D Go unww s m¢> umOu «aw Ohm u92¢u_u_zo—m e0 mace a as uOoT a go mama o co uon o no umwm a 93 wmem o @0 weao,o 90 Numb 0 at) uwou amu arm 90! wu:¢u—u_zo—m No Nana o «QImNNn.Q Na w@7«.9 uaIumua o No maaa a «QIw7-.o No NMMA o «OIwnaw e «c} awe; mum ohm no a an 5 mm QI » > as uwm~ o as mama a co mom~ a co w70m o >mo ohm nn.« nu « an OI » > oo wwmm o .9 mac“ 6 ca mama o no um~¢.o >ma o»m nnIn an.oI um «I h no one so wm~« a9 momm ac moon swa ohm mm a ma 9 NM AI ~ do ww~u «a uno— «o mama «o wuun. :ua ohm m7 « $7 «I 07 m: p ea mmmu oo um«~. oo wnua no mnmu >mo ohm 9990? 9°90)- 009°)- 7 M 7 N H N mchU x mmcdu .mw_h ca muow 5 0H 7&— oo N777 o 7 7ma 00 wnnw O N 7nd 0° unwv o N vnu zcw: mmcdu 0: auto 7 M 7 n M N mchu x mwcau .mwp— ucIuNnn I ma Nun oo u7wn I 7 Nu— OG whwn o M Nnn «Olmwfl7 o N Nua 2cm: mmtdu a: 91:9 7 n 1 N M N mchu x aficzu hmuh» Go uMmm 9 ea mud 99 mama o 7 mwa «& mmen o M mud Ow wmmn o N mun tam: mmzau oz 9;:u 7 M 7 N M N mmCAU x mmIJQ hmwbh Go mmNm I on man as uMwm I 7 nod 69 umo7 o M nm« 90 umem I N on“ 21w: mmCJU oz Omru 7 M 7 N M N mmcau r mmcqu .hmwbp «0 man” 0 ea Nun «a uvnn O 7 NW“ «9 uN~«.o M Nnu «9 umMM o N and zcuz mmcau 02 also AaLzoov .2. Waugh. 334 I. ~q ~ 9 n on o q n no.o v o _ on “I v u .... me n- M N .2. as m N mozzo_zjzo~m o z mmzoo z mmzoo ,ommz» mozzo_o_zo~m h z mmzoo z mmzoo ,omooo mo mo moon o oo m~z«.o oz mmzo o «o “no“ o oo mom mo no mozz.o oo mono o no u¢m«.o «o mod“ o oo ooo o No mono o oo uzoz.o oo moan o «o mono o z oom o No moon o oo mono o oo mmon o «o memo o v zoo m an mama.» oz who" o oo mquIo do mnm~.o m mom m «o mono.o zo ummo.o oo moom.o «o u«m«.o n moo a ~o moIo o «quwo~.o oo umoo_o «o mas“ o m mom m mo mmmm o oo on?» o «o mzmq o oo uuon.o u moo muoozz oz «co zuoo zoo ohm zoo opm zzmz mmzoo oz oozo mmooz> oz oz: uuoo zoo ohm >uo ohm zoo: mmzoo oz oozo moIo z n no a z n no o v u om o- z w an ...I M N N7 «I M N mozzu~z_zo~m » z mmzoo x mmzoo ommhp uoz¢o_u_zo_m p z mmmou x mmzoo omooo «a No uoom.o o. uqm~.o «o moan o «o mama o o« no“ mo mo unzo,o oo w~zm o «o moon o «o umo«.o o« mm“ n no unmo o oo mn~o_o no moon 9 oo uqoo o 1 non n no momo.o o. umom o «o u«o«.o oo uzoo o z zoo n mo mono o oo uno«.o oo mon~.o «o wood o m now n no u~nq o oo moo” o oo uno~.o no uno«.o o mm. o no mooo.o oo won... «o “no“ o «o moon o N mz~ o no wovn.o oo moon.o «o mood o oo unow.o N no“ muoozo oz zzz zmoo «am ohm zuo ohm zcuz mmzzo oz oozo mmooz> .02 «oz ouou «mu ohm zmo ohm zzuz mmzoo oz oozo o~.o v n I on.“ 9 n we «I v N on «I v m on_oI m a I oo «I n ~ mozzo.z.zo_m o z mmzoo z mmzoo pomp» uozzo_o_zo_m _ z mmzoo z mmzoo omuoo mo no moon o oo wwwu e do ova“ o oo won“ o o“ mom ma zo woos o oo moon o «o mood o «qumno.o oz Too o no uwoz.o oo woz«.o oo moom.o oo mono o z Now n no who" o oo uzo~.o oo umzz o oo mooz o 7 vow n a. momv.o oo oozq.o oo moon o oo whoo.o M Nam n «o moon o oquzno.o on unmo.o oo mono o M 7oz o oo mooo o oo muon.o do mom~.o oo moon I u ~«u o oo moo. o oo on»? o «o umman oo uovz I w you muoozo Oz «oz zuoo zzw com zmo ohm zcmz Imzoo oz oozo muozoz oz zz> zuoo «om opm zmo ohm zzmz mmzoo oz oozo .. o~.~I z m zo.« q n o~.oI v ~ mo.« 7 a on oI m u no o- n N mozzo_o.zo_m o z mmzoo x mmzoo .pmmop mozzo_zozo_m o z mmzoo x mmzoo .hmuho oz No when o oo mama o oo uvno.o oz ”zoo o oz “om mo oo mooo.o oo mowm.o oo won“ 9 oo umoo I o“ om" m «o moon o ozIwono o oo mzqo.o no uzuo o q no» m oo wooo o co moon o «o mooo.o do mono I 1 mod n mo unn~,o o. woo“ o oo unmm o oo uono o m Now m oo mooo.o oo mama o «o mooo_o oo mama I H mm“ m mo “no... go moo. o oo un~q_o oo movo o N moo m oo mooo o oo mnmn.o «o mono o oo moom I a zoo mmoozz oz zzz zuoo zoo o_m ouo ohm zzmz mIcoo oz oozo wooozz oz zoo zuoo ozo ohm >uo ohm zoo: mmzoo oz oozo oo.. 7 n o« o z M Na .0 7 N ...... «m n 7 N «a oI n N . do,“ m u mozzo.ozzo_m — z mmzzo z mmzoo homoo mozzo_z~zo_m p z mmzzo z mmzoo .omoop oo oo mooo o o. moon 9 «o momo.o «oImmmm I oo oom on oo wooo.o oo woom.o «o moo«.o oo unvu.I oo oo« o oo mooo o o. ozoo o «o u~oq o oo u~an.- z oou m oo wooo o oo moon.o «o mono.o «o moe«.I 7 mod n no moon 9 oo um~o,o «o u~m«.o oo moon.o m oo~ o oo wooo.o oo mono o «o moo«.o oo moon.I M zoo o oo mooo,o no moon 9 do mon«.z so umom.- m ooo m mo mono o «oImooo.o oo moon o oo moon.o a zoo mmoozz oz azo zuoo zoo ohm >uo ohm zzuz mmzzo oz oozo muzozo oz zoo ouoo zoo ohm zoo ohm zzuz -mzoo oz oozo @388 d. 339 335 on o v M v n .. m: N v w on o o u I no a o ~ mm o- m ~ wozzo_uozo_m » o mmcoo x mmzoo .hmupo oozoo_o.zo_m o z mmczo x mmzoo Poo». on oo moo. o co anon o «o won«.o oo moon.. on com o. no whom o oo omow o «o on“ .o «o mono o oo now n on moo... oo moon 0 «o ouno.o a. u«o«.I v onu o mo oomo.o «o umoo o «o uzmo o «o moon o v on“ o oo uooo.o co movo.o qo umzo o «o uo~«.I n own o no mood o oo moo“ o oo moon o «o wzom o m Now a mo woom o oo woo. o «o umv« o as man“ o a own a mo mono o oo mnoo.o «o moo~.o «o moon o o hum 33$, .2. ucz uuoo «am ohm zuo o3. :zu: mmzoo oz 9:: L3?» .oz of. zoo“. amu 3m 33 3m zzwz mmooo oz 9:: cm." w n on H o m on o v a I. no.~ z N on «I n o ~o o m ~ uozzo.o~zo_m p o mmzoo z mmzoo .omuho oozoo.o_zo_m p o mmzoo z mmooo poop» ha 90 maaa o a. mavn ,o «o mmMN 1° 09 muvw I an mnN 1.4 co maaa @ ea unwm @ «9 mm«« .0 e9 ummv I on nmm n co mooo.o co moon o a. ”no” a «o moon I 1 Now o oo uooo.o oo moon o no om~«.o «o won" I v on n mo moon 0 «o uo~o o «o uvou.o oo w~o~.o m ooN o oo uooo o oo oomz.o oo mono o oo mono I M nmu o oo oooo.o oo mono.o do umo~.o oo mono I m own o oo mooo_o oo omom o oo oo~m o «oIoooo I u now mooooz oz no, ouoo zoo ohm zoo o—m zzmz mmzoo oz oozo mooozo oz zoo oooo moo ohm zoo ohm zoo: mmooo oz oozo ...... om MN 7 M Mm» o 7 M .. nu_~ v ~ ~o o- z N u... «M MI M N «a on M N mozzo_o~zo_m p > -mooo x mmzoo .hmoph oozzo_o_zo_m » ; mmzoo . mmooo oomp— mo oo wooo.. oo uvqn.o «a mono o oo unvo.I o“ «nN o4 no oomo o oo moan o oo momo o «o oozq o o" no” n o. wooo.o so mono o oo mooo o «o menu I v «no u do moo» o «oIoozm o oo undo o «o oooo o 1 wow n «o uoom,o oo uzzz.o oo onn~.o o. uoqo o n ”on o «o mono o ooIoonu o oo moon o «o moo“ o H on o oo wooo o oo omn..o «o u~n«,o «o moon I u «on o no woo«.o oo momz o «o u~M«.o ”o moon o u n~u mmoozo oz zoo zooo zoo ohm >uo ohm zzoz mmzoo oz oozo mooooo oz oz) owoo zoo ohm ooo ohm zzoz mmzoo oz oozo on.o v M on.“ e M o No a. z N I: 3 z z a ~m.o n m .. on o n ~ oozzo_uozo_m — o mmzoo x mmzoo omooo oozoo_o_zo_m _ z mmzzo y mmzoo pooh» no no uzvz o oo umnw o «0 m««« o oo uon~.o on own mo oo who“ o oo woo” o «o mom" o ooIovvo o o. an n oo mooo o oo mono o a. mom" 9 oo mamM.I 7 wow o oo oooo o oo mono o _o mono o «o mom" - o moo n yo anon o oo oozm o «o wnv«.o oo on~«.o m oz“ o oo wooo o oo moon o «o oo~o o oo wood I m m m o no ovuq.o oo onou.o oo ammo o oo umqo o ~ zoo o mo moon o oo ooh“ o oo ummn.o oo onom o a mm 335.. oz an; ...uoo amu 3m zoo 3m 23: mood oz 8.8 3.33.. oz ”...; boo ozu 3m zoo ohm zzuz mood oz 9.5 .g.. oz n- z n o~ o- v M I. o..~I ' ~ am a w m mo.«I n a mm o n n mozco.o_zo_m » > mmoou z mmzoo .pmmpo oozzo_o_zo_m p > mmcoo z mmzoo oomph 3 no ummn o 09 mm: o co 33.4.. 3 33.6 3 I..." 3 oo 25... o 2. mm: o .3 mm: o «cImmmn I on on“ n «o u~o«.o «quoooI oo uvoo o no moon o v 'vo o oo uooo o oo mono o H... own .9 oo mnoo I v oo~ o ~o uz~«.o «oIoooo o oo moon o «o wood o n zoo m oo oooo o oo mono o «o manor. oo wooo.I M own o no moo. o oo uoo~ o oo uumo o oo undo o a zoo o mo oum~ o oo moon o «9 ovon o oo more o u moo mmoozo oz ozz ouoo mow ohm zoo opm zzoz mmzoo oz oozo wooooo oz zoo oooo zoo ohm >wo opm zoo: mwooo oz oozo oozoooov .2. 539 336 ma n m m mu341> mm341> Mu mw34¢> mw34¢> oz .02 .02 Oz uoz¢o_o_zo_m «o u«««.o «quooo.o mo oum«_o oo mooo.o do omvn.o oquooo.o mo mou«.o «quooo o m¢> umou «um ohm uozzo~u_zo_m no wuoqu oo mono o as uaao,o so moon 0 no unum.o co muno o no wood 9 oo mono.o ac» uwou «no ohm Q. 7.79 mozzooo.zo~m mo u~m«.o oo mmoa o «o uvvw.o oo woo“ o oo uooo o oo wonv.o No moon.o «oImmom o «co uwoo amu ohm uozzo_o_zo.m Mo w~«~.o oo mo- o ~o mood o «oIooov.o no un««.o «.Iuono.o zo wooo.o o. mooz.o zzo ouoo zzm ohm O O uozco_u_zo_m no owno o oo u~««.o mo moom.o oo wov«.o No mnwn.o co uneu o no unmo.o oo wv~«.o ¢c> ouoo «ow ohm Nm @I ON a m«.« h 99 u7oN om W77N. MOImnam co uwMN. >uo ahm ne.« Mo.“ mm OI b 3 a9 uo7h.o we mama o as m«7M.Q 09 wnw7 o >uo ohm nn.MI nn.«I meM h 00 wnmh. oo w7c7I «a maaa. co wncn. >wc 0pm hm oI Nn «I n7 «I h 99 unom. on who“. «@qumm. «a uNMw. >wa ohm no.N N0.N mo OI p a9 mmm7. De umnn. so u7n7. no wNmM >uc chm 0909)- 5436301> o'scam:> 64919at> P1V' mmcqu no u7mn do mnmu no unmu «a maaa. zen: mmCJU Q9 uu7M. «OImmNM on wwhnz co wwmw, sew: 7 7 M mmCJU no w7Mn Ga wnwm 2cm: 7 7 M mmCJU no wnm7 9a mama, a: wunmI «OINN7u. zcw: 7 7 M mmCJU co wNmn no maaa oo NNQQ. 09 Nana, zcwt NNM x mutau OH 7 M N 0666 hmwhh NoM NeM Non NoM MMCJU 02 onto NNH x MMCAU am 7 M N ”pmmbh «QM «QM «0M «QM mmcau oz omtu NNM x WMCJU ca unwv. oo mmwm. x mmcuu “hmmhh man QMN amu mam I40 02 omtu .wmmp» me mMN mmN mam MWCJu Oz omru NNM x mmcqu on 7 M N 0000 ”hmmhh 7mm 7MN 7MN 7MN mmcau 0: auto J} 05'5”?) III wm34¢> IO: oz n m o moooz oz on o ooooz .oz oo o o o mm34I> O: a >0IDIDO‘ magi O: uuzcu_u~zonm ea mac: 9 ea wmom.o 7° w«N«.o mo mNMw.o so NQQ@.Q ea mNmm a mo waao.e Q9 ~977.o ¢¢> umou mum cum in motto-u—zo~m Go maaa o No uM7M 9 an ana o oo uooo a 00 memo a co much a 09 maaa a Go m97n a at) umcu amu ohm 900. o... .. uuzcuum—zanm we maQM.e a: uw7u o No mNMN.Q am mmma w «o maM7.Q dmemH7 a we ummd 9 on mwM«.o at) umao «aw 0pm 46 4 wugcu_u~zo~m Mo wme o oo w—MN.O we wawm 9 «@Iummn.o 7e wom«.c OQ wwmn a 79 moun_o 09 mmwv o u¢> muou mum ohm coo wu=¢o_u~zo_m 79 mMuN o co mmm7.o ow Nome a on womm Q NO u«m«.@ so wmu~,o 7o un7u o ea man» a ¢¢> umau cam chm .hmupp MmN Man MMN MmN mmcgu 02 Oman .hmmhh «mu «MN «MN «MN tau Oz omzu .hmmhh amu emu omN amu x40 02 omzu .pmmop ..u fo- NMN NmN NNN mmcuu Oz omzu 77 SI 7 w on oI 7 N N~ a m N h > meJU x mmIJU am uNMn o chmnmw I @« «a mn7n a on mud“ n 7 «a won«.o on mmmw I M «a uunn.o «qumma I N >uo ohm 21w: M~.« 7 M mM.N 7 N Nm 9 M N h r mmcdu x mmtJU «a wm7a e «a mowq I co co maaa a «a WQMN I 7 «o mama a «a wnnu I M «a mNmn o o: mmww I N >mo Ohm ztmt mm nm n 7 M «a m 7 ~ MNIN M N h 7 mmcau x mucou co ummm o no qu~ @ on no Noun 0 we mnwn o q «almmrm e «a NMAN o m 90 umov a «a umnu o N >wo ohm 2cm: mm MM NI 7 M ma NI 7 N 99.9 M w p > mmcou x mwtou «a umNn a ma uw77 a go as wzma a «a mama o v «a mama Q Hquov~ o M «o ua7q a 09 whnu a N >uo ohm zcwc am 7 v m an a 7 N N7.«I n u p > mmcdu x mmcqo no mnaN o «QIw7Mm.e on «o mmmu @ «a mum« I 7 no wnmm.o no uku a m «o wMMN a 90 med a N >wa chm 2cm: mm 3:58 .hmw—h mWN mwu wwN mwN U 02 omzq 0:. magma. 337 mw34¢> wu34¢> an n n a mu341> mu m n m mw34¢> m« n n m mwaax> ,9: av 62 .77. a... w42¢u_u~za_m Na mama‘s as UNon 0 Me unMM o a. wMNN a No uwnn o 00 uuuw e «o u«no a aaIwmnN a ¢¢> maaa new ohm mu:¢u~u~zn~m Mo mm~«,o a. wahN.o No wmmN a so wnva 9 Mo W7mnvc a. mmmw e no u7- a DO U7M7.G Ni) Lwou amu chm wo:¢u—u~z:—m m9 uam« a as umuN a no unnh a on uwmw 0 N5 ummn a «aluwmm 0 Na wNnm a OS wa1~ a u¢> umOQ amu cpm muzcu—m_za_m No u7nM.e «aIW7mm a Na u~«n.. a: NW7« 9 No ua7q a «OIwn7~.Q No unu7 e on MONN a u¢> uwau xmw cum 0. muz¢u_m—=O_m Na uNnn.o Nalwann a NO umnv o 0e mum“ a No man« a «alu7ns 0 N9 uunm a as weed a ax> mwcu «an arm 7n 6 nN 7 «$.71 h > «a uNMqu 9'!“ (50¢ uluu a NIH m N u Nfdw one >wa ohm Wk.§ 7 M 7 N M N mmcgu x mmc4u hmwhn «a mum“ a on can an wma7.s 7 can oa uamn e M can «a uflwN O can xxw: mMGJu oz omzo 7 M 7 N n N mutau x mmxau “humph 06 “can 0 ON man «a maaa a 7 mum 09 www7 o n mum «a wnOH a N Ann zcw: mm¢4u Oz omtu 7 M 7 N M N mmzqu x mquu ,pmuhp De wNnn a mu 7am ea uocN a 7 van as wemn c M van a: w««« o N 7am 2cm: mmcau Oz emzu 7 M 7 N n N mmxuu x mmzdu .wmuph «a w~u«.e on Man «0 una« o 7 man «a man“ a M MNM «a mmNa a N Mun z¢wt mmxgu oz amzu 7 m 7 N n N mmIJU x mm¢40 ”pmmhh no mama 6 an sun on umNm o 7 can no mwNu a M oum we wwna a N Sun :th mm¢49 a: 63:0 mu m n m mw34¢> m mu34¢> « h rIw m mwaa¢> > main m a v m a 4 I mu34¢> a: a: 0: Oz wuz¢u~u—z:—m Ne WONN n dolwnmn.o No uMON a no wmmn a «a momw a navaww o u¢> uwou uxw ohm wu=¢u_u_za_m No mNon O a. mNnN a Mo unfin . a0 mw7m e «6 “ham 6 aoImwmm : as unnN a «almuam o ~¢> umcu mum cum wu=¢u_u_zc—m N9 maaa m «alun«7 a «o unwe e «aIuaNm a «a up?” a uQ-manm a N9 uwna Q uqun7n a u¢> umOu «aw a.m wuz¢u~u_::_. no wmau 9 ea uvau 0 Ne unnq 9 99 w7oN O No mNnN so wvmq @ m9 wnwa & as whN7 a at) Luau mum ohm um:¢u—u~za_m ma mama 9 «qummw a NO wqu a 96 u7nu a N9 um=« a «aqu7n a Na umna a «a umNo e n¢> Luau new ohm mm o M LN a N 7N.aI N » mchu hmwpp so unvu ed as qu7 a 7 06 wnmu a m a9 “Nau a >wn oh mwxgu a: amzu n« a m Mn N N MM «I N p x mmcdu «a ma«« 0 ea «9 wuufl. 7 99 uwnu a n Go “amu a N >uo Ohm mmcqu oz emZU MM u M kn.o N mm OI N p r ngu 96 wNmu a an an mama a 7 me mm7w a M 99 vaN a N I.. >ua oh mMcgu a: omzu «7 9. n «7 QI N nu eI N u MIJQ 95 NOMQ 5 ON 9: uun7 o 7 91 mmnN o m an umN« o >mo arm mm14u Oz Otto «n oI M 7n « N nn.N N _ mxau on uwon 6 ON 00 wmmn .9 7 99 www« o H CO umvN e N r >wa Ann mmc49 Dz omzu 33258 .2. mqmflfi 338 Mu n n a mm34¢> m4 mu34¢> M« b n m mw:4¢> mw34¢> 92 .Oz 02 O: D: 7 uuzcu~u_::_m co ufioo .o a. uuw7 .o u¢> muou raw bum uuzxo—u_za—m m¢> mwau amu Ohm wuzcu—u—z0_w No uN MN 9 we: Iummo a u¢> umOu mum ohm no .7 uuZtu—m—zu_m co maaa o co mNNM e ~¢> uwau mam ahm uuz:u_u—za_m no UNOM a «quam7.o «. wMMh.o «clubhn o we umuM a «olmnnn,a No mnnn o «OIwam a at? Luau saw chm 7 M 7 N M N mmCJU x mmCAU ”Mmuhh MoIm7a« I a« .7MM on wMMm .0 7 7MM an wamm .a m 7MM a: mama I 7MM :cw: mWCJU 0: auto 7 M 7 N M N mmcdu x mmcau ”thM— we wNNM @ ea MMM an unmu I 7 MMM «o w~««.o m MMM on unnu a MMM zen: mwcau 0: onto 7 M 7 N M N mmtqu x mmcqu .MMWMM «a ”and 0 ea MNM no mmuu o 7 mNM «9 W57" 5 M MNM «a mnnu o nNM amu: mmcdu oz omsu 7 M 7 N M N mm¢4u r mmcqu hump» «a mnou I On mNM no UNmu I 7 @NM «a man“ I M mNM so un«« I N QNM :Cw: thJU 0: gang 7 M 7 N M N mmcau x mmcqu Mmmhh «a uNMN a on mNM «a UNMN a 7 wNM «D m«7N o M wNM «a M7NN o N wNM zcw: ImtJJ Oz omzu ) m n n m a mm: 4... ma n m m mwaqt> > m n mIr A m u 2 4 I a: o: .0: muz£u_m_za—m a. wm«m.6 «QIunN7 0 Ho mans a «Glamor a at) mwou can o—m wu=¢u—u_zo—m Ma mn7d a 00 MN7M o ac> uuou «mm chm out. 7.0. uchu—u_=g—m De ween 5 aa mmnm a c1¢> muou «aw ohm wu=¢u~u_zu_m Ma uumu 0 0° uNmN @ Ma “NM7 a as uvmm o ¢¢> kmau umw arm wuzxu—u—zu—m “N uMmm - 5° waMM o a¢> mwou 8mm n—w P '> 7 M 7 N M N mmcdu x mm¢4u Mmuhh no wwNN a an «a N7MN a 7 «a wMNu a M «a u7NN a u zxw: mmEJU oz alto 7 M 7 N M N mmxAu A mmtdu thM_ «a “Med 0 cu 7NM ca u7Nm o 7 7NM do mnna a M 7NM no wmmm a N 7Nu zzu: mmcqu 0: onto 7 M 7 N H N mmcau x mmcgu pmuhp 00 uNwM I NM «9 Mama I 7 NM no mama I M Nu as wwuu a ~N1 ztw: mm N¢JU 62 onto 7 M 7 N M N umzau x mmgau pmw—h 05 mm n a on aNM a: umnM o 7 «NM 60 wmnm I M «NM 90 umwn 6 «NM 2cm: MMIJU O: at:u 7 M 7 N M N MMCJU x mm¢4u hnwhn «o uwMu o on anM «a mafia a 7 muM «a umna a M Man «9 uvqu o N mam 21w: m :40 Oz omxu $388 .2. qu<8 339 m« n w m mm341> « > musn «a m Agagt> .0 o: o: No UmNM a «alumna e co mace a a. uno7 e at} kmau can ohm 9. uuz:u~u—::_m No umM« o «QIMWM7 0 Ne w7MN a 99 WNW« a do wmnM a «.INNWN O «a w«mm a «Oumamw a 31> uuou mam Ohm uuz¢u_u_zc—m N5 umhm a as wmwu a No uhmM a an mmwwIa Na mMnM a MQINNMM 7 M9 mam“ 0 Na uNnM m K¢> uwau «mm chm wuz¢u_u—zo_m no umao s cc u0M7 - as wo®a,o on Mama 6 ac mace 0 ca mem o co maaa 5 as mnam 0 at, uwou 8am chm mu2¢u~u~=a—m Go uwoo a no uQnM 0 &9 mac. n as uwa 0 7e mdNu a o: unaw s cc meme 9 as umwm a fit) muou amu arm > 99 mmmm u do uNNv a >ua ohm 7 M 7 N M N mmcdu x mmzau bmwhh we wNMM. I an NWM as menu a 7 man «0 uan,o M nnM «a man“ I N NWM 2cm: mm¢49 o: omto 7 M 7 N M N mmzau x mMCJU pMmph «a wwnn.o OM MnM «a wow—.a 7 MMM «a wmwu o M MWM «a wa7« e N Zam: mmzqu 02M mmzu 7 M 7 N M N mchU x mmxdu wwwhb 9o mmmm a flu NmM «a wNuu a 7 NmM «a mnmqu M NnM a: wmam o N NnM :cw: mszu o: omzu 7 M 7 N M N mmZJu x mmcau .hmmMM so uNnn I on HnM «eIunmm.I 7 «nM «aIm7NM I M HnM «a maaa I «WM zcw: mucAU Dz omzu 7 M 7 N M N mmcdu x mMEJU hmwwh co unmn unM so wuwm I 7 anM as m7n« w M cnm “OI mNMM. I N QWM zcmt mean. 6: amtu m« n n m mw=4¢> MN m mw341> mu m w m mm34c> ma mw34c> mw34c> .92 o: a: D: a: wuz¢u_m_:D—I M9 mmvv 9 0a uumu a M. wman co umNh 9 Ne wmvm a «aIuaMm 6 aa maaa a on unow N am, uwcu «an ohm muzcu_u_z:~m Ia mw7q a co wNnN e no «man O Malwvmn 9 Ma mNMM a no mNaN a u¢> umcu amu awn wuzcu_u—zo_m N9 quq a an wnwn a Na uMm7 a as wvnN a Na m«nN G ac m7eM o NO wmw7 a no wwmd a m¢> gnaw «mu chm .70 co. muzcu_u~zo_m co u©&9 6 ac w«nM a Na muNN o Moluw7n a N9 mmOv @ am u77~ a as maaa a so un77 o 31> mwou amu ohm muzcu_u_23_m a: maaa a a: mba7 e M. W7NM a ma MHMM @ ea mesa @ 9e mmww 9 no maaa a s9 muMm a 31> mwou mam ohm 7 M 7 N M N mmzau x mmzuu pmuhh as wm7N o m7M as wnvn a 7“ 17M ea umMm a M m7M no NM7N I va 1cm: mmcgu 0: also 7 M 7 N I N nmzq; Mutqu hmwup as man» 0 ea N7M «a uwu« a 7 N7M Q0 uwcm a M N7M ea wM7h a N h M zcm: Mmigo Dz omtu 7 M 7 N N mmzau y mmcgu hmu: «a chq a an va «a man“ a 7 ”7M 05 m7~m a M w7m «0 uMMM a w7M :cfl. mmfu oz amzu 7 M 7 N M N mMCJU x mmrJu hmuhp =0 mmwm I an M7M Ga wnnN a 7 M7M as uaom a M M7M «a umwa I M MVM zzu: mm .9 02 onto 7 M 1 N M N mngu x mmzqu me»» no mean I an N7M 91 mum? s 7 N7M no mnma I M N7M «a mnNM I N N7M zzu: um Au 0: amtu A0388 .5”. Hugh. fi_‘ 340 , n 3 o v n . ~ I o- v N n a 8 o- n N 3:31.25“ mug”. x “med Eu: 353222“ b mmcd x mucnu 9mm: 2 no mm: a 3 mm: a 3 man... a 3m 2 as was» a a... N3... 9 3 fl; .9 S 33 a n3 n Na :3 a so 2.: o 3 #3 o v in n 3 an: . SIMS“ 9 2. mu: .. 3 ma: ... .. SM n 3 mm: a 3-33. m 2. 33 .o n in n 3 ”:5 a 3.5m. . :TuZm . 3 :2 a n ...: a 3 a“: o 3 ml: . so #3 .o In m. 3 33 o .3 “~55 3 ”mad 3 we: a non 33E. 9. ...; “so“. .5; Eu :5: mmcd oz 9.5 $3”; .2. 5:. .33 «an 2m >3 3m :3: mmzd 2. 9:5 7 a nu .T . n v m 3 .T . ~ n a 3 .a n N 355.55; 3...: x mmcfi. u5m: 3:51;...3 . > “mad 2 $5“. #3: 2 me man. a a. 3..“ a 2. at; o 3 2m 3 3 my.” . 2. unm~ a 3 mm: a 2. unnm o 3 3m n ~a “1: ¢ 3-3.2 a 3 mm: o 7 SM m up. mm: 9 firm: a 2. $1.. 3 u..." .. 7 5m ... 3 m2... . :1qu o 3 :2 a m 2n n M. “at .u 2. 33 a 3 “a: a on gun a m cum a no unnm... ..- mGn a 2. mm... . ~ u: a no mun“ o 2. 23 .o 3 m3“ ... 3 ”can... In 33.; 9. fit. .38 «um Em Ea: mmcd 2. 9:3 335, .9. «:5 $3 mum Em >2. em :5: mwzd oz 9:5 a... v M an o 7 n q . ~ , S o1 v n .... n ~ 3 o- m ~ $53.25; “mad x emmcd .5”: 353:.sz h _., mmcd x $5“. am: a" no wk. a a. 32 o co um: I in 2 no w~m~ o 2. mmmn a 2. we»; a as “a: a a. . Sn n ma :2 a «Twin a S “I" o .. {M n 2. wnmu 2. mm; .a 3 9.: a 2. “3.. a v G” n on we: a 3 mm?» .. 2. mm: - m In ... m. m3" a 3. mm; a E wwma . 3 mu: m n 1.3 a 2 £3 a Sign a E. $3 e w EM a 3 mo: .. 2. “in a 3 an: a 2. 3% 9 ~ SM 335. o: ...; b3 mam Em 25.. «mfg oz 2:... $3,; oz 3... $3 «an Em 3.. Eu :5: 2:: oz 9.5 5. _. n S a v m z _. a mm .o- v N . .. ~ I a- m u wuzcfltzoa mmcd x mag: .53» “2.31:3: » > 33... x muxd 5m: 3 use. a 2. man" a 2. m2" - 3 3m 3 mm mm: a 2. :3 a 3 um: .a o. “Cm a 3 can a ”a “$2. a es mg... . 3 we: I . 3n n ma unwu a 2. 3K . 3 new. a 2. an... a v ......” m 3 Us»... a 3.3: a 2. 32 .o n 3.” n 2 mama a 9. gm a 3 mm: s 2. #3 e m an m 3 EB. ,. 2. “3.. s so mam” - ~ mm" a 9 umnm . 2. was a 3 ”a ,u 2. £2 a u a: 33¢) a: ...; .33 amu 2m :5: mma .3 oz 25 $3.; .9. 2:. to“. ...; 2m >2 2“ :5: mmxd a: 2.5 v n E a q n 7 N nu « 7 N n N am a n n 323.25; mmcd y. $5... ..mm: 3.52:3; » > umcd x mmcd .3: m. 2. us: a 3 33 o 2 min I 3 3n 3 2. we... a 2. mm: a 3 m2: 9 Slum?“ - 3 mnm n 2. was. a 3 man“ o 2. m3“ - .. 3n n a. £2. ,0 3 my: a 3 uo- a as $3 - ' 3" n as 25.. .a 2. :3 s z. 32 n «3 m a. we: . 2. mam. a S we: ... no mm: . m an” m 3 ~23 a 3 #3 o 2. mm: - a 3m m Q 33 ... oo 33 a 2 mm”: ... 2. 2.3... a man .63.; 2. «E. $8 $3 Sm >2. Em 25.. mmcd oz 9.5 335, oz «E. :3 S: Em >3 Em 25.. mad 2. if. .....308 oz. mam: a... 7 M an ul 7 N a .69 Nl M N uuzzo—m—wfiim » 7. mmcI—o x mmIJU numb» mu Ma. ummu. ,9 on uan e «o mud.“ ,9 ea wann a am «am n No menu a co mNMM .a as wM7N a n5 mama m. 7 «an n N0 qun o MOIudMn ,a as w7mM .0 «a man" a M «an a ma uman a On. war... 0 «a mnMM .0 Nolw517.& N «NM mm...I_:> 6: up; umou can arm >mo ohm tau: mm ..u 02 cat”. «4 .ol 7 M ”N n. 7 N 07 D M N wu:¢u~u_2n:m _. 7. mmmgu x mmtqu ..mwr— an ac mama a an MMwM N no umnn a on. uNuM I an an». 1 m an. Noun. 0 a: uamn «o wnmu @ no w7mm I 7 mr. 4 m an. meme 9 no wvmu a «a m7.n« s on. wman I M mmM m an mass 9 as wa7n c we wflwn 0 Ga uwn« I N mun 3 mungm; .Oz m¢> umou mum ohm >ua aha 2cm: mmcgu oz dmzu 7.. 0| 7 M an al 7 N wm ,al M N mozzu—u—Eu—m .— > mm¢4u x mmggu hmwnp mu Mm. ww0« a an ummN e «a won« a «a UNMM a a; nn 0 No uuuMN .o co wnnn ... an. uN7M a «9 mm?“ o 7 mp. n N. uNwM m. we UMMN a an. muNn a «0 N77« 7 m anm m no w«m« a 90 chw a «o mwm— o co mmmm a N mun mun—4m; .02 «xx. uwsu mum arm >wa ohm zcmt mums“. Oz Omtu O! M M.” a 7 M N 9. M7 .N 7 N on N AM a M N uu:¢u_u_zu~m v MmCJQ .hmuhh wu:¢U—m_2c~m — > mm¢._o x mcho pmwhh m.— No mnoN .o 0. was« ... en mam mu Mo uq7m a so mnmN o «a mama a a0 wwnn a QM w.” 0 Ne mama 6 we u7sq o 7 an m we wan-m6 @o ummr. a as “up.“ 9 se wu77 I 7 mmM n 2. an: a .3 mama .a n 3n n no :3 o 3 man.» a .3 ude 3 mm: a n Y." a N9 wwwa a on. unNn ,e N Mom m Na. wwwn .a nquN7~ a sa uMmN 9 an. mama e N wNM mm24¢> o: ¢¢> muou «mu ohm mm¢4u 0: onto musz; .0: my; LNG”. mum ohm >wn chm tam: mmtau 02 euro M 70 «I 7 M N am «I 7 N N no DI M N 3235.3; _ > 3...: x 3:6 gum: 32:32.3; > > mucd x 3.3“. Gm: ma Mo w7Nu 7 a. wwwn a «a maaa a as waum o 9« Non nu No mean. a an M72. 5 ea wn7m O on. wNmm a on nun v. No um7N o no umua .a co wamN .9 «9 m3“ 9 7 Non n ma ummw a as “.24 e a... much a 3 mm: a 7 mmm n Mo mowN 7 ea mNom a «a mama o no wmmm a M Nam n Ne mama o «armorfl a so mm: a «a quu o M w.” m no unmN a a. mN7nd «a u7wq .O as wwnn a N NmM m Mb mama 0 09 mean a .2. won.“ a as wanw .o N an 33.: .2. 5; 33 can Eh >mo Em :5: muzd oz 2:... 32...; a: 5.; 38 «um Sm >2. Em :5: mm .3 c: 2:... ......88 .2. 59; APPENDIX K List of publications C. C. Sweeley, S. Gates and J. F. Holland, “Mass Chromatographic Approach to Quantitation of Compounds in Complex Biological Mixtures.” In: O.A. Mamer, W.J. Mitchell and C.R. Scriver (Editors), Application of Gas Chromatography-Mass Spectrometry to the Investigation oT'Human Disease, McGill University Montreal Child—£53718 Hospital Research Institute, p. 141 (1974). C. C. Sweeley, N.D. Young, J. F. Holland and S. C. Gates, “Rapid Computerized Identification of Compounds in Complex Biological Mixtures by Gas Chromatography-Mass Spectrometry,” ,1, Chrom., 99:507 (1974). S. C. Gates, N.D. Young, J. F. Holland and C. C. Sweeley, “Computer- Aided Qualitative Analysis of Complex Biological Mixtures by Combined Gas Chromatography-Mass Spectrometry.” In: A. Frigerio and N. Castagnoli (Editors), Advances in Mass Spectrometry in Biochemistry and Medicine, Vol. I, Spectrum Publications, New York, p. 483, 1976. R. W. Wilson, C.M. Wilson, S. C. Gates and J.V. Higgins, “ oc-Ketoadipic Aciduria: A Description of a New Metabolic Error in Lysine-Tryptophan Degradation.” Pediatrics Research, 9:522 (1975). S. C. Gates, C. C. Sweeley, N.D. Young and J.F. Holland, “Automated Multicomponent Analysis of Biological Mixtures by Gas Chromatography-Mass Spectrometry.” In: A. Frigerio (Editor), Advances in Mass Spectrometry in Biochemistry and Medicine, Vol. II, Spectrum Publications, New York, p. 171, 1976. 342 343 APPENDIX K (Cont’d.) C. C. Sweeley, S. C. Gates, R.H. Thompson, J. Harten, N. Dendramis and J. F. Holland, “Techniques for Quantitative Measurements by Mass Spectrometry.” In. Proceedings of the International Symposium on Quantitative Mass Spectrometry in Life Sciences, Elsevier, 1977. S. C. Gates and C. C. Sweeley, “Library Data for the Computer Identification of Organic Acids by Gas Chromatography- Mass Spectrometry.” Submitted to Biomed. Mass Spec. S. C. Gates, N. Dendramis, R.W. Wilson and AF. Kohrman, “Identification of a New Metabolite of L—homocitrulline.” Biochem. Med., in press. -r¢ LIST OF REFERENCES 41M1 51W1 58K1 58M1 59G1 60L1 60V1 61H1 64E1 64R1 64Sl 64W1 64W2 66D1 66Sl 67M1 REFERENCES A.J.P'. Martin and R.L.M. Synge, Biochem. J., 35:1358 (1941). RJ. Williams, U. Texas Publication No. 5109, p. 7, 1951. E. Kovats, Helv. Chim. Acta, 41:1915 (1958). LG. McWilliam and RA. Dewar, Nature (London), 181:760 (1958). ' RS. Gohlke, Anal. Chem., 31:535 (1959). J.E. Lovelock and SR Lipsky, J. Am. Chem, Soc., 82:431 (1960). ‘ W. J.A. VandenHeuvel, C. C. Sweeley and E. C. Horning, J. Am. Chem. Soc., 82:3481 (1960). D. Henneberg, Z. Anal. Chem., 183:12 (1961). P. Eneroth, K. Hellstrom and R. Ryhage, J. Lipid Res., 5:245 (1964). R. Ryhage, Anal Chem., 36:759 (1964). EA Szymanski (ed.), Biomedical Applications of Gas Chromatography, Plenum Press, New York, 1964. J.T. Watson and K. Biemann, Anal. Chem., 36:1135 (1964). C.M.Williams and C. C. Sweeley, in Biomedical Applications of Gas Chromatography, H.A. Szymanski (ed.), Plenum Press, New York, p.225, 1964. C.E. Dalgliesh, E.C. Horning, M.G. Horning, K.L. Knox and K. Yarger, Biochem. J., 101:792 (1966). C. C. Sweeley, W.H. Elliott, I. Fries and R. Ryhage, Anal. Chem., 38:1549 (1966). E.A. Murphy and H. Abbey, J. Chron. Dis., 20:79 (1967). 344 68H1 68J1 68K1 68P 1 68T1 69W1 '70A1 70B1 70H1 70P1 7081 70 $2 70 W1 70 Y1 71A1 345 M. G. Horning, in Biomedical Ajplications of Gas Chromatography, Vol. 2, H.A. Szymanski (ed.T, Plenum Press, New York, p.53, 1968. R.L. Jolley and M.L. Freeman, Clin. Chem., 14:538 (1968). S. Katz, A. Confer, C.D. Scott, C.A. Burtis, M. Freeman, R.L. Jolley, N. Lee, S.A. McKee, B.E. Maryanoff, W.W. Pitt and K. S. Warren, An Annotated Bibliography of Low Molecular Weight Constituents of Human Urine,‘ USAEC Report ORNL-TM-2394, Oak Ridge, 1968. L. Pauling, Science, 168:265 (1968). R. Tham, L. Nystrom and B. Holmstedt, Biochem. Pharmacol., 17:1735 (1968). C.M. Williams, A.H. Porter, M. Greer, K.N. Scott and C. C. Sweeley, Biochem. Med., 3:164 (1969). J.R. Althaus, K. Biemann, J. Biller, P. F. Donaghue, D.A. Evans, H.H. Forster, H.S. Hertz, C.E. Hignite, R.C. Murphy, G. Preti and V. Reinhold, Experentia, 26:714 (1970). C.A.Burtis, G. Goldstein and C.D. Scott, Clin. Chem. 16:201 (1970). RA. Hites and K. Biemann, Anal. Chem., 42:855 (1970). W.W. Pitt, Jr., C.D. Scott, W.F. Johnson and G. Jones, Jr. , Clin. Chem., 16:657 (1970). C.D. Scott, D.D. Chilcote and W.W. Pitt, Jr., Clin. Chem., 16:637 (1970). ‘ C. C. Sweeley, B.D. Ray, W.I. Wood, J. F. Holland and M. I. Krichevsky, Anal. Chem., 42:1505 (1970). G. Z. Williams, D.S. Young, M.R. Stein and E. Cotlove, Clin. Chem.,, 16:1016 (1970). D.S. Young, Am. J. Clin. Path. , 53:803 (1970). P. Armitage, Statistical Methods in Medical Research, John Wiley and Sons, New York, p.406, 1971. 71C1 71H1 71H2 71H3 71H4 71J1 71131 71M1 71M2 7101 71P1 71R1 71V1 '71Y1 7121 72B1 72 C1 346 J. C. Crawhall, O. Mamer, S. Tjoa and J. C. Claveau, Clin. Chim. Acta, 34:47 (1971). E.K. Harris and D. L. DeMets, Clin. Chem., 17 :983 (1971). E. C. Horning and M.G. Horning, J. Chrom. _S_<;_i., 9:129 (1971). E.C. Homing and M.G. Horning, Clin. Chem. , 17:802 (1971). F. Hutterer, J. Roboz, L. Sarkozi, A. Ruhig and P. Bacchin, Cli . Chem., 17:789 (1971). E. Jellum, O. Stokke and L. Eldjarn, Sgand. J, Clin. Lab. Invest, 27:273 (1971). N. C. Law, V. Aandahl, H.M. Fales and G. W.A. Milne, Clin, Chim. Acta, 32:221 (1971). M.L. Moss, C.A. Horton and J. C. White, Annu. Rev. Biochem., 40:573 (1971). J.E. Mrochek, W. C. Butts, W.T. Rainey, Jr. and C.A. Burtis, Clin. Chem. , 17:72 (1971). R.W.A. Oliver and S.A. Oliver, The Analysis of Children’s Urine, Heyden, London, 1971. ‘ L. Pauling, A.B. Robinson, R. Teranashi and P. Cary, Proc. Nat. Acad. Sci., USIL 68:2374 (1971). ' J.W. Rosevear, K.J. Pfaff and E.A. Moffitt, Clin. Chem., 17:721 (1971). J.F. van de Calseyde, R.J.H. Scholtis, N.A. Schmidt and C.J.'J.A. Leyten, Clin. Chim. Acta, 32:361 (1971). D.S. Young, J.A. Epley and P. Goldman, Clin. Chem., 17:765 (1971). A. Zlatkis and H.M. Liebich, Clin. Chem., 17:592 (1971). w. C. Butts, Anal. Biochem., 46:187 (1972). B.J. Cerimele, D. C. Clapp, G. C. Cokinos, B.D. Obermeyer and RL. Wolen, Clin. Chem. , 18:744 (1972). 72C2 72 C3 72 C4 72 C5 72G1 72 G2 72G3 72H1 72J1 72P1 72 P2 72R1 72R2 72T1 72W1 7331 7332 347 RA. Chalmers and R.W.E. Watts, Analyst, 97:951 (1972). E.A. Chalmers and R.W. E. Watts, Analyst, 97:224 (1972). RA. Chalmers and R.W.E. Watts, Analyst, 97:958 (1972). D.D. Chilcote and J.E. Mrochek, Clin. Chem., 18:778 (1972). D. Gompertz and G.H. Draffan, Clin. Chim. Acta, 37:405 (1972). I D. Gompertz and G.H. Draffan, (in. Chim. Acta, 40:5 (1972). RR. Grams, E.A. Johnson and E.S. Benson, Am. J. Clin. Path., 55:188 (1972). S. Hansen, T.L. Perry, D. Lesk and L. Gibson, Clin. Chim. Acta, 39:71 (1972). E. Jellum, O. Stokke and L. Eldjarn, Clin. Chem., 18:800 (1972). J.E. Pettersen, E. Jellum and L. Eldjarn, Clin. Chim. Acta, 38:17 (1972). W.W. Pitt, C.D. Scott and G. Jones, Clin. Chem., 18:767 (1972). A.H. Reed, D. C. Cannon, V.J. Pileggi and J. W. Winkelman, Clin. Biochem., 6:29 (1972). A.H. Reed, D. C. Cannon, J.W. Winkelman, Y.P. Bhasin, B.J. Henry and V.J. Pileggi, Clin. Chem., 18:57 (1972). K. Tanaka, Clin. Chim. Acta, 43:151 (1972). P. Wlnkel, J. Lyngbye and K. Jorgensen, Scand. J. Clin. Lab. Invest., 30:339 (1972). L. Baczynskyj, J. Duchamp, J.F. Zieserl and U. Axen, Anal. Chem., 4522784 (1973). K. Blau, G.K. Summer, H. C. Newsome, C.H. Edwards and O.A. Mamer, Clin. Chim. Acta, 45:197 (1973). 73C1 73C2 73C3 73D1 73E1 73G1 73H1 73H2 73H3 73J1 73K1 7 3K2 73L1 73M1 7 3M2 7 3N1 73N2 73P1 348 D.D. Chilcote, Clin. Chem., 19:826 (1973). D.D. Chilcote and C.D. Scott, Anal. Chem., 45:721 (1973). V.J. Cirillo, J. Chrom., 81:197 (1973). M. Duran, D. Ketting, P.K. deBree, C. van der Heiden and S.R. Wadman, Clin. Chim. Acta, 49:117 (1973). L. Eldjarn, E. Jellum and O. Stokke, Clin. Chim. Acta, 49:311 (1973). S.L. Grotch, Anal. Chem., 45:2 (1973). SR. Heller, H.M. Fales and G. W.A. Milne, Organ. Mass Spec., 7:107 (1973). J.F. Holland, C. C. Sweeley, B.E. Thrush, B.E. Teets and MA. Bieber, Anal. Chem., 45:308 (1973). W. F. Holmes, W.H. Holland, B.L. Shore, D.M. Bier and W.R. Sherman, Anal. Chem., 45:2063 (1973). D.J. Jenden and R.W. Silverman, J. Chrom. Sci., 11:602 (1973). S.Katz, W.W. Pitt, Jr. and G. Jones, Jr., Clin. Chem., 19:817 (1973). K.-S. Kwok, R. Venkataraghavan and F. W. McLafferty, é Am. Chem. Soc., 95:4185 (1973). G. Lancaster, P. Lamm, C.R. Scriver, S.S. Tjoa and O.A. Mamer, Clin. Chim. Acta, 48:279 (1973). B. F. Maume, P. Bournot, J. C. Lhuguenot, C. Baron, F. Barbier, G. Maume, M. Frost and P. Padieu, Anal. Chem., 45:1073 (1973). B.F. Maume and J.A. Luyten, J. Chrom. Sci., 112607 (1973). H. Nau and K. Biemann, Anal. Lett., 6:1071 (1973). A. Nonaka, Anal. Chem., 45:483 (1973). I.-B. Peetre, Chromatographia, 6:257 (1973). u \ 1 I o a u (.n ('9 n . c IJI 73P2 73R1 73R2 73S1 7382 7 383 7384 73W1 73W2 73W3 73W4 73Z1 74A1 74B1 74B2 74B3 349 J.E. Pettersen, S. Landaas and L. Eldjarn, Clin. Chim. Acta, 48:213 (1973). R. Reimendal and J. Sjovall, Anal. Chem., 45:1083 (1973). A.B. Robinson, P. Cary, B. Dore, I. Keaveny, L. Brenneman, M. Turner and L. Pauling, Int= Res. Comm. System, 73-3, 30-9-1 (1971). C.D. Scott, D.D. Chilcote, S. Katz and W. W. Pitt, Jr., J. Chrom. Sci., 11:96 (1973). ' C.D. Scott and N.E. Lee, J. Chrom., 83:383 (1973). C.R. Scriver, C. L. Clow and P. Lamm, Clin. Biochem., 6:142 (1973). O. Stokke, E. Jellum, L. Eldjarn and R. Schnitler, Clin. Chim. Acta, 45:391 (1973). J. T. Watson, D.R. Pelster, B.J. Sweetman, J. C. Frolich and J.A. Oates, Anal. Chem., 45:2071 (1973). J.O. Westgard and M.R. Hunt, Clin. Chem., 19:49 (1973). T.A. Witten, S.P. Levine, M.T. Killian, P.J.R. Boyle and S.P. Markey, Clin. Chem., 19:963 (1973). T.A. Witten, S.P. Levine, J. 0. King and s. P. Markey, gin. Chem., 19:586 (1973). A. Zlatkis, H.A. Lichtenstein, A. Tishbee, F. Shunbo and H.M. Liebich, J. Chrom. Sci., 11:299 (1973). M. Axelson, T. Cronholm, T. Curstedt, R. Reimendal and J. Sjovall, Chromatographia, 7:502 (1974). J.D. Baty and A.P. Wade, Anal. Biochem., 57:27 (1974). W. Bertsch, R.C. Chang and A. Zlatkis, J. Chrom. Sci., 12:175 (1974). J.E. Biller and K. Biemann, Anal. Lett., 7:515 (1974). 74C1 74C2 74 C3 74C4 74D1 74D2 74D3 74E 1 74G1 74H1 74H2 74K1 7 4K2 74L1 350 R.A. Chalmers, S. Bickle and R.W.E. Watts, Clin. Chim. Acta, 52:31 (1974). R.A. Chalmers, A.M. Lawson and R. W. Watts, Clin. Chim. Acta, 52:43 (1974). R.A. Chalmers and R.W.E. Watts, Clin. Chim. Acta, 55:281 (1974). " C.E. Costello, H.S. Hertz, T. Sakai and K. Biemann, Clin. Chem., 20:255 (1974). J. Dosman, J. C. Crawhall, G.A. Klassen, O.A. Mamer and P. Neuman, Clin. Chim. Acta, 51:93 (1974). W.J. Dreyer, A. Kuppermann, H.G. Boettger, C.E. Giffin, D.D. Norris, S. L. Grotch and LP. Theard, Clin. Chem., 20:998 (1974). K.M. Dubowski, Clin. Chem., 20:966 (1974). L. Eldjarn, E. Jellum and O. Stokke, in Advances in Mass Spectrometry in Biochemistry and Medicine, A. Frigerio and N. Castagnoli (eds.), Raven Press, New York, p.287, 1974. ‘ I. Gan, J. Korth and B. Halpern, J. Chrom., 92:435 (1974). E.K. Harris, Clin. Chem., 20:1535 (1974). E. C. Horning, M. G. Horning, J. Szafranek, P. VanHout, A.L. German, J.P Thenot and C.D. Pfaffenberger, J. Chrom., 91:367 (1974). B.J. Kimble, B.E. Cox, R.V. McPherron, R.W. Olsen, E. Roitman, F. C. Walls and A.L. Burlingame, J. Chrom. Sci., 12:647 (1974). P.T. Kissinger, L.J. Felice, R.M Riggin, L.A. Pachla and D. C. Wenke, Clin. Chem., 20:992 (1974). C.R. Lee and R.J. Pollitt, in Mass Spectrometry in Biochemistry and Medicine, A. Frigerio and N. Castagnoli, (eds.), Raven Press, New York, p.365 , 1974. 74L2 74M1 74M2 74M3 74M4 74M5 74N1 74N2 74N3 74R1 74R2 7481 7482 74Z1 75A1 75A2 351 J.A. Luyten and G.A.F.M. Rutten, J. Chrom., 91:393 (1974). F. W. McLafferty, R.H. Hertel and R.D. Villivock, Org. Mass Spec., 9:690 (1974). O.A. Mamer and S.S. Tjoa, Clin. Chim. Acta, 55:199 (1974). S.P. Markey, W.G. Urban, A.J. Keyser and S.L Goodman, in Advances in Mass Spectrometry, A.R. West (ed.), Vol. 6, p.187, 1974. f L. W. Mayron, E. Kaplan, S. Ailing and J. Bechtel, Clin. Chem., 20:172 (1974). J.E. Mrochek and W.T. Rainey, Jr., Anal. Biochem., J 57:173 (1974). N. Narasimhachari, K. Leiner and C. Brown, gin. Chim. Acta, 62:245 (1974). H. Nau and K. Biemann, Anal. Chem_., 46:426 (1974). M. Novotny, M. L. McConnell, M. L. Lee and R. Farlow, Clin. Chem. , 20:1105 (1974). R.L. Reece, Perspectives in Biology and Medicine, Winter, p.227, 1974. A.B. Robinson and L. Pauling, Clin. Chem., 20:961 (1974). RE. Summons, W.E. Pereira, W.E. Reynolds, T. C. Rindfleisch and A.M. Duffield, Anal. Chem., 46:582 (1974). C. C. Sweeley, N.D. Young, J.F. Holland and S. C. Gates, J. Chrom., 99:507 (1974). A. Zlatkis, W. Bertsch, D.A. Bafus and H.M. Liebich, J. Chrom., 91:379 (1974). F. P. Abramson, Anal. Chem., 47:45 (1975). D.A. Applegarth and P.M. Ross, Clin. Chim. Acta, 64:83 (1975). 75B1 75B2 75D1 75E 1 75G1 75H1 7 5H2 75H3 75H4 75K1 75K2 75K3 75K4 75K5 75 L1 75 L2 75L3 75L4 352 F.W. Bultitude and S.J. NeWham, Clin. Chem., 21:1329 (1975). R.W. Burnett, Clin. Chem., 21:1935 (1975).} H. Dirren, A.B. Robinson and L. Pauling, Clin. Chem., 21:1970 (1975). L. Eldjarn, E. Jellum and O. Stokke, Clin. Chem. , 21:63 (1975). E.M. Gindler, Clin. Chem., 21:309 (1975). E.K. Harris, Clin. Chem., 21:1457 (1975). B. Hedfjall and R. Ryhage, Anal. Chem., 47:666 (1975). M. G. Horning, J. Nowlin, C.M. Butler, K. Lertratanangkoon, K. Sommer and R. Hill, Clin. Chem., 21:1282 (1975). J.A. Hoskins and R.J. Pollitt, J. Chrom., 109:436 (1975). B. J. Kimble, F. C. Walls, R.W. Olsen and A.L. Burlingame, 23rd Annual Conference on Mass Spectrometry and Allied _T_____opics, Houston, p. 503, 1975. P.T. Kissinger, R.M. Riggin, B.L. Alcorn and L.-D. Rau, Biochem. Med., 13:299 (1975). T. Kitagwa, B. A. Smith and E. S. Brown, Clin. Chem. ., 21: 735 (1975). B.A. Knights, M. Legendre, J. L. Laseter and J.S. Storer, Clin. Chem., 21:888 (1975). B.R. Kowalski, Anal. Chem., 47:1153A (1975). 8. Landaas, Clin. Chim. Acta, 64:143 (1975). A.M. Lawson, Clin. Chem., 21:803 (1975). H.M. Liebich, O. Al-Babbili, A. ZlatkiS and K. Kim, Clin. Chem., 21:1294 (1975).) S. Lindstedt and G. Steen, Clin. Chem., 21:1964 (1975). 75N1 7501 75P1 75P2 75P3 7581 75T1 75T2 75T3 75W1 75W2 75W3 75Y1 75Y2 75Z1 76B1 ‘..---'— ?$'§fide—M.—Elu ' 353 M. Novotny and R. Farlow, J. Chrom., 103:1 (1975). B. Oberholtzer, C.B.S. Wood, T. Palmer and B.M. Harrison, Clin. Chim. Acta, 62:299 (1975). G.M. Peysna, F. W. McLafferty, R. Venkataraghavan and H.E. Dayringer, Anal. Chem., 47:1161 (1975). LR. Politzer, S. Githens, B.J. Dowty and J.L. Laseter, J. Chrom. Sci., 13:378 (1975). H. Przyrembel, D. Bachmann, L Lombeck, K. Becker, U. Wendel, S.K. Wadman and H.J. Bremer, Clin. Chim. Acta, 58:257 (1975). F.W. Sundrman, Jr., Clin. Chem., 21:1873 (1975). K. Tanaka, in Biology of Brain Dysfunction, Vol. 3, G.E. Gaull (ed.), Plenum Press, New York, p.145, 1975. J.A. Thompson and S.P. Markey, Anal. Chem., 47:1313 (1975). J.A. Thompson, S.P. Markey and P.V. Fennessey, Clin. Chem., 21:1892 (1975). R.W.E. Watts, R.A. Chalmers and A.M. Lawson, The Lancet, Feb. 15, 368, 1975. M. Werner and W. L. Marsh, C.R. C. Critical Reviews 111’ Clinical Laboratory Sciences, p.81, 1975. R.M. Wilson, C.M. Wilson, 8. C. Gates and J.V. Higgins, Ped. Res., 9:522 (1975). D.S. Young, New Engl. J. Med., 292:795 (1975). N.D. Young, J.F. Holland, J.N. Gerber and C. C. Sweeley, Anal. Chem., 47:2373 (1975). F. Zambotti, K. Blau, G.S. King, 8. Campbell and M. Sandler, Clin. Chim. Agta, 61:247 (1975). T.H. Bindel, P.V. Fennessey, B.S. Miles and S.L Goodman, Clin. Chim. Aster 662209 (1976). 7 6B2 76C1 76C2 76D1 76D2 76D3 7 6G1 7 6G2 76G3 76H1 7 6H2 76L1 76L2 7 6M1 '76N1 354 L. Bjorkman, C. McLean and G. Steen, Clin. Chem., 22:49 (1976). RA. Chalmers, M.J.R. Healy, A.M. Lawson, J.T. Hart and R.W.E. Watts, Clin. Chem., 22:1292 (1976). R.A. Chalmers, M.J.R. Healy, A.M. Lawson and R.W.E. Watts, Clin. Chem., 22:1288 (1976). E.G. deJong, J. van Bekkum, H.A. van’t Klooster and J. Freudenthal, preprint, 1977. R6. Dromey, Anal. Chem., 48:1464 (1976)° R.G. Dromey, M. J. Stefik, T. C. Rindfleisch and A.M. Duffield, Anal.‘ Chem., 48:1368 (1976). S. C. Gates, N.D. Young, J. F. Holland and C. C. Sweeley, in Advances in Mass Spectrometry in Biochemistry and Medicine, Spectrum Publications, New York, p.483, 1976. N. Gregersen, Clin. Chim. Acta, 70:417 (1976). S. C. Gates, MSSMET User Instructions and Technical Documentation for MSSMET. Unpublished, both available from the author or Dr. Sweeley. 1976. E.K. Harris, Clin. Chem., 22:1343 (1976). R.H. Horrocks, E.J. Hindle, P.A. Lawson, D.H. Orrell and A.J. Poole, Clin. Chim. Acta, 69:93 (1976). A.M. Lawson, R.A. Chalmers and R.W.E. Watts, Clin. Chem., 22:1233 (1976). S. Lindstedt, K. Norberg, G. Steen and E. Wahl, Clin. Chem., 22:1330 (1976). R.D. Malcolm and R. Leonards, Clin. Chem., 22:623 (1976). E. Nakamura, L.E. Rosenberg and K. Tanaka, Clin. Chim. Acta, 68:127 (1976). I. . . l I l n - n n a . u n a . u n a IN. ,. \4 n . n L . u 'l ‘ u u. '6 JV 7 601 76P1 76P2 76R1 7681 7682 76T1 76Y1 76Y2 77B1 77B2 77C1 77F1 77G1 7762 7701 -' " TV.” 1."; *k‘lh-Wq -2;., _. ' ‘ 355 D.M. Olsen, B.L. Kane and P.H. Proctor, New Eng. J. Med., 294:925 (1976). LA. Politzer, B.J. Dowty and J. L. Laseter, Clin. Chem., 22:1775 (1976). H. Przyrembel, U. Wendel, K. Becker, H.J. Bremer, L. Bruinvis, D. Ketting and S.K. Wadman, Clin. Chim. Acta, 66:227 (1976). J.L Routh and W.D. Paul, Clin. Chem., 22:837 (1976). D.M. Schoengold, R.H. deFiore, D.J. Miner and P. B. ; Hamilton, Clin. Chem., 22:16 (1976). i D.M. Schoengold, RH. deFiore, D. J. Miner and P.B. Hamilton, Clin. Chem., 22:19 (1976). B.E. Thiers, G.T. Wu, A.H. Reed and L.K. Oliver, Clin. Chem., 22:176 (1976). T. Yamamoto, K. Doi, Y. Takeuchi, M. Baba and M. Tanaka, Clin. Chim. Acta, 68:241 (1976). D.S. Young, Clin. Chem., 22:1555 (1976). B.E. Blaisdell, Anal. Chem., 49:180 (1977). B.E. Blaisdell, private communication. R.A. Chalmers, A.M. Lawson, R.W.E. Watts, Clin. Chem., 23:903 (1977) C. Fenselau, Anal. Chem., 49:563A (1977). S. C. Gates, C. C. Sweeley, N.D. Young and J. F. Holland, in Advances in Mass Spectrometry in Biochemistry and Medicine, Vol. II, in press. S. C. Gates and C. C. Sweeley, submitted to Biomed. Mass Spec. 8. Ohmori, M. Ikeda, S. Kira and M. Ogata, Anal. Chem., preprint. 356 7781 C. C. Sweeley and N. Dendramis, private communication. 77T1 J.A. Thompson, Clin. Chem., 23:901 (1977) IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII uIAluminum;(wmmmmymmmH