“Tm LIBRARY 7...“? Michigan State ‘I" L University This is to certify that the thesis entitled THE USE OF MINISTRS AND MITOCHONDRIAL DNA TO IDENTIFY HANDLERS OF PIPE BOMBS presented by Stefanie Lee Kremer has been accepted towards fulfillment of the requirements for the MS. degree in Criminal Justice -Q" a Major ProTessor’s Signature 3/ A/f/ O 8’ Date MSU is an affinnative-action, equal-opportunity employer ..-u-9o-.--I--.-.-.—.---n-uqlr.-u-n-o-u-u-n-.-o-g-n—.—.-.--—.-—.—.-.— -.---.-.—-—.-c-.----—-a—--—.--u-—---— PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5/08 KlProj/Acc8Pres/ClRC/DateDue.indd THE USE OF MINISTRS AND MITOCHONDRIAL DNA TO IDENTIFY HANDLERS OF PIPE BOMBS By Stefanie Lee Kremer A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Criminal Justice 2008 Abstract THE USE OF MINISTRS AND MITOCHONDRIAL DNA TO IDENTIFY HANDLERS OF PIPE BOMBS By Stefanie Lee Kremer The deflagration of pipe bombs produces very high temperatures, which, in combination with the general nature of DNA from shed skin cells, means that only degraded DNA is likely to remain on the resultant bomb fragments. Further, because the bomb surface has only touch DNA, low copy number (LCN) techniques must be utilized during analysis. Previous research employed short tandem repeat (STR) (Esslinger et al., 2004) and mitochondrial DNA (mtDNA) (Gehring, 2004) analyses to identify the handlers of pipe bombs. The rate of obtaining an STR profile was very low, while increased success was garnered with mtDNA. The goal of the current research was to use miniSTRs to better identify individuals who handle or assemble pipe bombs. In this research, 17 volunteers were asked to touch two sets of pipe bomb components, one made of PVC and the other of steel, for a total of 34 bombs, which were then deflagrated. DNA was amplified using two sets of multiplexed miniSTR primers as well as mtDNA primers. MtDNA profiles were more likely to produce correct assignments than miniSTRs. Further, when data from both miniSTR and mtDNA profiles were combined, the number of correctly assigned bombs was even higher. These results indicate that both nuclear and mtDNA should be used in conjunction when DNA that is degraded and in low quantity is encountered on items of evidence. Acknowledgements I would like to thank those who made this research possible. Thank you to the members of the Michigan State Police Bomb Squad, especially First Lieutenant Shawn Stallworth and Detective Sergeant Timothy Ketvirtis who were instrumental in coordinating the logistics of deflagrating all of the bombs in this project. My gratitude goes to my thesis advisor, Dr. David Foran, for his guidance during the research and the amount of time spent revising this manuscript. Further, I would like to thank Shane Hoffmann for his assistance with collecting the bomb fragments. Finally, thank you to Dr. Mahesh Nalla for your time, input, and guidance as one of my committee members. I would also like to thank my family and friends whose support and encouragement during the last few years provided me with the strength to complete this part of my academic career. I hope I have made them proud. iii Table of Contents List of Tables .................................................................................................................. v List of Figures ................................................................................................................ vi Introduction ..................................................................................................................... 1 Pipe bombs as evidence ............................................................................................... 3 STRs and miniSTRs .................................................................................................... 4 Low copy number DNA .............................................................................................. 5 DNA from pipe bombs ................................................................................................ 6 Research goals ............................................................................................................. 9 Materials and Methods .................................................................................................. 11 Obtaining and decontaminating pipe bomb materials ................................................. 11 Handling of containers by subjects ............................................................................ 11 Deflagration .............................................................................................................. 12 DNA isolation and extraction .................................................................................... 12 Characterization of bomb DNA and reference samples using miniSTRs .................... 13 Characterization of bomb and reference samples using mtDNA ................................. 15 Determination of genetic profiles and assignations .................................................... 17 Results .......................................................................................................................... 19 Pipe bomb deflagration and DNA isolation ................................................................ 19 MiniSTR profiles and bomb assignation .................................................................... 20 MtDNA sequencing and bomb assignation ................................................................ 25 Bomb assignations using both miniSTRs and mtDNA ............................................... 27 MiniSTR amplicon size vs. number of alleles amplified ............................................ 31 Examination of PVC and steel pipe bombs ................................................................ 33 Discussion ..................................................................................................................... 36 Conclusions ................................................................................................................... 50 Appendix A ................................................................................................................... 51 References ..................................................................................................................... 52 iv List of Tables Table 1. Primer pairs used to amplify mtDNA from pipe bombs ................................... 16 Table 2. Bomb assignations using miniSTRs ................................................................ 21 Table 3. Summery of bomb assignations using miniSTRs ............................................. 22 Table 4. Summery of bomb assignations using mtDNA ................................................ 25 Table 5. Bomb assignations using miniSTRs, mtDNA, and both mtDNA and miniSTRs. .............................................................................................................................. 28 Table 6. Summery of bomb assignations using both mtDNA and miniSTRs ................. 30 Table 7. Summery of the number of alleles amplified per locus .................................... 32 Table 8. Comparison of PVC and steel pipe bomb assignations .................................... 34 Table 9. Comparison of PVC and steel pipes combining miniSTR and mtDNA data 35 List of Figures Figure l. Post-blast debris ............................................................................................ 19 Figure 2. Correctly and incorrectly assigned pipe bombs at 50 RFU using miniSTRs ...22 Figure 3. Correctly and incorrectly assigned pipe bombs at 100 RFU using miniSTRs .23 Figure 4. Correctly and incorrectly assigned pipe bombs at 1000 RF U using miniSTRs. .............................................................................................................................. 24 Figure 5. Correctly and incorrectly assigned pipe bombs using mtDNA ......................... 26 Figure 6. Average number of alleles amplified per locus at 50 RFU in bomb profiles. ..32 Figure 7. Average number of correct alleles per locus at 50 RFU when bomb profiles were compared to reference profiles ...................................................................... 33 vi Introduction Improvised explosive devices (IEDs) are used by many rogue organizations, militia groups, and individuals to cause destruction and panic. Between January 2001 and February 2006, over 18000 explosive incidents occurred in the United States, including almost 3400 bombing incidents that caused 409 injuries, 56 deaths, and over $25 million in damages (Bureau of Alcohol, Tobacco, Firearms, and Explosives, 2006). IEDs are also problematic on the world stage. Between July 2003 and April 2008, over 1600 American soldiers were killed by IEDs while in Iraq (iCasualties.org, April 2008). The increasing use of IEDs has created the need to identify the person or persons who handled, assembled, and deflagrated a device. IEDs can be made from various materials and employ myriad configurations. Over 40% of IEDs worldwide are pipe bombs (Burke, 2007). High profile incidents involving pipe bombs include mail bombings by Theodore Kaczynski (the Unabomber), the Olympic Park Bombings by Eric Rudolph in 1996, and the Columbine High School massacre in 1999. Pipe bombs are hazardous because of the high-velocity fragments that are produced following deflagration; fiagments from the container can reach speeds of 20000 feet per second (Lenz, 1965). They can be assembled from materials that are purchased without difficulty at hardware and sporting goods stores and typically consist of a length of pipe, end caps, explosive, and detonator mechanism. The explosive is usually black or smokeless powder sold to individuals who make their own cartridges and shotgun shells for hunting purposes. Smokeless powders can be subdivided into three groups: single base (containing nitrocellulose (NC)), double base (NC and nitroglycerine (NG)), and triple base (NC, NG, and nitroguanidine) (Beveridge, 1998). Single and double base smokeless powders are commercially available and therefore encountered in IEDs. Pipe bombs may also contain projectiles, such as nails, bolts, or shot, to increase the amount of injury and damage they cause. The method of detonation can be as simple as a fuse or as complex as a switch or timing system. Several types of improvised switches that have been used to activate IEDs include mousetraps, aluminum foil, and Clothespins (Thurman, 2006). Timing delays on pipe bombs allow the perpetrator to leave the scene before the bomb deflagrates. These too can be improvised from common objects including wind-up clocks and timers and items such as cell phones and pagers that can be remotely activated (Thurman, 2006). The relative ease of obtaining the materials as well as the general stability of the explosive means pipe bombs can be effortlessly transported with a low risk of accidental deflagration. Metal is the container material used most often for making pipe bombs (National Research Council, 1998). Between 1992 and 1994, 485 bombing incidents occurred in the United States that utilized metal pipe containers and black or smokeless powder. Dm'ing the same time period, there were 105 incidents in which plastic pipes were found (National Research Council, 1998). Both metal and plastic pipes are regulme employed for construction purposes and can be purchased without difficulty. Steel pipes are generally galvanized, i.e., coated with a thin layer of zinc (to prevent rust) that will peel off at temperatures above 200°C (American Galvanizers Association, 2000). Plastic pipe bomb containers are usually manufactured fiom polyvinyl chloride (PVC). Steel pipes and end caps typically have threaded fittings, whereas PVC pipes and end caps have smooth fittings that require an adhesive such as PVC cement or cyanoacrylate to glue them together. Pipe bombs as evidence When pipe bomb evidence is gathered, certain class characteristics such as the type or brand of the explosive can be determined. Some manufacturers of smokeless powders add a specific unique chemical, or taggant, to the powder that can be detected following a deflagration. Further, the general morphology of the powder is a characteristic that allows for brand identification (Beveridge, 1998). These class characteristics may be helpful for identifying suspects; however, individualizing evidence provides a definitive association between the perpetrator and the crime scene. Elements left on a pipe bomb that could definitively identify the handler include fingerprints and DNA from the person or persons who assembled or handled the device; however, fingerprints are not likely to be found on deflagrated pipe bombs due to the intense heat produced during the deflagration (Beveridge, 1998). This makes the search for DNA on deflagrated pipe bombs that much more imperative. DNA is present in every cell in the body, except mature red blood cells. While touching a surface, skin cells are sloughed off and can adhere to it. Van Oorshot and Jones (1997) demonstrated that not only could DNA be recovered from latent fingerprints, but also items that had been touched by individuals could be matched back to those individuals. This type of DNA analysis has become increasingly common and is often referred to as “touch DNA” or “trace DNA” analysis. Researchers have shown that DNA can be deposited on a variety of surfaces. Balogh et al. (2003) found that fingerprints deposited on paper contain both nuclear DNA (nDNA) and mitochondrial DNA (mtDNA). Other surfaces such as glass, plastic, ceramic, and vinyl (van Oorshot and Jones, 1997) can also retain DNA afier being touched. Therefore, it is possible that DNA from skin cells can be transferred to both PVC and metal pipes during the assembly of a pipe bomb. S TRS and miniSTRs The standard method for examining nDNA in forensic applications is short tandem repeat (STR) analysis. STRs consist of tandemly repeated segments of DNA, each of which is typically two to six bases long (Butler, 2005). Individuals contain varying numbers of the tandem repeats at many loci, and it is these repeat numbers that can be used to identify a person. The STR loci used for forensic identification are included in several commercial DNA-typing kits, including PowerPlex® 16 System by Promega and AmpFlSTR Identifiler® PCR Amplification Kit by Applied Biosystems. Both kits utilize the polymerase chain reaction (PCR) to make copies of the STR loci. All of the primers used to amplify the loci are contained within a single reaction mixture (multiplex), reducing the number of PCRs that need to be performed. In the PowerPlex l6 kit, the 16 loci range in size from 100 base pairs (bp) to 450 bp (Promega Corporation, 2007), while the range in the Identifiler kit is from 100 bp to 375 bp (Applied Biosystems, 2001). One primer used to amplify each locus has a chemical dye attached to it A laser excites the dye during the detection stage and a detector capttn‘es the wavelength emitted. The dyes have different ranges of emission, and thus the detector can differentiate them. In order to distinguish the loci detected using a given dye, the loci are separated by size, which is influenced by both the size of the STR itself and the amount of flanking DNA amplified on either side of it. The latter can be increased by moving the primers away from the repeat region, which allows for loci using the same dye to occupy their own size range, preventing them from overlapping in size. When high quality DNA is amplified during STR analysis, the increase in locus size does not usually affect the level of the profile produced. However, in DNA that is degraded, the larger loci may not amplify. MiniSTRs were designed so that primers anneal closer to the repeat region and therefore yield profiles utilizing much smaller segments of DNA (Wiegand and Kleiber, 2001; Butler et al., 2003). They range from approximately 50 bp to 300 bp (Krenke, 2002), which is particularly beneficial when attempting to amplify degraded DNA. Multiple miniSTR primer sets have been combined to produce a “miniplex” in which three to six loci are amplified in the same PCR (Butler et al., 2003). Opel et a1. (2006) showed that the “miniplexes” are more successful at producing a full STR profile with degraded DNA from human bone than using traditional STR analysis. Low copy number DNA Findley et a1. (1997) showed that it is possible to obtain STR results from a single hmnan cell. The optimum amount of DNA added to the STR PCR is around 1 ng (Promega Corporation, 2007; Applied Biosystems, 2001). Given that a normal diploid human cell contains approximately 6 pg of nDNA, around 167 cells are needed for an optimal reaction, however this amount may not be available. When the amount of genomic DNA present in a sample is less than 100 pg, it is commonly referred to as low copy number (LCN) DNA (Gill et al., 2000). After touching an object there may be only a few cells from which DNA can be obtained, thus rendering it LCN. Findley et a1. (1997) identified several problems that occur in LCN analysis. Stochastic effects materialize when random loci or alleles are sampled more than others leading to peak height imbalance or complete allelic drop—out. It has been suggested that the amplification and analysis of LCN DNA can be made more sensitive by increasing the number of PCR cycles (Wiegand and Kleiber, 1997; van Hoofth et al., 1998); however, this could decrease the accuracy of the results due to over-amplification of exogenous DNA, which may even be amplified in place of the target DNA. Taberlet et al. (1996) proposed that alleles only be considered as part of a LCN profile if they appear in at least two PCR replicates. Budowle et al. (2001) expanded on this and suggested several approaches to increase sensitivity without increasing cycle number. They include: 1) reducing the PCR volume; 2) post-PCR filtration to remove ions that compete with DNA when being injected into the capillary; 3) use of formamide with low conductivity; 4) adding more amplified product to the formamide; and 5) increasing injection time. Finally, it should be noted that STR profiles produced from LCN DNA should not be used to exclude individuals; because of the increased amount of allelic drop-in and drop- out, the accuracy of the results may be low. DNA fiom pipe bombs The deflagration of pipe bombs produces very high temperatures for a short period of time. It is not known what effect this heat has on DNA, however past research has explored DNA degradation following exposure to high temperatures. Threadgold and Brown (2003) showed that DNA from wheat seeds degrades at temperatures above 200°C. Wheat seeds were placed (Triticum aestivum L.) in an oven at temperatures ranging from 150°C to 250°C for different time periods (15 to 300 minutes). They found that at temperatures above 200°C and at times longer than 15 minutes, the DNA degraded to a point where a 246 bp portion of a nuclear gene and a 181 bp portion of a mitochondrial gene could not be amplified. Smokeless powder ignites at 315°C, causing a rapid change in temperature during deflagration of a pipe bomb (Lenz, 1965). Although the temperature produced during deflagration is not known and the length of time the bomb spends at that temperature is likely very short, it seems possible that the DNA deposited onto the surface of the bomb degrades. Two groups of researchers have analyzed the DNA remaining on pipe bomb evidence following deflagration. Esslinger et a1. (2004) isolated DNA from deflagrated pipe bombs and analyzed it using standard STRs. The study design included 20 pipe bombs that had been decontaminated using 10% bleach and UV irradiation and were then handled by subjects for approximately 30 seconds. The bombs were deflagrated in a hole in the ground that was covered with a large rock to contain the fragments. Resulting DNA was amplified at nine STR loci and the sex marker amelogenin. The difference between using unconcentrated DNA and concentrated DNA prior to amplification was also compared. No profiles contained all ten loci using the unconcentrated DNA, while concentration prior to PCR resulted in one full profile. One partial profile was recovered and six active profiles were found on the pipe bombs using unconcentrated DNA, and two partial and five active profiles were recovered using concentrated DNA. Further, the level of bomb fragmentation and fragment recovery influenced the ability to obtain a profile, with the more intact bombs yielding more complete profiles. Bombs that were categorized as being highly fragmented with few recovered pieces yielded no profiles at all. The findings indicated that nDNA can survive the heat produced during a bomb deflagration; however most of the DNA is too degraded to be amplified using standard STR primers. Gehring (2004) showed that mtDNA could be obtained fi'om deflagrated pipe bombs. MtDNA is a circular genome found in mitochondria, which can number between 80 and 680 in a cell (Robin and Wong, 1988) in contrast to STRs that are present in only two copies per cell. This, in combination with its apparently protected location within the mitochondrion (F oran, 2006), may explain why mtDNA is more likely to amplify than nDNA in degraded materials. MtDNA is maternally inherited, thus a mtDNA haplotype is shared among siblings and other maternal relatives. That reduces its usefulness as an individualizing characteristic, although it provides for greater accessibility to reference samples if a suspect or victim is not available. Gehring (2004) prepared 36 pipe bombs in a manner similar to Esslinger et al. (2004). Following deflagration in an enclosed brick room and subsequent DNA extraction, the hypervariable regions of the mtDNA genome were sequenced. Bombs were assigned blindly to study participants. Eighteen of 36 bombs were correctly assigned to the individual who handled the bomb, and seven more were assigned to a subset of three individuals who shared the same mtDNA haplotype, thus there was a 69% success rate in correctly assigning the donor to a profile from a pipe bomb. Further, the research showed a trend between the level of fragmentation of the deflagrated pipe bomb and the quantity of DNA obtained, with higher levels of fragmentation resulting in lower amounts of recovered DNA. The studies by Gehring (2004) and Esslinger et al. (2004) together revealed important points. First, it was possible to recover DNA from a deflagrated pipe bomb. This indicated that the heat produced during the deflagration was not high enough to completely destroy or remove DNA deposited on the pipe by the handler. Second, the DNA that remained following the deflagration was of a sufficient quality to yield DNA profiles. Third, sequencing of mtDNA was more successful in identifying the handler of a pipe bomb than were standard STR analyses. Finally, both studies established the connection between the level of fragmentation of a bomb and the success of obtaining positive results from the fragments. Research goals The objective of the research detailed here was to determine, using both nDNA and mtDNA, the identity of persons who assemble pipe bombs. Distinguishing individuals by mtDNA analysis is promising, however, STRs are currently the most widely used method of DNA identification in crime laboratories and mtDNA is not individualizing. MiniSTRs allow degraded evidence to be analyzed with many of the same loci as traditional STR analysis. Further, the reagents and instrumentation required for miniSTR analysis are the same as those for analyzing standard STRs, permitting integration of the miniSTR profiles into existing DNA databases. Therefore, the first goal was to investigate the efiicacy of miniSTR analysis in the identification of assemblers of pipe bomb devices. Also explored was the relationship between the size of the loci analyzed and the success rate in obtaining alleles from each. It was hypothesized that alleles from smaller loci would be more likely to amplify than alleles from larger loci. The second goal was to compare the success rate of obtaining a correct mtDNA profile to the success rate of obtaining a correct miniSTR profile. Also, the usefulness of each kind of profile in identifying the individual who handled the pipe prior to deflagration was examined. Further, both methods were used in conjunction to determine the overall success of assigning a pipe bomb to an individual. The final goal was to investigate the profiles produced from metal and PVC pipes. Esslinger et al. (2004) did not find a correlation between the container material and the ability to obtain a STR profile. The current study however used both mtDNA and nDNA to develop profiles; therefore more information was available to determine if a difference exists between the two types of containers. 10 Materials and Methods Obtaining and decontaminating pipe bomb materials IEDs were assembled from l-foot pieces of PVC or galvanized steel pipe (1 inch diameter) and PVC or steel end caps, purchased at local hardware stores. A 1/4 inch hole was drilled in the center of one end cap for each device. A PVC end cap was affixed to one end of a PVC pipe using PVC cement and allowed to dry. Pipe pieces were soaked for 1 hour in 10% household bleach, rinsed with distilled water, UV irradiated for 5 minutes, turned, and irradiated for an additional 5 minutes. Surfaces of the container were then wiped with ELIMINase® (Decon Laboratories, Inc., Bryn Mawr, PA), according to the manufacturer’s instructions and rinsed twice with sterile water to remove residual ELIMINase®. The pieces were dried in a laminar flow hood and stored in paper bags. Eighteen devices of each type were assembled. Handling of containers by subjects The use of human subjects as handlers of experimental bombs was approved by the Michigan State University Committee on Research Involving Human Subjects (IRB# 06-601). Subjects signed a consent form prior to participation. They randomly selected one PVC and one metal pipe assembly, removed them from the paper bag, handled each for 30 seconds, and then placed them back in the bag. Buccal swabs as DNA reference samples were also provided. Random identification numbers were assigned to each pipe bomb as well as the reference samples, and the rest of the experiments used blind scoring for analysis. One unhandled PVC and steel pipe were controls. 11 Deflagration Pipe bombs were transported to the Lansing Fire Fighting Facility in Lansing, MI. There, a member of the Michigan State Police Bomb Squad filled the bombs with Green Dot Smokeless Shotshell Powder (Alliant Powder Co., Radford, VA), placed an approximately 1.5 foot long safety fuse into the hole in the end cap, and attached the second end cap to the device; PVC pieces were glued together using PVC cement and steel pieces were assembled using the threaded ends. F acemasks and latex gloves were worn when handling the bomb components. The bomb was placed in the brick room, the fuse lit, and the door closed. The length of fuse took approximately 30 to 45 seconds to burn. Following deflagration, two other investigators gathered the pieces and placed them in a new paper bag. Between deflagrations, the room was swept to remove any debris. One of the investigators who gathered the pipe bomb pieces performed the rest of the analyses in this study. DNA isolation and extraction Bombs were processed individually to avoid cross-contamination. Bomb fragments were removed fi'om the paper bag under a laminar flow hood. A double-swab technique (Sweet et al., 1997) was used to recover DNA from the fragments. A cotton swab was moistened with 100 uL digestion buffer (20 mM Tris, 50 mM EDTA, 0.1% SDS, pH 7.5), swiped over the outside surface of all fragments and placed in a 1.5 mL sterile tube. A dry swab was then swiped over the surfaces to collect remaining moisture and placed in the same tube. For the metal bombs, only those fi'agments that retained the galvanized coating were swabbed. Four hundred microliters digestion buffer and 6 uL l2 proteinase K (20 mg/mL) were added to the tube. The contents were vortexed and incubated overnight at 55°C. The swabs were placed in a spin basket in a new 1.5 mL tube and centrifuged for 1 minute at 14000 revolutions per minute (rpm) to remove the liquid, which was then combined with the liquid remaining in the first tube. An equal volume of phenol was added, vortexed, and centrifuged at 14000 rpm for 15 minutes. The aqueous layer was transferred to a new tube and an equal volume of chloroform was added. The tube was vortexed and centrifuged at 14000 rpm for 15 minutes. The aqueous layer was placed on a Microcon® YM-30 spin column (Millipore Corporation, Billerica, MA) and centrifuged at 14000 x g for 15 minutes. The DNA on the column was washed twice with 100 uL TE (10 mM Tris, 1 mM EDTA, pH 7.5) and centrifuged after each wash at 14000 x g for 5 minutes. The DNA was resuspended in 20 pl. TE and stored at -20°C. Reference buccal swabs were processed in the same manner. Characterization of bomb DNA and reference samples using miniSTRs DNA obtained from each bomb and the reference samples was amplified using two sets of miniplexed primers: miniSGM (http://www.cstl.nist.gov/div83 l/strbase /miniSTR.htm) and miniNCOl (Coble and Butler, 2005). MiniSGM PCR reactions included 1 [LL Gold ST*R 10X Buffer (Promega, Madison, WI), 2.2 uL miniSGM primers (1 uM, NTST), 1.5 uL BSA (10 mg/uL), 0.25 pL AmpliTaq Gold® DNA Polymerase (5 U/uL, Applied Biosystems, Foster City, CA), 3.5 uL water, and 2 uL 1:10 diluted DNA template. Therrnocycle conditions included an initial denatmation step at 95°C for 10 minutes, followed by 42 cycles of 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute. A final extension step at 60°C for 45 minutes was then performed. 13 Reference samples underwent the same PCR procedure except that l uL 1:10 diluted template DNA was added and 32 cycles of the denaturation, annealing, and extension steps were performed. Reference and bomb samples were amplified at different times to avoid cross-contamination. Amplification using miniN C01 primers was in reactions containing 2 uL Gold ST*R 10X Buffer, 2 uL miniNCOl primers (NIST'), 1.5 uL BSA, 0.2 uL AmpliTaq Gold® DNA Polymerase, 2 uL 1:10 DNA template, and 2.5 uL water. Reference sample reactions contained the same except 1 uL 1:10 diluted DNA template was added. Cycling parameters were the same as the miniSGM reactions. All sets of reactions included a positive and negative control, as well as a reagent blank. Pipe bomb sample PCRs were performed in triplicate, while reference samples were amplified once. The PCR product was added to a Montage® PCR Unit (Millipore Corporation) along with 300 uL TE and the unit was centrifuged for 15 minutes at 1000 x g. The DNA was washed twice with 100 uL TE and centrifuged at 1000 x g for 5 minutes. The resultant DNA was resuspended in 10 uL TE. During the data collection phase, an unknown error occurred which resulted in some samples producing unreliable data. Those samples and all remaining ones were purified using an UltraCleanT“ PCR Clean- Up Kit (MO BIO Laboratories, Inc., Carlsbad, CA) according to manufacturer’s instructions. Following the extra clean-up step, the DNA was concentrated to a volume of 10 uL by adding 3.3 uL 3M NaOAc and 100 uL 95% ethanol and centrifuged for 5 minutes at 13200 rpm. The ethanol was removed and the resultant DNA was vacuumed dry and resuspended in 10 uL TE. Appendix A indicates the method used to clean each PCR. 14 Two microliters of the PCR product from the bombs and 0.5 uL GeneScanTM 500 LIZ® Size Standard (Applied Biosystems) were added to 23 uL deionized formamide. Reference samples were prepared by mixing 1 uL PCR product, 0.5 uL GeneScanTM 500 LIZ® Size Standard, and 24 uL formamide. The allelic ladder samples contained 23 IL formamide, 0.5 uL GeneScanTM 500 LIZ® Size Standard, and 2 uL of either miniSGM or miniNCOl allelic ladders (NIST). DNAs were analyzed using capillary electrophoresis on an ABI PRISM® 310 Genetic Analyzer (Applied Biosystems). Samples were electrOphoresed without denaturation using the following parameters: 5 seconds injection at 15 kV, 28 minutes run time at 15 kV, and a rim temperature of 60°C through a 47 cm x 50 um capillary. A 1X Running Buffer (Applied Biosystems) and POP-4TM Polymer for the 310 Genetic Analyzer (Applied Biosystems) were utilized. The resulting data were analyzed using GeneMapperID Software v. 3.2 (Applied Biosystems). The analysis method for the miniSGM loci was Microsatellite Default (Applied Biosystems) and for the miniNCOl loci was Microsatellite NC, created by the primary investigator. The allele panels for miniSGM and miniNCOl were mini-SGM and NC01 respectively, both created by the primary investigator based on panels provided at http://www.cstl.nist.gov/div831/strbase/miniSTR.htm. The miniplexed samples employed the CE_G5__HID_GS500 size standard setting and the D-33 Matrix Standard file (Applied Biosystems). Characterization of bomb and reference samples using mtDNA Three regions of the mitochondrial genome from each bomb sample were PCR ammified: the first halfof hyper-variable region 1 (HVl-l), the second half of hyper- 15 variable region 1 (HV1-2), and hyper-variable region 2 (HV 2). Primers are shown in Table 1; F82 was developed at the Forensic Biology Laboratory at Michigan State University and the remainder was developed at AFDIL (Edson et al., 2004). Table 1. Primer pairs used to amplify mtDNA from pipe bombs. HVl-l Non-nested Primer Pair F15989 R16322 5’ CCCAAAGCTAAGATTCTAAT 5’ TGGCI‘TT AT GTACI‘ ATGTAC Semi—nested Primer Pair F15989 R1625] 5’ CCCAAAGCI‘AAGATTCTAAT 5’ GGAGTTGCAGTTGATGT HV1-2 Non-nested Primer Pair F1 6144 R16410 5’ TGACCACCTGTAGTACATAA 5’ GAGGATGGTGGTCAAGGGAC Semi-nested Primer Pair F1 6190 R16410 5’ CCCCATGCTTACAAGCAAGT 5’ GAGGATGGTGGTCAAGGGAC HV2 Non-nested Primer Pair F82 R484 5’ ATAGCATTGCGAGACGCTGG 5’ TGAGATTAGTAGTATGGGAG Semi-nested Primer Pair F155 R484 5’ TATTTATCGCACCI‘ACGTTC 5’ TGAGAT‘TAGTAGTATGGGAG Reactions were performed in 20 uL volumes and included 1 uL 1:10 diluted DNA template, 2 uM forward and reverse primer, 1.5 uL BSA (10 mg/uL), 0.2 uL AmpliTaq Gold® DNA Polymerase, 200 uM dNTPs, 2 uL GeneAmp 10X PCR Buffer II (Applied Biosystems), and 2.5 mM MgC12. PCR conditions included a 12-minute 94°C denaturing step, followed by 38 cycles of 94°C for 30 seconds, 55°C for 1 minute and 72°C for 1 minute, with a final extension step of 72°C for 5 minutes. Five microliters of the PCR product was electrophoresed on a 1.5% agarose gel. If a band was seen but contained 16 insufficient DNA for sequencing, the product was re-amplified with the non-nested primer pair for an additional 10 to 20 cycles. If bands were absent, the sample was amplified for 24 cycles with the corresponding semi-nested primer pair. Reference samples were amplified using primers F15989 and R569 for 32 cycles and the same PCR conditions as above. PCR products were purified with a Montage® PCR Filter Unit as described previously. PCR products were sequenced utilizing a CEO DTCS Quick Start Kit (Beckman Coulter, Fullerton, CA) according to the manufacturer’s instructions in a 10 p.L reaction volume. Sequencing reactions of bomb samples included the same primers that were used to amplify the DNA. Reference samples were sequenced with primers F15989, R16410, F15, and either R484 or R589. Sequencing reactions were purified according to manufacturer’s instructions and analyzed with a CEO 8000 Genetic Analysis System (Beckman Coulter) and a separation time of 60 minutes. Sequences were aligned with the software program Geneious Pro 3.0.6 (Drummond et al., 2007). Determination of genetic profiles and assignations Pipe bomb STR profiles were established by including alleles found in at least two of the three PCR replicates. Further, three peak height threshold values (50 RFU, 100 RF U, and 1000 RF U) were applied and a profile was developed for every bomb. Bomb profiles were compared to those of the subjects, and an individual (or individuals) was chosen as the most likely to have contributed the profile based on the number of loci with the same alleles. The bomb assignations were then compared to the true handler identity to determine correctness. l7 In a separate process using mtDNA, bomb samples were assigned to individuals based upon matching haplotypes. To assess the power of mtDNA and miniSTR analyses to determine the handler, the results were used in combination to make assignations according to the following rules: 1. If a single individual was determined using mtDNA, that pipe bomb was assigned to that individual. 2. If mtDNA assigned the bomb to more than one individual, the list of possible handlers determined from the miniSTR data was consulted, and the individual(s) found to overlap these two lists was chosen. If no overlap existed, the two lists were merged, thereby assigning the bomb to the group created from both lists. 3. If the mtDNA failed to assign a bomb, the list of possible handlers that had been determined fiom the miniSTRs was used. 18 Results Pipe bomb deflagration and DNA isolation Fragmentation levels of the deflagrated bombs varied from low to complete. Steel bombs tended to produce larger, more intact fragments whereas PVC bombs yielded fragments that were much smaller (Figure 1). Eight of the steel bombs resulted in a large fiagment that consisted of about 90% of the original length of the pipe. The other larger fragments of the remaining bombs were between 25% and 75% of the original size of the bomb (exemplified in Figure 1A). All PVC bombs but one were highly fragmented, consisted of mostly small pieces (exemplified in Figure 13), and produced some fragments that were too small to swab (less than one cm), which were not collected. One PVC bomb, 17P, did not fragment; only the end cap was blown off. Figure 1. Post-blast debris. (A) represents the typical pieces collected from a steel pipe bomb, while (B) shows the remains of a PVC pipe bomb. The swabs used to collect DNA from the bomb surfaces also collected residue from the burned powder, and from the steel pipes, flakes of the galvanizing layer, thus most of the swabs were discolored. The organic DNA extraction from PVC pipes removed most, if not all, of the discoloration. This was usually not the case with the steel 19 pipes however; many of the organic extractions resulted in a pink colored aqueous layer, which was not completely removed following filtration on Microcon® YM-30 columns. It was hypothesized that the color resulted fiom a reaction between metal particles and the phenol used in the organic extraction. Preliminary experiments indicated that the DNA from steel pipes amplified as successfully as DNA from PVC pipes, so no further action was taken to either identify or remove the discoloration. MiniSTR profiles and bomb assignation STR profiles were generated utilizing three peak height threshold values: 50, 100, and 1000 RF U. Every bomb had at least one allele that amplified two or more times in the triplicate reactions at 50 and 100 RF U. At a threshold value of 1000 RFU, no alleles amplified more than once from three of the bombs (3 P, 4P, and 138). There were no complete STR profiles produced from any of the bombs at any of the threshold values; each profile had some alleles missing, and often loci did not amplify. All STR profiles exhibited some extraneous alleles (not originating from the handler) that were seen in duplicate reactions, and thus were reproducible. The source of these alleles could not be determined; many of them did not correspond to the investigators participating in the study. However, all but one bomb that exhibited extraneous alleles had at least one allele in common with an investigator, and therefore the investigators could not be excluded as a source of contamination. The bombs, individuals they were assigned to, and the individuals who actually handled them are shown in Table 2. Using a threshold of 50 RFU, 16 (47.1%) of the bombs were correctly assigned (F igtn'e 2 and Table 3), half of which (23.5%) were assigned to either single individuals or a set of individuals, the latter encompassing 20 Table 2. Bomb assignations using miniSTRs. The bombs (P = PVC; S = steel) and study participants (“1ndiv.”) are listed along with the individual or sets of individuals to whom each bomb was assigned. Individuals that were correctly associated with a bomb are indicated in bold. Bomb lndiv. 50 RFU 100 RFU 1000 RFU IP 313 313,920 920 313,485 2P 124 009,177,211,875,920 009,177,211,875,920 211,875 3P 009 4P 307 406 124,211,215, 124,211,215,313,398, 313,398,485, SP 211 211,875 485,522,622,875,920 522,622,875,920 6P 398 398 398 398,622 124,211,215,3l3, 124,211,215, 398,485,522,622, 124,211,215,313,398, 313,398,485, 7P 522 875,920 485,522,622,875,920 522,622,875,920 8? 209 920 920 209,920 9P 736 l77,209,398,406,736 l77,209,398,406,736 177,406,446,736 10? 215 211,485 211,485 211 llP 622 485,875 485 009,211 009,124,177,211,313, 12P 875 009,211,398,485 398,446,485,522,875,920 398,920 13P 920 622,875 875 398,622 14P 446 875 875 209,446,485,875 15P 485 406,485 406,485 211,875 16P 177 177 177 177,485 l7P 406 406 406 211 IS 313 313 313 211,485 zs 124 177,211,209 177,211,215 177,211,313,522 38 009 485,736 211,485,736 211 48 307 307 307 009,307,485 SS 211 211 211 211 68 398 485 485 177,398,485 78 522 875 875 522 88 209 622 622 209,622 211,215,307, 98 736 215,307,398,875 215,307,398,736,875 313,398,209 177,209,211,522, 209,211,622, 108 215 622,875,920 209,211,622,875,920 875,920 118 622 485 485 209 128 875 211,406,875 211,875 211,875 138 920 148 446 398,406,446,485 398,406,446,485 406,446 158 485 485 485 485 168 177 177 177 177,211,875 l77,211,398,406, 177,211,398, 178 406 736,875 177,211,398,406,736,875 406,736,875 21 Table 3. Summery of bomb assignations using miniSTRs. The number of pipe bomb assignations, as well as the percentage is indicated for each category and RFU value. 50 RFU 100 RFU 1000 RFU Correctly assigned to a single person 8 (23.5%) 8 (23.5%) 3 (8.8%) Correctly assigned to a subset of people 8 (23.5%) 9 (26.5%) 15 (44.1%) Incorrectly assigned to a single person 7 (20.6%) 9 (26.5%) 4 (11.8%) 9 2 Incorrectly assrgned to a subset of (26.5%) 5 (14_7%) 9 (26.5%) people Not assigned (5.9%) 3 (8.8%) 3 (8.8%) 2 — 10 people. Sixteen bombs (47.1%) were incorrectly assigned, with seven (20.6%) assigned to a single individual and nine (26.5%) assigned to a subset (2 — 7) of the study participants. Two (5.9%) could not be assigned due to the lack of amplified alleles. Not assigned 5.9% Correctly assigned-single 23.5% Incorrectly assigned-multiple 26.5% Correctly assigned-multiple Incorrectly 23.5% assigned-single 20.6% Figure 2. Correctly and incorrectly assigned pipe bombs at 50 RFU using miniSTRs. Segments of the chart represent the percentage of bombs that were assigned to single or multiple individuals. When the peak height threshold value was raised to 100 RFU, 17 bombs (50%) were correctly assigned, including eight (23.5%) to a single individual and nine (26.5%) 22 to a set of 2 — 11 individuals (Figure 3 and Table 3). Fourteen (41.2%) bombs were incorrectly assigned, with nine (26.5%) to single individuals and five ( 14.7%) to a group of 2 — 5 individuals. Three (8.8%) bombs could not be assigned. Twenty-three bombs had the same classification at both 50 and 100 RFU, whereas six had different classifications. Two bombs, 12F and 9S, were incorrectly assigned to a set of donors at a threshold of 50 RF U, but were correctly assigned to a set of donors using the 100 RFU threshold. Not assigned C orrectl assi ned- 8.8% y g single 23.5% Incorrectly assigned- multiple 14.7% Incorrectly assigned Correctly assigned— singlc multiple 26.5% 26.5% Figure 3. Correctly and incorrectly assigned pipe bombs at 100 RFU using miniSTRs. Segments of the chart represent the percentage of bombs that were assigned to single or multiple individuals. Using a threshold value of 1000 RFU, 18 (52.9%) pipe bombs were correctly assigned, with three (8.8%) assigned to a single individual and 15 (44.1%) assigned to sets of 2 — 10 individuals (Figure 4 and Table 3). Thirteen (38.2%) bombs were incorrectly assigned—four (1 1.8%) to a single individual and nine (26.5%) to a group of 23 2 — 6 individuals. Three (8.8%) bombs could not be assigned. With the increase in threshold vedue from 100 to 1000 RFU, 15 bombs stayed in the same classification whereas 19 changed. Five went from being incorrectly assigned to a single individual to being correctly assigned to a set of individuals ranging in size fiom 2 — 4. One bomb was incorrectly assigned to a single individual at 100 RFU but correctly assigned to a single individual at 1000 RFU. The remainder of the bombs that changed classifications either went from being correctly assigned at 100 RFU to being incorrectly assigned at 1000 RFU, or the number of individuals to which they were assigned changed and the accuracy stayed the same. Correctly Not assigned assigned-single 8.8% 8.8% Incorrectly assigned-multiple 26.5% Correctly assigned-multiple 44.1% Incorrectly assigned-single 1 1.8% Figure 4. Correctly and incorrectly assigned pipe bombs at 1000 RFU using miniSTRs. Segments of the chart represent the percentage of bombs that were assigned to single or multiple individuals. The two control pipes that were deflagrated had alleles amplify. The steel control pipe had 11 reproducible alleles at both the 50 and 100 RF U thresholds and seven at 1000 24 RFU. The PVC control pipe had two reproducible alleles at 50 RFU and one each at 100 and 1000 RF U. These results, like the extraneous alleles observed on the handled bombs, could not be completely attributed to any of the investigators in the study. For example, the steel control pipe produced alleles 8, 9, and 11 at D16S539. One investigator had a 9 allele at this locus; however the other two alleles were not attributable to any of the investigators. MtDNA sequencing and bomb assignation HVl sequences were obtained for all bombs and HV2 sequences were acquired for all but one bomb. Eleven of the 34 bombs were correctly assigned to a single individual (32.4%) and ten others were correctly assigned to a group of two people (29.4%) (Figure 5 and Table 4). Nine bombs could not be assigned because a haplotype was produced that did not match any of the subjects (26.5%). Fewer mis-assignments were made using mtDNA (11.8%) than miniSTRs (38.3% to 44.1%). Table 4. Summery of bomb assignations using mtDNA. The number of pipe bomb assignations as well as the percentage of the total is indicated for each category. Number Percent of bombs Correctly assigned to a single person 11 32.4% Correctly assigned to a subset of people 10 29.4% Incorrectly assigned to a single person 2 5.9% Incorrectly assigned to a subset ofjeople 2 5.9% Not assigned 9 26.5% 25 Not assigned 26.5% Correctly assigned-single 32.4% Incorrectly assigned-multiple 5.9% Incorrectly assigned-single 59% Correctly assigned-multiple 29.4% Figure 5. Correctly and incorrectly assigned pipe bombs using mtDNA. Segments of the chart represent the percentage of bombs that were assigned to single or multiple individuals. One PVC and one steel bomb were swabbed prior to handling by subjects to see if any preexisting DNA that may have been on the bombs was destroyed, and neither bomb generated a mitochondrial PCR product. The two control bombs that were deflagrated each yielded a mtDNA profile. The PVC bomb produced a mixture of haplotypes from unknown individuals: C and T at 16179, 16224 and 16311. The steel control pipe bomb had mixtures of C and T at 199, 204, and 250, which is consistent with the profile of individual 307 and an unknown source. Twenty-four of the 34 experimental bombs showed signs of mtDNA contamination in their profiles. Some of the contamination could potentially be attributed to the investigators, as a mixture of haplotypes that was consistent with originating from both a study participant and an investigator was present. Six of the 24 presumably contaminated bombs were correctly assigned to a single 26 individual and eight were correctly assigned to a group of two individuals, as a minor profile was subtracted. Seven of the nine bombs that were not assigned showed mtDNA contamination. Three bombs had mixtures of haplotypes consistent with having been contaminated prior to deflagration by the investigator responsible for filling the pipe bombs with powder and placing them in the deflagration room. The investigator who performed the remainder of the analyses in the study may have contaminated three samples during the set-up of PCR, as contamination was not found in all three sequences that were produced. Two of these samples, 2S and HP, had a mixture in HV 1, with both a C and T at 16179, 16291 and 16356, which was consistent with coming from the primary investigator and the subject, but no contamination was found in HV2. A DNA sample from bomb 5P showed a mixture of bases only in HV 2, at 309.1C, consistent with the primary investigator. The same investigator may also have contaminated four other samples, however the step when this occurred could not be determined because contamination was found in all three sequences produced, thus contamination could have happened during any step. Fourteen samples contained an extra haplotype or haplotypes that were not attributable to any of the investigators, based on the mixture of bases found. Five of these had only one position with a mixture of bases that was different fiom the subjects’ haplotype. The remainder of the bombs had between two and six positions that had a mixture of bases different than their corresponding reference samples. Bomb assignations using both miniST Rs and mtDNA MiniSTRs and mtDNA were used in combination to assign pipe bombs to the individuals shown in Table 5. At a peak height threshold of 50 RF U, 17 bombs (50%) 27 be" be“ as_ mm4.sb_ sea ss_ n_N.ss_ sac aa_ new.oms ems.mwv amNrmmw mam.n_~ mav.ooe mwv.oos ema.wam mw4 am_ .wam._n~ .aov.wam .eov.mom eve msw.-m.evv mam.-m.e44 new.mwv.evv.oon mew mew warwvv ova av_ ans cue cue «No.wam mew mam.~Ne cue o~a am_ o~a.mam oms.msa mwv.wam oma.wam oaa.msa mw4.wam - mam aN_ .Nmm.mw4.844 .__N.ooo .Nmm.wwv.844 .nn~.ooo .mam.m_m._nm .wam.m_m._cm .ss_.sm_.aoo .san.vmn.soo "Ne Nae «no mwwrmoo nae mam.mwv ”Ne NNe can m_~ m_u m_n _EN mw4.__~ mwv.__N m_~ m_N ao_ cos woe cow ems.844.ooe.ss_ ens.ao4 ens.eoe woe ems ea .wom.aom.s5_ .wsm.ao~.sac as" oaa.aov.a¢~ oma.oo¢.acn oma.ac~ ems oma eoe.ae~ mom aw o~8.msw oma.msw omo.msw oma.msw ommrmsw.mmo.~Nn oma.maw - «mm ma .mma.-m .NN8.~Nm .Nme.anm .mma.-m.mw4.wam .me.»¢m.m_m .mme.-m .mwe.waM.m_m .mw4.wom.m_m .mwe.wom.m_m .mzm.m_N.n_N.4Nc .m_~.~_~.¢m_ .mwe.wam.mnm .m_~.~n~.em_ .m_N.__~.4NE .WEN.E_N.4N_ .m_m._sm.4m_ man man man -8.aam man can ama.wam wam no ~_~ __N __N oma.msm ona.msw.NNo.NNm mam.__~ a_~.ssc __N am .mme.~Nm.mmv.mam .mw4.mam.m_m .m_m.m_~.~a~.em_ .m_~.~_u.v~z - - woe - - woe - saw me - - - - - - - soc am wearaaw omo.msw oma.msw WNwr__N omo.msm omo.msw - van AN .n_~.ss_.aoc .__~.ssn.aco .__N.sa~.aoo .__N.ssc.aoo nan nan m_n mw4.n_m oma omo.n_m mam mcm a_ 58% $E§ 55% adage :bSS admfi e§e%a ignegm neaacmooo_ eeaaamoc_ egaaaMOW