WIWINWIWWIWIHHIWWMIIIMIMI

       

ITHS

THESIS

n1

4.535251%

This is to certify that the
thesis entitled

Predicting Mitochondrial DNA Quality Based on Bone
Weathering and Type

presented by
Lisa M. Misner

has been accepted towards fulfillment
of the requirements for the

 

 

 

 

Master of degree in Forensic Science
Science
Wm Professors Signature/
5 / /J/ 0 C/
/ / ,—
Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY
Michigan State
University

 

 

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIHC/DateDue.p65~p.15

PREDICTING MTDNA QUALITY BASED ON BONE WEATHERING AND TYPE
By

Lisa M. Misner

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTERS OF SCIENCE
Forensic Science

2004

ABSTRACT
PREDICTING MTDNA QUALITY BASED ON BONE WEATHERING AND TYPE
By

Lisa M. Misner

Investigation of human skeletal remains often requires DNA analysis by the
forensic biologist. This study was designed to determine if the degradative state of a
skeleton, bone and/or type of bone can be used to predict the presence of DNA and its
qualitative and quantitative condition, reducing the time and expense of analysis. Eighty-
nine bone samples originating from 36 skeletons, found at an abandoned cemetery in
Pittsburgh, Pennsylvania dating to the mid-1 8008 were tested. QuantiBlot-based
quantification of mtDNA was not successful due to very low DNA yields, thus PCR-
based analysis of mtDNA was used. Amplification of 220 bp fragments was most
successful, followed by 107 bp, then larger fragments. The weathering condition of
whole skeletons and individual bones were not useful in predicting successful DNA
amplification, while sex of the individual may play a role in the preservation of DNA,
with DNA from male skeletons amplifying more often. The best indicator of the quality
of mtDNA obtainable from a bone sample, however, was the type of bone from which
DNA was extracted. Amplification occurred most frequently from femurs, followed
closely by ribs and distantly by pelvises. DNA sequencing affirmed the skeletal origin of
the DNA. The findings indicate that knowledge of bone type is the single most important
factor in predicting success of DNA analysis, while other factors, particularly the

appearance of bone, are of minimal use.

. .- I l 111:.112I. ,.-.- 2. ... .. 2......fll ix... i . . , a .- l..~.c. .w .. .
.- E {I‘li'ilill'ltg 1.1.x... .2“: lu.....ko.lvlu::r...t....blit 2.. tli .....t-t.'.t.- .I..!..”uv rt} if. i .I.. ilLtt.

In.

.. .i U I..: .. d: I. . .
.... L 3". ... 2...?113

i
l
2
l
a
i
{i-
E-

n . . .
,ri.tII I‘ll... .(tl. t C.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. v
LIST OF FIGURES ................................................................................. vi
INTRODUCTION ................................................................................... 1
Bone Degradation ............................................................................. 2
DNA Degradation in Skeletal Remains .................................................. 5
Description of the Voegtly Samples ........................................................ 9
Research Goals ............................................................................. 13
METHODS AND MATERIALSI4
Summary of Samples .......................................................................... 14
Preparation of Bones for Subsequent DNA Extraction ............................... 15
DNA Extraction ............................................................................ 19
DNA Quantification Using QuantiBlot ................................................. 19
DNA Amplification of Bone Samples ..................................................... 20
DNA Sequencing of Bone Samples ........................................................ 22
Statistical Methods ......................................................................... 23
RESULTS ............................................................................................ 24
Sampling .................................................................................... 24
Observations during Bone Preparation .................................................. 24
DNA Extraction ............................................................................ 25
DNA Quantification Using QuantiBlot ................................................. 26
DNA Amplification .......................................................................... 28
Skeletal Weathering and Amplification Success. 35
Individual Weathering and Amplification Success .................................... 36
Bone Type and Amplification Success .................................................. 38
Sex Estimates and Amplification Success .............................................. 39
Age and Amplification Success .......................................................... 42
DNA Sequencing ............................................................................. 43
DISCUSSION ....................................................................................... 46
DNA Pellet Color ............................................................................. 46
DNA Quantification .......................................................................... 46
DNA Amplification ........................................................................ 48
DNA Quality ................................................................................ 50
DNA Sequencing ........................................................................... 51
Weathering Condition and Amplifiable DNA .......................................... 52
Bone Type and Amplifiable DNA ....................................................... 59
Sex and Amplifiable DNA ................................................................ 60
Age and Amplifiable DNA ............................................................... 61
Conclusions ................................................................................. 63

iii

REFERENCES ................
........................................................... 66

iv

LIST OF TABLES
Summary of Bones Analyzed ........................................................................ 17
MtDNA Probes Used in QuantiBlot Procedures ................................................ 20
MtDNA Primers Used for Amplification and Sequencing. 21
Improvement in Inhibition ......................................................................... 31

Overall PCR Results ................................................................................ 33

Sequence Polymorphisms In Individuals Generating PCR products form More Than One
Bone ................................................................................................... 45

LIST OF FIGURES

Microscopic Comparison of Compact and Spongy Bone at 100x .......................... 5
Excavation Site and Site Map of Grave Locations of the Voegtly Cemetery ............ 12
Quantification of mtDNA Using QuantiBlot .................................................. 28
Typical Amplification Results: Positive, Negative, and Inhibited ......................... 29
Amplification Results in Comparison to Skeletal Weathering Stage ...................... 35

Amplification Results in Comparison to Skeletal Weathering Stage and Bone Type. . .36

Amplification Results in Comparison to Individual Bone Weathering Stage ............. 37
Amplification Results in Comparison to Individual Bone Weathering Stage and Bone
Type ................................................................................................. 38
Amplification Results in Comparison to Bone Type ......................................... 39
Amplification Results in Comparison to Sex of the Individual ............................. 40
Amplification Results in Comparison to Estimated Age .................................... 42
Electropherogram Depicting Typical 5’ End Data ........................................... 44

vi

Introduction

An important goal of any forensic investigation involving unidentified human
remains is positive identification. Forensic specialists including anthropologists and
odontologists may evaluate the remains, estimating the individual’s sex, stature, age at
death, and ancestry. Any identifiable characteristics such as fingerprints, if flesh is still
intact, unique skeletal features and dental arrangement are also noted. However, when
antemortem reference records are not available for comparison or remains are fragmented
or otherwise in a state in which definitive conclusions cannot be made as to the person’s
identity, DNA analysis may be required. Indeed, analysis of DNA from human skeletal
remains as a means of identification has been used in numerous cases, beginning in 1989
(Paabo et al., 1989). Advancements in techniques and applications occurred in the
1990’s (Boles er al., 1995; Hagelberg et al., 1991; Primorac et al., 1996; reviewed in
O’Rourke et al., 2000), and in 1991 the Armed Forces DNA Identification laboratory was
established for the identification of the remains of US. military personnel (Holland and
Parsons, 1999). Ten years later a high-throughput method was developed for the analysis
of the thousands of bones connected with the attacks on the World Trade Center (Holland
et al., 2003). Bones encountered by the forensic biologist often vary in their degree of
degradation, and if more than one bone is available, the scientist may choose which bone
to analyze based on its appearance. Although bone degradation and DNA degradation
are separate processes, any relationship between the two has not been well defined.
Knowledge about such a relationship would help the forensic biologist predict the success
of analyses involving skeletal remains. Currently, this proceeds through a series of trial-

and-error processes until results are obtained or the samples are abandoned, potentially

wasting laboratory resources and the analyst’s time. Knowing the likelihood of obtaining
usable data based on the type and condition of the bone would be beneficial to the

forensic biologist as it would expedite the analysis procedure and reduce costs.

Bone Degradation

Though the structure and composition of skeletal material allows it to persist long
after soft tissue has decomposed, it is still subject to degradative processes. Several
studies have addressed the issue of bone degradation and the factors that affect it. At the
microscopic level, Shackleford (1996) analyzed the effects of soil condition and
geological age of the skeleton on the histology of seven femurs from Native Americans
found in Alabama. Five of the femurs were excavated from soil described as a mixture of
sand, silt, clay, and shell mound, and were estimated to have been buried during the
Mississippian period (500—1000 years ago). Two femurs were estimated to have
originated from the Archaic period (3000—5000 years ago); one was removed from a dry,
protected bluff, the other was surrounded by a shell mound. Sections of each femur were
analyzed using electron microscopy and electron diffraction. All seven samples showed
intact mineral constituents similar to that of modern bone, indicating that neither the age
of the skeleton nor the composition of the surrounding soil greatly affected the
preservation of histological structure.

Solomon and Hasse (1967) analyzed the histology of bone from sites in Israel that
varied in age, soil composition, and moisture content. A total of six bones, two humeri,
three femurs, and one metatarsal were analyzed originating from 3500 BC. to 200 AD.

The bones were excavated fi'om one of three different soil types: sand/clay from a humid

climate; loess (fine-grained deposit of wind blown dust), lime, and salts from a dry
climate; and clay, granite, powder, and salts from a dry climate. Sections of each bone
were examined using various stains to compare the histological structure of these samples
to fresh bone. Results indicated that histological structure did not vary among bones
from the same soil condition regardless of age, further confirming that the time since
death of the individual has little effect on bone preservation. However, a difference was
noted among bones removed from soils of similar composition but differing humidity.
The histological structure of samples from humid climates showed a greater degree of
degradation than those of from dry climates, indicating that exposure to moisture had a
negative effect on bone preservation.

In an overview of the processes of bone diagenesis, Hedges (2002) summarized
the histological and biochemical changes that occur as a bone degrades. One correlation
observed was with loss of collagen, the porosity of the bone decreased and the
crystallinity increased. Hedges noted that microbial attack (both bacterial and flmgal), as
determined by histological modification (e. g. re-configuration of the mineral phase), was
the major cause of collagen loss, with severely attacked bone having lost at least 80% of
its original collagen. In addition, low temperature was found to prevent microbial attack,
and extreme hydrology (either very wet or very dry) also inhibited microbes. Analyzing
the effects of soil conditions, soils of neutral pH had calcium and phosphate
concentrations similar to those found in hydroxyapatite (Calo(PO4)6(OH)2; the major
bone mineral), which led to a slower dissolution rate. In contrast, soils of low pH had
lower calcium and phosphate concentrations and dissolution occurred faster, as the

protons from the soils replaced the calcium ions in the hydroxyapatite of the bone.

At a macroscopic level, Micozzi (1991) summarized the results of several studies
which indicated that the survivability of a bone increases with increasing density, and that
various factors such as age, size, and bone type affect density. With respect to age, there
is a biphasic relationship to bone density, with young and very old individuals having less
dense bones. The density of bone is related to its microstructure. There are two main
microstructures of bone, compact bone and cancellous (spongy) bone (Figure 1).
Compact bone is composed of a system of osteons in which compact concentric layers
(lamellae) of hydroxyapatite and collagen are arranged around a central Haversion canal.
Between the layers are small spaces (lacunae) where the mature bone cells (osteocytes)
reside. Spongy bone consists of trabeculae (spicules) of bone separated by large open
spaces, and osteocytes are found within lacunae of the trabeculae. Every bone contains
both compact and spongy bone structure though the distribution of each varies depending
on bone type. For example, all bones have compact bone along the cortex, however, in
long bones, such as the femur, fibula, tibia, humerus, radius, and ulna compact bone
continues throughout the thickness from the cortex to the medulla all along the shaft of
the bone, while trabecular bone is present beyond the cortex at the articular ends. In
contrast, flat bones, such as the rib and pelvis, have spongy trabecular bone beyond the

cortex in all portions of the bone.

Figure 1. Microscopic comparison of compact and spongy bone at 100X.

    

    

Compactbnegis'fr’epresented onttii’eleft L=lainEIIa, HCéHaversion canal, La=laguna.
Spongy is represented on the right. T=trabecula. (Figures taken from McCutchen, 2003)
Pastron and Clewlow as cited by Micozzi (1991) ranked the survivability of the
long bones removed from prehistoric human burials found within caves and rock shelters
of a rock wall in New Mexico. They found that the femur was the most durable,
followed by the tibia, fibula, humerus, ulna, and radius. Such conclusions indicate that
surface area of a bone plays a role in its survivability; and indeed Gill-King (1997)
described surface area as an “important factor” in the rate of bone degradation. Bones
with a larger surface area to volume ratio have more contact with the surrounding soil,
making them more susceptible to ion exchange and the actions of the microorganisms

within.

DNA Degradation in Skeletal Remains

As skeletal material undergoes the process of degradation, the organic molecules
within the bone, including DNA, also degrade (Gotherstrom et al., 2002). Studies have
been carried out to analyze the degradation of the DNA molecule in bone under various

conditions (Parsons and Weedn, 1996; Smith et al., 2003). High temperature, high

moisture, and acidic pH accelerate DNA degradation. In addition, chemical
modifications occur, such as oxidation, single-strand breaks, hydrolysis, and cross-
linking from free radical attack and UV damage. Microorganisms also affect the stability
of DNA, producing enzymes such as endo- and exonucleases which directly attack DNA
(Rogan and Salvo, 1990). As long as soft tissue persists, additional enzymatic attack can
occur from autolytic enzymes.

Although the factors that contribute to the degradation of skeletal material and
DNA have been studied individually by numerous researchers, the way in which the two
might correlate is not well understood. A limited number of studies have been carried out
however, which address the issue (Gotherstrom et al., 2002; Tuross, 1994; others
reviewed in Kaestle and Horsburgh, 2002; Parsons and Weedn, 1997; and O’Rourke et
al., 2000). In general, researchers have found that the likelihood of obtaining usable
DNA from a sample can be predicted, to a degree, fiom the gross morphology of the
sample, with harder, more intact bones providing usable DNA more often than softer,
more brittle bones (Kaestle and Horsburgh, 2002; and Parsons and Weedn, 1997 and
references within). In addition, a correlation was found between the microscopic
preservation of bone samples and DNA recovery, with usable DNA being present in
bones with better preserved microstructure (O’Rourke et al., 2000).

Tuross (1994) addressed the issues of time since death and burial conditions by
analyzing three collections of bone: a set of three human femurs obtained during surgical
procedures; bone samples originating from elephant and wildebeest, which were
collected once a year for fifteen years after the animals’ deaths; and a variety of human

skeletal collections from North and Central America, which had been buried up to 7000

years. DNAs extracted from the samples were quantified by ethidium staining in 4-20%
acrylamide gels and compared to modern samples. Results showed the largest difference
in DNA content occurred between bones that had been buried immediately (the human
skeletons) and those which were exposed to surface conditions (the elephant and
wildebeest). Although the unburied samples contained nearly one tenth the amount of
DNA of the buried samples, the amount of DNA from the unburied samples remained
fairly stable as the years passed, thus suggesting DNA is more stable once the deceased is
fully skeletonized. The author proposes this is due to binding of the DNA to
hydroxyapatite, which would protect the DNA from hydrolysis. It should be noted that
since the DNA extract was quantified via gel analysis, it is unclear whether or not the
total amount of DNA extracted was indigenous to the individual, as bacterial and/or
fungal DNA may have been coextracted. In addition, comparisons were made among
bone samples of three different species, each likely having different microstructures, not
to mention the large difference in time since death and in climate, each of which
represents a variable that should be considered when evaluating these results. Several
studies since Tuross’ have shown that environmental conditions play a larger role than
time since death in relation to the quality of DNA obtainable from bones (as reviewed by
Kaestle and Horsburgh, 2002)

Gotherstrom et al. (2002) focused on finding a reliable, fast, and inexpensive way
of predicting the preservation of DNA in ancient bones and teeth. Two sets of bones
were utilized: one modern, experimentally degraded set of bovid bones and one set of
ancient horse bones; both sets were pulverized for analysis. Three measurements were

taken from the bones: the crystallinity index for hydroxyapatite as determine by X-ray

diffraction, the presence of mitochondrial DNA (mtDNA) as detected by the polymerase
chain reaction (PCR), and the amount of extracted collagen. The results of their study
suggested a correlation between crystallinity of the hydroxyapatite and the presence of
DNA, with amplifiable DNA decreasing as crystallinity increases, thus supporting
Tuross’ (1994) suggestion that DNA stability was related to hydroxyapatite binding. A
similar relationship between the presence of DNA and the amount of collagen left in the
bone was also noted. Based on the research of other scientists, the authors hypothesized
that DNA degradation may be related to mineral sorption, and that once hydroxyapatite
no longer protects the DNA molecule, it is vulnerable to degradative processes.

Although studies such as these have begun to address the correlation between
bone degradation and the amount of DNA degradation, they do not provide the large-
scale, statistically testable data necessary to draw significant conclusions. The majority
of these studies were performed with small sample sizes, comparing samples from
different geographic locations, species, and time since death. Most of the skeletal
material used in these studies was of animal origin, and in some cases, weathering
conditions were mimicked in a laboratory setting. While these studies were helpful in
raising new questions and providing suggestions for a relationship between bone and
DNA degradation, a single, large, aged set of skeletal remains is needed to solidify any
correlations.

As described by Gotherstrom et al., 2002 and other authors (Gilbert et al., 2003;
Kalmar et al., 2000; Paabo er al., 1989), mitochondrial DNA (mtDNA) analysis is the
primary mode of DNA typing of aged skeletal remains. Mitochondria are cytoplasmic

organelles containing a small, maternally-inherited, circular piece of DNA distinct from

the linear chromosomes of the nucleus. Thousands of mitochondria can exist in a single
cell, and as such there are thousands of copies of mtDNA in each cell. This is far
different then the two copies (maternal and paternal) of a DNA locus that are analyzed in
most forensic analyses. Human mtDNA was first sequenced in its entirety in 1981
(Anderson et al.) and is approximately 16569 base pairs (bp) in length. The majority of
genes in mtDNA encode proteins essential to mitochondrial function, thus these areas are
highly conserved among individuals, and are generally not useful for identification. An
approximately 1125 bp stretch of non-coding DNA known as the control region contains
two regions which mutate frequently and therefore vary among individuals. The
sequence variation of these two areas, known as hypervariable region I and H (HVI and
HVII), are generally the focus of forensic analysis (reviewed by Holland and Parsons,
1999). Typically, this involves amplification of both hypervariable regions using PCR
followed by a sequencing procedure to determine the exact DNA sequence, which is then
compared to the Anderson reference sequence. The nucleotide differences between the
sample and the reference sequence constitute the mtDNA profile for that sample, which

can then be compared back to a known sample or a maternal relative for positive

identification.

Description of the Voegtly Samples

Bones for this study were obtained from Dr. Doug Ubelaker of The Smithsonian
Institution and sections collected by Jennifer Drier of George Washington University. As
described by Ubelaker et al. (2003), the skeletal material originated from a burial site,

known as the Voegtly Cemetery, in what is now northern Pittsburgh, Pennsylvania. In

1787, an area north of the Alleghany River was designated as Old Alleghany Town.
During the early to mid-1800’s the area was settled mainly by Swiss-Gennan immigrants.
A prominent millwright family, by the name of Voegtly, together with their Swiss
neighbors, bought 161 acres along the Alleghany River in Old Alleghany Town in 1822.
In 1833, after a donation of land from the Voegtly family, The First Evangelical Church
of Alleghany (Voegtly Church) was constructed. The adjacent cemetery was utilized
from that point until 1861, when land for a larger cemetery was acquired atop the nearby
Troy Hill, just east of the original cemetery. A series of expansions and modifications to
the Church occurred over the next century, and in the process, the original cemetery,
without permanent markers, was built over.

From 1881 to the early 1900’s the church experienced a fluctuation in
membership, with continuous decline from the 1950’s on. In 1959, it was announced that
the Voegtly Church was in the right-of-way for the new East Street Valley Expressway
(I-279/I-579). When the last surviving contact with the Voegtly family died in 1972, the
parsonage was destroyed, and in November of 1984, the Pennsylvania Department of
Transportation acquired the property. In May of 1985 the congregation filed for
dissolution and the Voegtly Church was officially disbanded and destroyed the same
year. Two years later, in the spring of 1987, the company in charge of construction
monitoring discovered five coffins and the associated skeletal remains. A full-scale
archaeological excavation was carried out from June to September of 1987 (a site map of
the grave locations is shown in Figure 2), in which all remaining skeletal material was
removed, recorded (each burial was assigned a consecutive identification number upon

discovery), packaged, and taken to The Smithsonian Institution for processing.

10

Data collection at the Smithsonian began in 1993 following the Standards for Data
Collection from Human Skeleton Remains (Buikstra and Ubelaker, 1994). At the
laboratory the remains were unpacked, cleaned, and estimations of sex, age, and living
stature, were recorded when possible. Estimates of sex were made by evaluating cranial
and pelvic morphology when possible. When these more reliable features were not
available or in poor condition, skeletal robusticity or presence of particular artifacts found
with the remains (e. g. earrings, remnants of a moustache) was used. In general, age was
estimated by eruption of teeth and dental wear, and stature was estimated by femur
length. In addition to these estimates, the degree of weathering was assessed for each
individual burial based on a classification system taken from Behrensmeyer (1978). The
system is based on a scale from 0 to 5 and briefly includes:

Stage 0: Bone surface shows no signs of cracking or flaking due to weathering.

Stage 1: Bone shows some cracking, usually longitudinal in long bones.

Stage 2: Some cracking and flaking is apparent, especially on the outermost

concentric thin layers of the bone.

Stage 3: Bone surface has rough patches of weathered compact bone; external

concentric layers have been removed, but weathering does not penetrate deeper

than 1.0-1.5 mm.

Stage 4: Bone surface is coarse and splinters may exist; weathering reaches into

inner cavities.

Stage 5: Bone has large splinters and is easily broken; original bone shape may

be undeterrninable.

11

A detailed record of the data collected at the Smithsonian is available in Ubelaker et al.
(2003).

Figure 2. Excavation Site and Site Map of Grave Locations of the Voegtly Cemetery

   
   

   

. u. '_ 11A.
I __ ‘. - 9 ~

(Figures taken from Ubelaker et al., 2003) 1

According to the church records, which were translated from their original
German text, 823 individuals were buried in the original Voegtly Cemetery between 1833
and 1861. Not all of these records contained data on name, sex, or age. In most cases,
this lack of information corresponded to records from infants and children. Sixteen of the
records with missing information were marked as “not seen”, probably indicating a child
that was miscarried or stillborn. Of the 799 records in which age was recorded, 311
(38.9%) were less than 1 year old, 192 (24%) were between 1 and 5 years, and 66 (8.3%)
were between 5 and 20 years of age. The sex of 774 individuals was recorded, 365
(47.2%) were females and 409 (52.8%) were males.

Of the 724 individuals recovered from the abandoned cemetery, 208 bones from
88 individuals were sampled for analysis. Two small sections (approximately two to four
cm3) were removed from each bone. One set of sections was sent to Brazil for
histological examination; the other set was reserved for DNA analysis at Michigan State

University.

Research Goals

The goal of this study was to utilize the large set of skeletal remains recovered
from the Voegtly Cemetery to determine if a correlation exists between bone type, its
level of degradation, and the qualitative and/or quantitative condition of the DNA within.
The specimens from the Voegtly Cemetery provide an ideal sample set for a controlled
study. Unlike previous studies where comparisons were made between bones of varying
age, species, and geographical regions, and sample sizes were small (e. g., Gotherstrom et
al., 2002; Tuross, 1994), these samples provide a large sample set in which these
variables are controlled. Being buried in the same cemetery, the remains were likely to
be exposed to the same soil environment, and through knowledge of church records, all
bones had been exposed to this environment for roughly 150 years. Also, the history of
the church and its members indicate that the majority of the individuals buried in the
cemetery were Swiss-German immigrants, thus making ethnicity a controlled variable as
well.

Skeletal remains are the most likely type of sample requiring analysis of the
forensic biologist well after an individual’s death. Currently, multiple days are needed to
assess if usable DNA is obtainable from a particular sample. Without prior knowledge of
the likelihood of successful analysis of skeletal material, the scientist may be required to
repeat the procedure multiple times on different bones (or portions of) until results are
obtained. Being able to make predictions of the amount and quality of DNA from aged
bone based on the type of skeletal material (long bone, flat bone, short, or irregular) and
its level of weathering, would minimize time and expenses spent on analysis and

maximize efficiency.

13

Materials and Methods
Summary of Samples

A subset of 89 bones from the total set obtained from The Smithsonian Institution
was analyzed in this study. Table 1 lists the bones analyzed by the burial number
assigned during their recovery. Each bone is described based on information obtained
from The Smithsonian Institution, including sex, estimated age, and one of five
weathering stages (titled “skeletal weathering stage” herein). As detailed above and by
Ubelaker et al. (2003), the weathering stages originally assigned to the samples by the
anthropologists at The Smithsonian considered the condition of the skeleton as a whole.
In addition to this, a new staging system was developed at MSU and assigned to each
bone based on visual inspection for the DNA study. This individual bone staging system
is separated into four categories:

Stage 1: Bone surface shows minimal flaking. The bone piece is still whole, with

no large pieces broken off.

Stage 2: Bone surface shows some flaking, pieces of bone are coming off in
sheets. Small pieces of bone are breaking away from the cut piece.

Stage 3: The bone is fragmented into several pieces. At least one large piece is

still present.

Stage 4: The bone is extensively fragmented. No large pieces are present.

The bones used in this study spanned five of the six weathering stages assigned by
Ubelaker et al. (2003): 6 from stage 1, 23 from stage 2, 22 from stage 3, 18 from stage 4,
and 20 from stage 5. The three samples from the single individual classified as stage 0

were not tested due to the insufficient sample size and nonexistence of comparison

14

samples. A total of 2 crania, 28 femurs, l fibula, 25 pelves, and 33 ribs were tested,
coming from a total of 36 individuals. Of these, 11 were female, 17 were male, and l
was labeled as “possibly male”. This individual had large mastoid processes and
supraorbital ridges, but female sex could not be ruled out. With seven samples sex could

not be determined.

Preparation of Bones for Subsequent DNA Extraction

Approximately 1 cm3 of bone was cut from the source section using a Dremel
MultiPro tool (model no. 395-7 6) with cut-off wheel attachment number 409. Cut pieces
were collected in a 17 X 100 mm polypropylene tube (FisherScientific) and immersed in
1-2 ml of filter-sterilized wash buffer (1% SDS, 25 mM EDTA, filter sterilized using a
0.2 pm syringe filter (FischerScientific)) and 0.1 mg/ml proteinase K and incubated for
one hour at room temperature. Following the incubation, the wash buffer was poured off
and each sample was washed with 1ml of sterile deO six consecutive times. Excess
water remaining in the tube was removed with a sterilized Pasteur pipette. Samples were
dried using compressed air passed through a 0.45 pm vent filter (Millipore) for 15—30
min.

Dried samples were ground to a powder using an IKA A11 Basic Grinder with a
tungsten blade (IKA Works, Inc.). The mass of each sample was recorded both before
and after grinding, and ground samples were collected in a 2 ml microcentrifuge tube.
The grinding mill was cleaned with 10% bleach followed by 70% EtOH after the
grinding of each sample and exposed to UV light at 2500 Joules/cm2 when the next

sample was fi'om the same burial number, and thus the same individual. Between

15

Table 1. Summary of Bones Analyzed
Headings denote burial number, bone type, age, sex (male (M), possibly male (M?), or
female (F)), skeletal weathering stage (0—5), and individual bone weathering stage (1—4).

16

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Burial Bone Skeletal Bone Weathering
Number Type Age Sex Weatherigq Stage Stage
027 Rib 22 M 5 3
027 Fern ur 22 M 5 1
030 Femur 40-55 M 5 3
034 Pelvis 35-45 M 4 2
034 Rib 35-45 M 4 4
034 Femur 35-45 M 4 1
047 Pelvis 11 F 4 2
047 Rib 11 F 4 4
047 Rib 11 F 4 1
047 Fem ur 1 1 F 4 2
1 1 1 Pelvis 30-35 M 2 1
1 1 1 Rib 30—35 M 2 3
1 1 1 Femur 30-35 M 2 1
114 Femur 10.5 F 5 2
124 Pelvis 28-35 M 3 3
124 Rib 28-35 M 3 4
124 Femur 28-35 M 3 2
126 Femur 25-30 F 3 1
126 Pelvis 25-30 F 3 3
126 Rib 25-30 F 3 3
132 Rib 25-30 M 2 3
132 Femur 25-30 M 2 1
1 32 Pelvis 25-30 M 2 2
164 Femur 22-26 M 3 1
164 Rib 22-26 M 3 3
164 Pelvis 22-26 M ° 3 2
167 Rib 15-16 M? 2 2
167 Pelvis 15-16 M? 2 2
167 Femur 1 516 M? 2 1
192 Rib 60-80 M 2 1
192 Pelvis 60-80 M 2 2
192 Femur 60-80 M 2 2
203 Fibula 1 .9mos. 1 2
203 Cranium 1 .9mos. 1 2
203 Rib 1.9mos. 1 2
256 Pelvis 35-45 M 3 2
256 Rib 35-45 M 3 4
256 . Femur 35-45 M 3 3
260 Rib 3 4 2
260 Femur 3 4 2
322 Femur 20-24 M 2 2
322 Rib 20—24 M 2 2
322 Pelvis 20-24 M 2 2
328 Pelvis 40-45 M 2 2
328 Rib 40-45 M 2 2
328 Femur 40-45 M 2 1

 

 

 

 

 

 

l7

 

Table 1. Continued

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

331 Rib 15 M 5 3
331 Femur 1 5 M 5 1
345 Rib Adult M 3 4
345 Femur Adult M 3 1
345 Rib Adult M 3 2
345 Pelvis Adult M 3 3
348 Rib 27-35 M 3 4
348 Femur 27-35 M 3 1
348 Pelvis 27-35 M 3 3
349 Femur 5.8 5 3
355 Rib 4 5 4
355 Femur 4 5 3
381 Femur 25-30 M 2 1
381 Rib 25-30 M 2 2
381 Pelvis 25-30 M 2 3
389 Rib 5 3
389 Pelvis 5 3
402 Pelvis 25-40 F 5 3
4098 Femur 30-40 M 3 2
4098 Rib 30-40 M 3 3
409B Pelvis 30-40 M 3 3
447 Rib 30-40 F 4 3
447 Femur 30-40 F 4 2
448 Femur 25-35 F 5 2
448 Rib 25-35 F 5 4
489 Pelvis 1.5 2 1
489 Rib 1.5 2 1
529A Rib 18-25 F 4 2
529A Femur 18-25 F 4 1
529A Pelvis 18-25 F 4 2
539 Rib 30-40 F 4 4
539 Pelvis 30-40 F 4 2
540 Rib 32-45 F 5 4
540 Pelvis 32-45 F 5 2
545 Femur 25-32 F 1 2
545 Pelvis 25-32 F 1 3
545 Rib 25-32 F 1 4
546 Pelvis 18-21 F 4 2
546 Rib 18-21 F 4 3
686 Cranium 4-5.5 5 2
704 Femur 45-60 M 5 2
704 Pelvis 45-60 M 5 1
704 Rib 45—60 M 5 3

 

l8

 

 

samples of differing burial numbers, the mill was disassembled and with 10% bleach
followed by 70% EtOH to remove any trapped bone dust, then reassembled and exposed

to UV light at 4167 J/cmz.

DNA Extraction

Four hundred microliters of digestion buffer (20 mM Tris, 100 mM EDTA, 0.1%
SDS) and 0.4 mg/ml proteinase K was added to each ground bone sample and incubated
overnight at 56°C. After incubation, one volume of phenol was added, each sample was
vortexed then centrifuged at 14000 rpm at 4°C for 5 min. The aqueous layer was
removed to a clean microcentrifuge tube. If the phenol layer appeared dark brown, 3
second or third phenol extraction was performed. Following the phenol extraction, one
volume of chloroform was added, the sample was vortexed and centrifuged as above, and
the aqueous layer was removed to a new microcentrifuge tube. DNA was precipitated
with 40 ul of 3 M sodium acetate and 800 pl of 95% EtOH and incubated at -20°C for at
least one hour. Samples were centrifuged at 14000 rpm at 4°C for 20 min, after which the
supernatant was removed and the pellets were vacuum-dried for approximately 20 min.
Each sample was resuspended in TE buffer (10 mM Tris, 1 mM EDTA), using 1 til/mg of

ground bone, and stored at -20°C.

DNA Quantification Using QuantiBlot
To determine the quantity of mtDNA extracted from each sample, the QuantiBlot
Human DNA Quanitification Kit (Perkin Ehner) was used with modifications. Three

biotinylated probes were created to target human mtDNA (Table 2) (Promega) in place of

19

the QuantiBlot D17Zl Probe provided in the kit, which targets nuclear DNA.
Optimization of the mtDNA probes led to the use of 30 pl of the individual probe (100
pmol/ul) for each membrane. In addition, the hybridization temperature was brought
down from 50°C to 40°C and all other steps were performed at 24°C. One microliter of
10 ng/ul K562 DNA High Molecular Weight (Promega) was used as a positive control.
Raccoon (Procyon lotor), red fox (Vulpes vulpes), badger (T axidea taxus), and salmon
(Oncorhynchus sp.) DNA were used to test for species specificity. The DNA
concentration of these samples were estimated by running them next to 10 ng of cell line
DNA on a 1% agarose gel and then diluted to a concentration similar to the cell line
source. Thirty microliters of bone extract was loaded onto the membrane. All other steps
were followed as described in the QuantiBlot kit protocol; the chemiluminescent method
of detection was employed.

Table 2. mtDNA probes used in QuantiBlot procedures

 

 

 

 

 

 

Name Sequence

DLoopl 1F 5’ CT ATCACCCTATTAACCACTC 3’
DLoop31R 5’ GAGTGGTTAATAGGGTGATAG 3’
C01 6281F 5’ AACAGTCTACCCTCCCTTAG 3’

 

F =forward, R=reverse, the numbers correspond to the position of the 5’ base of the probe
in the complete human mtDNA sequence (Anderson et al., 1981), and DLoop and C01

(cytochrome c oxidase complex 1) refer to the region of the mtDNA to which the probe is
targeted.

DNA Amplification of Bone Samples

Preliminary amplification reactions contained 1 U Taq DNA polymerase
(Promega), 0.2 mM each of dNTP (Promega), 2 umol each of mtDNA-specific primers
(Genosys) (Table 3), 10X PCR buffer (Promega), 10X MgClz (Promega), and 1 ul of

sample DNA in a total volume of 20 pl. In early experiments, two concentrations of

20

 

sample DNA were used, one undiluted and one 1:20 dilution of the bone extract. As
experiments progressed, only the 1:20 dilution was used. Also, as the experiments
proceeded, HotMaster Taq (Eppendorf), along with the provided buffer, was used instead
of standard Taq, In addition, 5X HotM Enhancer (Eppendorf) was included in later
reactions to reduce PCR inhibition. The amplification reaction consisted of denaturation
at 94°C for 2 min, followed by 38 cycles of denaturation at 94°C for 30 s, annealing at
56°C for 45 s, and extension at 72°C for 45 s. The last cycle was followed by an
additional extension step at 72°C for 5 min. Fragment size and amplification quality were
visualized by running 5 pl of the PCR product, or amplicon, on a 2% agarose gel
followed by staining with ethidium bromide.

Table 3. mtDNA primers used for amplification and segrencing

 

 

 

 

 

 

 

 

 

 

 

Name Sequence Region of Size of product
mtDNA

F155 5’TATTTATCGCACCTACGTTC 3’ HVII 329 bp
R484 5’TGAGATTAGTAGTATGGGAG 3’ HVII

F82 5’ATAGCATTGCGAGACGCTGG3’ mm 402 bp

F1 6144 5’TGACCACCTGTAGTACATAA 3’ HVI 107 bp
R16251 5’GGAGTTGCAGTTGATGT 3’ HVI

F 16190 5’CCCCATGCTTACAAGCAAGT 3’ HVI 220 bp
R16410 5’GAGGATGGTGGTCAAGGGGAC 3’ HVI

 

 

 

F =forward, R=reverse, the numbers refer to the position of the 5’ base of the primer in
the complete human mtDNA sequence. Headings denote the name of the primer it was
assigned when designed, the exact sequence, the region of mtDNA it targets, and the size
of the amplicon resulting from the paired forward and reverse primers.

Initially, PCR was performed using F16190 and R16410 for each sample. If a
sample did not produce the 220 bp amplicon, F16144 and R16251 were used, to produce
a 107 bp amplicon; however, if amplification of the 220 bp amplicon was successful,

amplification of progressively larger amplicons (329 bp followed by 402 bp) was

attempted. When needed, an additional amplification reaction was performed to increase

21

 

the amount of product to be used in sequencing reactions. In these reactions, 1 pl of the
original PCR product was used as a template and the number of cycles was reduced to 20.
If a reaction appeared to be inhibited (there was no amplification band and no
primer dimer formed), the sample was purified using a Microcon-100 spin column
(Millipore) with 300 pl of TE buffer at the recommended speed and time. The samples
were washed two more times with TE buffer then the retenate was resuspended to the

starting volume using TE.

DNA Sequencing of Bone Samples

When amplification was successful, the remaining 15 pl of PCR product was
purified using a Microcon—30 spin column (Millipore) with 300 pl of TE and centrifuged
per the manufacturer’s instruction. The samples were washed two additional times with
TE before retentate was brought back to 15 pl with TE.

Up to 5 pl of template DNA was sequenced using the BeckSeq kit (Beckman-
Coulter). Per the manufacturer’s recommendations, 50—100 femtograrns were used for
sequencing reactions; estimations of concentration were made from the yield gel of the
original amplification. The primers used for sequencing were the same as those used for
amplification, and as a modification to the protocol provided with the kit, the total
reaction volume was reduced to 10 pl. The sequencing reaction was followed as outlined
in the kit manual. Following sequencing, 4 pl of stop solution (1.5M NaAc, 50 mM
EDTA, 2.5 mg/ml glycogen) and 30 pl of cold 95% EtOH was added to each reaction,
which was then centrifuged at 14000rpm for 15 min. The supernatant was removed and

the sample was washed twice with 200 p1 of cold 70% EtOH with a 2 min, 14000rpm

22

spin between washes. After the final wash, the supernatant was removed and the sample
was vacuum-dried for approximately 10 min, then resuspended in 40 p1 Sample Loading
Solution (SLS) (Beckman-Coulter). Samples were sequenced on a CEQ 8000 Genetic
Sequence Analyzer (Beckman-Coulter), using the LF R-1-60 program (capillary
temperature 50°C, denature 120 s at 90°C, inject 15 s at 2.0kV, and separate 60 min at
4.2kV). Resulting sequences were aligned using BioEdit Sequence Alignment Editor
(Hall, 2004) and compared to the Anderson reference sequence. Any polymorphisms
within a sequence were entered for comparison to the FBI mtDNA database to determine

the frequency of the specific base change(s) (Monson et al., 2002).

Statistical Methods

Statistical analysis was performed using Microsoft Excel 2002. Single factor
Analysis of Variance (ANOVA) was chosen to examine the effect of weathering stage,
sex, age, and bone type on the amplification results. ANOVA is used when the
independent variables (in this case, age, sex, weathering condition, and bone type) have
discrete categories and the dependent variable (amplification success) is continuous or
binary (yes/no). This method of analysis involves comparing the difference among the
means of two or more sample sets, while considering the variance around each mean.
The analysis tests the hypothesis that the means of the groups are equal. For the analysis,
the age of the individual was categorized into 5 groups: infant (less than 1 year old), child
(1-13 years old), adolescent (14-17 years old), young adult (18-24 years old), adult (25-

49 years old), and older adult (50+). Results were considered significant at p<.05.

23

Results
Sampling

Samples were chosen based on the number and bone type available for a
particular burial number. As a result, there was a bias in sampling towards individuals
with more than one bone available for analysis. Those individuals with at least three
bone types, preferably femur, rib, and pelvis, were selected first, followed by those with
only two of these bones. One individual (203) had three bone types available (fibula,
cranium, and rib), two of which did not correspond directly to the set categories. The
fibula was grouped into the femur category for statistical analysis due to its similar
structure as a compact bone. The cranium remained in a category of its own. Five
individuals from skeletal weathering stage five (30, 114, 349, 402, 686) were analyzed
despite there being only one bone type for each individual. These samples were included

to bring the sample size at that stage to a number closer to that of the other stages.

Observations during Bone Preparation

In general all of the bones were brown in color on the outer surface. Sections of
bone originating from femurs appeared white beyond the cortical layer. In contrast, all
pieces originating from pelves and ribs were brown throughout. During the cutting
process, sections of femur came away in solid pieces, with the outer layer flaking off at
times. The porous nature of the pelvic bone led to the production of several smaller
pieces separating from the main sample when contact was made with the Dremel tool.
For the most part, sections cut from ribs separated in one piece from the source sample,

with the outermost layer flaking off at times.

24

During the washing procedure, small amounts of dirt were removed in the first
wash step, with all of the rinse steps with water remaining clear. A difference in the
amount of dirt removed was noticed only for the rib and pelvis samples from individual
704. With these samples, the wash solution turned opaquely brown and did not
completely clear until the second or third water rinse. This individual had been rated as a
stage five with the skeletal weathering stage criteria. Using the individual bone
weathering stage criteria and rating each bone separately, the rib was a stage three, and
the pelvis was a stage one. Despite the noted difference in the amount of dirt removed
from these two samples, amplification of DNA was not affected, as both samples
produced an amplicon.

During the milling process, the femurs were the hardest bone type to crush. In
most cases, the sections were broken down into smaller pieces ranging from 1 mm to
powder. Both ribs and pelves were more friable than the femurs and thus ground more
easily. The size of rib and pelvis pieces after grinding ranged from powdered to 1—2 mm
slivers of bone. The cortical layer of rib and pelvis sections tended to break off from the
main part of the bone in sheets. If this outer layer was present at the grinding stage, it
was not broken down into powder, but retained its fibrous structure and was broken into
slivers of bone approximately 0.5 mm in diameter and 2 mm in length. The mass of the
sample before and after grinding was recorded. On average there was a 25% loss of bone
during the grinding procedure and transfer to tubes for extraction.

DNA Extraction
Many samples needed to be extracted two or three times with phenol before the

phenol layer no longer appeared dark brown. In almost all instances, the samples which

25

required multiple phenol extractions came from either rib or pelvis. The color of the
DNA pellet after extraction was noted for each sample, which varied in translucency and
color; out of solution DNA should be colorless if it is clean. Translucent samples ranged
from colorless to light yellow or light brown, while opaque samples ranged in color from
white to light yellow or light brown. On occasion some samples would have a darker
brown spot in an area of the pellet, presumably residual dirt. In general, DNAs from
femurs were lighter in color than those from pelves or ribs, with only one femur sample
having a dark brown spot in an otherwise clear pellet. The darker, brown pellets resulted
exclusively from extractions of pelves and ribs. Pellet color did not seem to affect
amplification in a predictable manner though, as rib extracts amplified significantly more

often than pelvis extracts (details presented below).

DNA Quantification Using QuantiBlot

DNA yields from seven samples were analyzed using QuantiBlot. Optimization
of the system with the mtDNA probes designed for this experiment was completed using
1 pl and 10 pl of human cell line DNA at a concentration of 10 ng/ pl. Testing probe
concentration, it was found that faint detection of the DNA was possible with the mtDNA
probes DLoopl 1F, DLoop31R, and C01 6281F, beginning at a 10 pmol/pl concentration
and increasing with the 100 pmol/pl concentration using the manufacturer’s suggested
hybridization and first wash step temperature of 50°C. When the hybridization
temperature was dropped to 45°C and the first wash step was dropped to 24°C, detection
of DNA improved greatly, hence even lower hybridization temperatures were tested. The

SDS in the hybridization solution began to precipitate at a 30°C hybridization

26

temperature, although hybridization did occur. Increasing the hybridization temperature
to 40°C, prevented the SDS from precipitating and hybridization was improved compared
to the 45°C experiment. The human positive control DNA hybridized to all three probes
at all hybridization temperatures.

DNAs from raccoon, red fox, badger, and salmon were used to test the specificity
of each of the mtDNA probes; DLoopl 1F and DLoop3 1R are designed to be human-
specific, while COl 6281F is designed to be mammal-specific. Preliminary results
indicated that at 40°C, raccoon, red fox, and salmon DNA hybridized to the DLoopl 1F
and DLoop31R probes, while all these species plus badger hybridized to the C01 6281F
probe. These DNAs were at a much higher concentration than the human control DNA
however, and diluting them to 10 ng/ pl decreased the amount of hybridization, while
raising this temperature to 45°C prevented hybridization completely when DLoopl 1F and
CO 6281F were used, but not DLoop31R. Therefore, DLoopl 1F and C01 6281F were
used in the quantification procedures for preliminary testing on seven bone samples;
DLoop31R was not used due to its non-specific binding. In addition, the 40°C
hybridization temperature was used to improve the sensitivity of the probes.

Six of the samples tested did not generate an interpretable signal under these
conditions, as the strength of the signal was not above the lowest standard, and thus was
considered to be below the lower limit for the sensitivity of the experiment (Figure 3).
Only DNA extracted from the femur of burial number 47 produced a positive result: a
dark band, stronger in intensity than the most concentrated standard. Given this rather
surprising result, it was felt that this result may have been artifactual (see Discussion),

and in combination with the negative results obtained for the 6 other samples, as well as

27

the large amount of each sample required for each analysis, quantification via QuantiBlot
was not continued.

Figure 3. Quantification of mtDNA Using QuantiBlot

DLOOpllF DLoopllF/C016281F cor 6281F
A CS 2 . CS 3

fiFflUOUfl>

 

MtDNA quantification using the DLoopl 1F and C01 6281F probes separately and
together (DLoopl lF/COl 6281F). CS=Concentration Standard, A—G represents serial
1:2 dilutions of the human cell line DNA beginning with 10 ng. Lane A in columns 1—3
represents the positive control human cell line DNA (1 pl loaded). Lanes D—F in lanes
1-3 contain 30 p1 of DNA from bone samples 47 femur, 345 femur, and 489 rib. Results
from 47 femur represent usable data with a band greater in intensity than the most
concentrated standard. Results from 345 femur and 489 rib were not above the lowest
standard, and thus were considered to be below the lower limit for the sensitivity of the
experiment.
DNA Amplification

MtDNA fragments were successfully amplified from 55 of the 89 samples.
Typical PCR results can be seen in Figure 4. A result was positive when a band for the
targeted amplicon was observed (positive band, PB) and negative when no band was
observed, but primer activity (PA) was present. Primer activity (including a process

known as primer-dimer formation) stems from the two PCR primers interacting to form a

small PCR product. The product should be seen even when when no bone DNA is added.

28

If primer activity does not occur and only the primers themselves are seen (P), something
is inhibiting the PCR reaction.

On occasion the reagent blanks from the bone extraction procedures produced
positive results after PCR. Each time this occurred, the mtDNA from the reagent blank
was sequenced along with the bone samples that were processed following the last
negative reagent blank. In no instances did mtDNA from a reagent blank match any of
the sequences of the bone samples or laboratory personnel involved in sample processing.

Figure 4. Typical Amplification Results: Positive, Negative, and Inhibited

 

Lanes 1—8 represent amplification results from bone samples using the 220 bp primer set.
Lanes 1-4 and 7 are positive (target band (PB) present). Lanes 5, 6, and 8 represent
inhibited results (no primer activity (PA) observed). Lane 9 is the positive control. Lane
10 is the negative control.

Originally, 31 bones were amplified using standard Taq polymerase, of which
only one produced a band using the 220 bp primer set. In an effort to improve
amplification success, standard T aq was replaced with HotMaster Taq. An additional 6
samples of this original group of 31 produced a 220 bp amplicon. During these
preliminary experiments it was noted that a 1 :20 dilution of the extracted DNA improved
PCR results; in some cases amplification occurred at a 1:20 dilution, but not with
undiluted DNA. Further, if amplification did occur with the undiluted DNA,

amplification also occurred at the l :20 dilution, but with increased band intensity.

Therefore only the 1220 DNA dilution was used for the remaining experiments.

29

In addition to the change in T aq, a PCR enhancer was added that might help
reduce PCR inhibition. As a result, six samples lost PCR inhibition, three of which
successfully amplified. Samples that continued to show PCR inhibition were cleaned via
a Microcon-100 spin column. Eight of the twelve remaining inhibited samples
successfully amplified following this procedure. Results are shown in Table 4. In the
end, four samples out of the total 89 remained inhibited after all steps to remove PCR
inhibition were performed, three were pelves (328, 345, and 402) and one was a rib
(539). Of interest, only one femur (409B) inhibited PCR before any improvement steps
were taken, which was overcome with the addition of the enhancer. One sample, 489 rib,
inhibited PCR after the addition of the enhancer, but the sample was lost in the Microcon
step, therefore the final result for this sample is not known; however the available

information for this sample was included in the overall PCR results.

30

Table 4. Improvement in Inhibition

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Burial Bone Type Inhibited Amplification Skeletal Individual
Number Stage Bone Stage
047 Rib No Yes“ 4 4
124 Rib No Yes* 3 4
126 Pelvis No No 3 3
126 Rib No No 3 3
132 Pelvis No Yes* 2 2
192 Pelvis No Yes 2 2
322 Pelvis No Yes 2 2
328 Pelvis Yes No* 2 2
328 Rib No Yes* 2 2
345 Pelvis Yes No* 3 3
402 Pelvis Yes No* 5 3
409B Rib No Yes* 3 3
409B Femur No Yes 3 2
489 Rib Yes No" 2 l
539 Rib Yes No* 4 4
545 Pelvis No No l 3
545 Rib No Yes* 1 4
704 Rib No Yes" 5 3
704 Pelvis No Yes* 5 l

 

 

 

 

 

 

 

List of the 18 samples which initially showed PCR inhibition arranged by burial number
and bone type. Inhibited column indicates whether PCR reactions with these samples
remained inhibited following dilution of the DNA extract. * indicates samples that were
processed with a Microcon-100 spin colmnn. ** indicates the sample for which final
results are not known as the sample was lost during the Microcon step.

The final PCR results, arranged by burial number and bone type, are shown in
Table 5. If amplification was successful, it was noted by the maximum fragment length
obtainable. Samples that were subject to Microcon-100 processing are denoted by a star.
Thirty-four samples produced the 220 bp amplicon, all of which were additionally
analyzed using the 329 bp set of primers. Five samples (203 cranium, 164 femur and rib,
and 345 rib l and 2) produced the 329 bp amplicon. PCR of those five samples with the
402 bp primers was carried out, but negative results were obtained from all samples. Of

the 50 samples that did not produce a 220 bp band, but did not inhibit PCR, 21 produced

a 107 bp amplicon.

31

Table 5. Overall PCR Results

Headings denote burial number, bone type, and the maximum obtainable amplicon size.
Those samples in which an amplicon was not produced and there was no primer activity
are labeled as “inhibited”. Samples which did not produce an amplicon but showed
primer activity were considered negative results and all boxes were left blank.

* indicates samples that were processed with a Microcon-100 spin column.

32

Table 5.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Burial General

Number bone 107 bp 220 bp 329 bp Inhibited
027 Femur X

027 Rib X

030 Femur X

034 Femur X’

034 Pelvis

034 Rib X

047 Femur

047 Pelvis X

047 Rib X‘

047 Rib

11 1 Femur X

11 1 Rib X

1 1 1 Pelvis X

1 14 Femur X

124 Femur X

124 Pelvis X

124 Rib X“

126 Femur

126 Pelvis

126 Rib

132 Femur X

132 Pelvis X‘

1 32 Rib X

164 Femur X
164 Pelvis

164 Rib X
167 Femur X

167 Pelvis

167 Rib

192 Femur X

192 Pelvis X

192 Rib

203 Cranium X
203 Fibula X

203 Rib

256 Femur

256 Pelvis

256 Rib

260 Femur X

260 Rib X

322 Femur X

322 Pelvis X

322 Rib X

328 Femur X

328 Pelvis X'
328 Rib X*

 

33

 

Table 5. Continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

331 Femur X

331 Rib X

345 Femur

345 Pelvis X"
345 Rib

345 Rib

348 Femur

348 Pelvis

348 Rib X

349 Femur X

355 Femur

355 Rib

381 Femur X

381 Pelvis

381 Rib X

389 Pelvis

389 Rib

402 Pelvis X*
4098 Femur X

4098 Pelvis

409B Rib X’

447 Femur X

447 Rib X

448 Femur X

448 Rib X

489 Pelvis

489 Rib X
529A Femur X

529A Pelvis

529A Rib X

539 Pelvis

539 Rib X“
540 Pelvis X

540 Rib

545 Femur X

545 Pelvis

545 Rib X“

546 Pelvis X

546 Rib

686 Cranium X

704 Femur X

704 Pelvis X“

704 Rib X*

Total 21 29 5

 

34

 

Skeletal Weathering Stage and Amplification Success

DNA amplification of a sample was considered positive regardless of the
amplicon size produced. Amplification results broken down by skeletal weathering
stage, which was based on the condition of the whole skeleton, are displayed in Figure 5
and show that there is no obvious relationship between weathering stage and
amplification success. Bones from stage five, the most weathered, amplified the most
often at 70% (14 out of 20), followed by stage two at 69.57% (16 out of 23), stage one at
66.67% (4 out of 6), and stage four at 61.11% (11 out of 18). Samples from stage three
amplified the least often, at 45.45% (10 out of 22). The difference in amplification
among all weathering stage categories is not significant (p=0.460).

Figure 5. Amplification Results in Comparison to Skeletal Weathering Stage

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8

c 30

0

g 20

Pg 10 IND

g 0 rIYes

Z 1 3 4 5
Weathering Stage

Graph depicting amplification success (yes) and failure (no) arranged by the five skeletal
weathering stages.

Figure 6 shows the same data divided into bone type. Amplification success is
similar across the five weathering stages, with and pelvis and rib samples amplifying the
least often at stage one (0 out of l and 1 out of 2, respectively), while femur samples
amplifying the least often at stage three (42.86%, 3 out of 7). There is no significant
difference in the pelvis, rib, or femur data among the five stages (p =0.604, p=0.956, and

p=0.076, respectively).

35

Figure 6. Amplification Results in Comparison to Skeletal Weathering Stage and
Bone Type

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 10
8 3 . _
9.; 5 I E i? .
a , ls.- Iglgl
g. , _ I I I II I I:e::I - u
2,. a“ II II II II we
552.295.295.2-252es‘529.yes
U 0
1 2 3 4 5
Bone Type by Stage

Graph depicting amplification success (yes) and failure (no) arranged by bone type within
each of the five skeletal weathering stages.
Individual Bone Weathering Stage and Amplification Success

Figure 7 shows amplification results based on the individual bone weathering
criteria in which each bone was staged individually. Amplification success across the
stages (without considering bone type) was similar, with the highest at stage two
(71.43%, 25 out of 35) followed by stage four (63.63%, 7 out of 11), stage one (63.16%,
12 out of 19), and stage three (45.83%, 11 out of 24). Statistical analysis comparing all
four weathering stages shows the variance in amplification success across the stages is

not significantly different (p=0.269).

36

Figure 7. Amplification Results in Comparison to Individual Bone Weathering
Stage

 

 

 

 

 

 

 

 

 

 

33

8 40

[n 30

8 2D

8 10 I N0
E I Yes
:3 0 -

Z

 

Weathering Stage

Graph depicting amplification success (yes) and failure (no) in relation to the four
individual bone weathering stages.

Amplification success of the different bone types varies across the four
weathering stages (Figure 8), however the differences are not significant. For example,
stage two femurs have the highest amplification success at 90.91% (10 out of 11),
followed by those at stage one at 76.92% (10 out of 13), and stage three at 50% (2 out of
4) (p=0.213). The pelvis samples show a decrease in amplification success from stage
one to stage three (as with femurs, none of the pelvis pieces analyzed in this study were
categorized as stage four) (66.67%, 2 out of 3; 46.15%, 6 out of 13; and 11.11%, 1 out of
9 for stages one, two, and three, respectively) (p=0.130). Rib samples have the highest
amplification success at stage two (75%, 6 out of 8), followed by stage three (72.73%, 8

out of 11), stage four (63.64%, 7 out of 11), and stage one (0%, 0 out of 3) (p=0.111).

37

Figure 8. Amplification Results in Comparison to Individual Bone Weathering
Stage and Bone Type

 

 

 

 

 

  

 

 

 

    

 

 

 

 

 

8 15
C.
O
E 10.
O
a 5.
g B No
Z 0‘

g I Yes

'E

a

U

1 2 3 4
Bone Type by Stage

Graph depicting amplification success (yes) and failure (no) arranged by bone type within
each of the four individual bone weathering stages.
Bone Type and Amplification Success
The strongest influence on amplification success appears to originate from bone

type. As shown in Figure 9, amplification occurred in the 2 crania samples (100%),
79.31% (23 out of 29) of the femur/fibula samples, 36% (9 out of 25) of the pelvis
samples, and 63.64% (21 out of 33) of the rib samples. Across bone types there is strong
statistical evidence (p=0.006) suggesting a difference in amplification success, with the
greatest difference between femurs and pelves (p=0.0009), followed by ribs and pelves

=0.037). There is no significant difference between ribs and femurs (p=0.181). Crania

were not included in the comparisons due to the low sample number.

38

Figure 9. Amplification Results in Comparison to Bone Type

 

 

 

 

 

 

 

 

 

 

 

 

 

     

h”; 40
8
CD I!)
”5 20
2 .NO
E 10 V IYes
3 0_ , ,
Z cranium femur fibula pelvis rib
Bone

Graph depicting amplification success (yes) and failure (no) arranged by bone type.

Sex Estimates and Amplification Success

Amplification results for each individual are shown in Figure 10 by the four sex
categories (female, male, “possible male”, and undetermined). Amplification occurred in
57.14% (4 out of 7) of the undetermined individuals, 63.64% (7 out of 11) of the females,
and 94.12% (16 out of 17) of the males. The one individual classified as “possibly male”
(M?) also amplified. There is no significant difference in amplification success among
the four categories (p=0. 120). However, removing the undetermined individuals fiom the
analysis reveals a significant difference in amplification success between males and
females (p=0.041), with male samples amplifying more often. Adding the “possible

male” to the male category drops the p-value to 0.034.

39

Figure 10. Amplification Results in Comparison to Sex of the Individual

 

 

 

 

 

 

 

 

 

 

 

 

(I)
E 20
:1
E
.Z
'0
E
“E I No
.8 I Yes
E
2".)
Z undetermined F M M?
Sex

Graph depicting amplification success (yes) and failure (no) in relation to the
anthropologically estimated sex of the individual. Four categories were used to describe
the individuals sampled in this study: undetermined (sex of the individual could not be
determined), female (F), male (M), and possible male (M7).

Amplification success for each bone was compared with respect to sex and
skeletal weathering stage. Six bones were classified as stage one, three originating from
undetermined individuals and three from females, with each category having an
amplification success of 66.67% (2 out of 3); no male skeletons were classified as stage
one, eliminating a comparison between sexes. At stage two, 25% (1 out of 4) of bones
from the undetermined category amplified, 83.33% (15 out of 18) of bones fiom male
skeletons amplified, and 33.33% (1 out of 3) bones from the “possible male” amplified;
no female skeletons were designated as stage two, nullifying a statistical comparison
between sexes. Bones from male skeletons at stage three amplified 52.63% (10 out of
19) of the time, whereas bones from female skeletons at stage three amplified 0% (out of
3) (p=0.097). All the bones from undetermined individuals at stage four amplified,
53.85% (7 of the 13) of female bones amplified, and 66.67% 2 out of 3 male bones

amplified (p=0.710). Finally, at stage five, amplification occurred in 40% (2 out of 5) of

40

bones from undetermined individuals, 66.67% (4 out of 6) of female bones, and 100%
(out of 8) of male bones, providing no evidence for a significant difference in
amplification success between the two sexes (p=0.088). Thus, no significant difference
between the sexes existed at stage three and stage five, while comparisons could not be
made at stage one and two.

Similar comparisons were made for the individual bone weathering categories.
However, unlike the skeletal weathering categories, these included bones from male and
female individuals in all stages. Thirty-three percent (1 out of 3) of the female bones
from stage one amplified and 78.57% (11 out of 14, one of which was from the “possible
male”) of male bones amplified (p=0.134). Stage two bones from females amplified
72.72% of the time (8 out of 11) and from males 66.67% of the time (12 out of 18, two of
which were from the “possible male”) (p=0.743). Female bones at stage three amplified
16.67% of the time (1 out of 6), while male bones amplified 64.29% of the time (9 out of
14) (p=0.054). At the most weathered stage, 60% (3 out of 5) of the female and 80% (4
out of 5) of the male bones amplified (p=0.545). Thus, at all four stages of the bone
staging system, there is no significant difference in amplification success between male
and female bones.

Amplification results in comparison to bone type and sex of the individual were
also analyzed. In all, 7 femurs, 8 pelves, and 10 ribs were sampled from females; 18
femurs, 15 pelves, and 18 ribs were sampled from males. The male numbers include
three bones, one of each type, that were collected from the “possible male”. Neither of

the two crania nor the single fibula was given a sex classification. There is no significant

41

difference in amplification success between the sexes when comparing femurs, pelves, or

ribs (p=0.372, p=0.912, and p=0.065, respectively).

Age and Amplification Success

The effect of estimated age of the individual on amplification success is displayed
in Figure 11. There is no significant difference in the amplification success across the
five age categories (p=0.748). The amplification results of each age category are as
follows: 6 out of 8 (75%) children, 2 out of 2 (100%) adolescents, 5 out of 5 (100%)
young adults, 14 out of 19 (73.68%) adults, and 1 out of 1 (100%) older adult amplified.
The age of one individual (from burial number 389) was not estimated at the time of
recovery and therefore is not included in the statistical analysis.

Figure 11. Amplification Results in Comparison to Estimated Age

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0)
ES 20
3
'o
">' 15
‘6
E
I. 10
2 I No
B 5 I Yes
E
E 0_ ,V ‘ .

undetermined child adolescent young adult adult older adult

Age Category

Graph depicts amplification success (yes) and failure (no) of each bone sampled in
relation to the five age categories. The age of one individual was not estimated.
Amplification success in relation to age and bone type was also analyzed.
Successful amplification from femurs occurred 66.67% (4 out of 6) of the time in
children, 100% in adolescents, young adults, and older adults (out of 2, 4, and 1 sample,

respectively), and 75% (12 out of 16) of the time in adults. With ribs, 33.33% (2 out 6)

42

of samples from children amplified, as did 50% (1 out of 2) from adolescents, 80% (4 out
of 5) from young adults, 77.78% (14 out of 18) from adults, and 0% (0 out of 1) from
older adults. Pelvis samples amplified 50% of the time from children (1 out of 2) and
young adults (2 out of 4), 0% (out of 1) from adolescents, 31.25% (5 out of 16) from
adults, and 100% (out of 1) from older adults. A significant difference is not observed
among the age categories regardless of bone type (p=0.674, 0.172, and 0.626 for femur,
rib, and pelvis, respectively).

The effect of age on amplification success shows no significant difference among
the age categories at any of the five stages of the skeletal weathering system (p=1, 0.057,
0.451, 0.948, and 0.624, respectively). Similarly, when the same analysis is performed
based on the bone weathering stages, no significant difference among age classes is found
at weathering stages two through four (p=0.150, 0.465, and 0.695, respectively); however
at weathering stage one there is a significant difference in amplification success among
the age categories (p=0.026). With this stage, bones from adolescent to adult amplified
the best (100%, 100%, and 70%, respectively), while bones fi'om the child and older adult

categories did not amplify at all.

DNA Sequencing

To confirm that PCR product generated in these experiments actually originated
from the bone being tested, positive samples were processed for sequencing. Sequences
were produced for any individual that produced PCR products fi'om more than one bone
so that sequences comparisons within an individual could be made. Sequencing was

performed on 31 PCR products, ten 107 bp amplicons and twenty-one 220 bp amplicons,

43

using the same primers as amplification. The 5’ end of both the forward and reverse
reactions produced poor quality data which, in some cases, meant that overlapping
sequences between the two reactions was not possible (Figure 12). A full sequence of the
targeted region was obtained from 15 (48.39%) samples, none of which originated from a
107 bp fragment.

Figure 12. Electropherogram Depicting Typical 5’ End Data

30 40 50
PAC CATAGGACT TCA CAN A G NC A A G T A

«gay.

Portion of electropherogram showing poor quality data at the 5’ end. Peaks correspond
to each of the four bases: adenine (A), guanine (G), thymine (T), and cytosine (C). N
represents a situation in which one of the four bases could not be called. Bases are
number by tens along the top. Analyzable data begin at base 50.

 

Forward and reverse sequences were aligned with the Anderson reference
sequence and compared to determine if the DNA extracted from the different bones
originated from the same source. Fifteen of the 21 (71.43%) 220 bp samples produced
analyzable sequence from both the forward and reverse primers. Four (19.05%) 220 bp
samples (27 femur, 192 pelvis, 448 femur, and 448 rib), produced analyzable sequence
from only one DNA strand. In each case, the readable strand originated from the forward
primer, with the exception of 448 rib. Two (9.52%) 220 bp samples (345 rib and 4098
rib) did not produce an analyzable sequence from either primer. Eight of the ten (80%)
107 bp samples did not produce analyzable sequences. Two (20%) 107 bp samples (111

femur and 260 femur) gave readable sequence with the forward primer only.

Comparable sequences, those in which there was at least a partial sequence from
more than one bone of an individual, were obtained from eight individuals (Table 6).
Sequences within an individual were consistent in all of these. Three individuals (27,
164, and 203) differed from the Anderson reference sequence by one base across the
analyzable section, one individual (447) differed by two bases, and four samples (192,
448, 529A, and 704) did not differ from the reference sequence. All sequence
polymorphisms were unique to the mtDNA database.

Table 6. Sequence Polymorphisms In Individuals Generating PCR products form
More Than One Bone

 

 

 

 

 

 

 

 

 

 

Burial Number Sequence Sequence Polymorphism
and Bone Type Interval
27 femurF 16121-16232
rib F/R 16272-16376 16340C
164 femur F/R 16173-16216
rib F/R 16247-16385 16267+T
192 femur F/R 16167-16390 None
pelvis F
203 fibula F/R 16167-16388 16270+T
cranium F/R
447 femur F/R 16169-16260
rib F/R 16299-16378 16248T, 16340C
448 femur F 16177-16382 None
rib R '
529A femur F/R 16169-16387 None
rib F/R
704 femur F/R 16186-16378 None
rib F/R

 

 

 

 

The sequence interval available for comparison between at least two bones of an
individual, with polymorphisms fiom the human mtDNA reference sequence noted.
F=forward, R=reverse sequence.

45

Discussion

The goal of this study was to determine if there is a relationship between the
outward appearance of skeletal material (considering bone type and level of weathering)
and the quality of the DNA within. Being able to predict the amount of obtainable DNA
and its level of degradation (i.e., an appropriate target sized amplicon) would assist the

forensic scientist in providing a more efficient and productive analysis of the material.

DNA Pellet Color

One step of DNA preparation after extraction is to bring the DNA out of solution
and centrifuge it down to a pellet. Out of solution, DNA is colorless or white if it is
clean. Discolored pellets may be dirty and inhibit PCR. The DNA pellets of the Voegtly
samples were sometimes discolored; however, there was no apparent trend between the
degree of discoloration and amplification success. Although DNA pellets from femur
extractions were often lighter and DNAs from femurs amplified more often, pellets from
pelvis and rib extracts showed equal amounts of brown discoloration, yet DNAs from ribs
amplified significantly more often than that from pelves. Therefore, pellet color does not

seem to be a good indicator of predicting amplification success.

DNA Quantification

In a typical forensic lab setting, nuclear DNA yield for a particular sample is
estimated before amplification so that extracts can be diluted for optimal results; the ABI
QuantiBlot system is a method commonly employed for the quantification of nuclear

DNA. However, no such method exists for mtDNA. Modifications were made to the

46

QuantiBlot procedure to allow for the quantification of mtDNA, including the design of
three mtDNA probes and reduction of the hybridization and wash temperatures.
Although the modified procedure did not prove useful for DNA quantification of the
Voegtly samples, the procedure was optimized and shown to specifically detect human
mtDNA using the DLoopl 1F probe and mammalian mtDNA with the CO] 6281 probe.
Oddly, the DLoop3lR probe did not bind specifically to human mtDNA, but also bound
to raccoon, salmon, red fox, and badger DNA, even at elevated hybridization
temperatures. This is surprising as DLoop31R is the exact complement of DLoopl 1F;
the reason why one would bind specifically but not the other is unknown. The single
interpretable signal from a bone extract may have been artifactual, as DNA from that
extract (47 femur) was not amplifiable. If the result was not artifactual, and was indeed
of human origin, then the DNA from that extract should amplify given the intensity of the
band observed (Figure 3). However, amplification was not successful, therefore the
source of this of the DNA could not be confirmed and could be the result of
contamination.

In addition, the 10 ng/ p1 concentration of the human cell line DNA given by the
manufacturer describes the total DNA concentration, both mitochondrial and nuclear.
Without knowing the exact mtDNA concentration of standards used, only comparative
results can be made (e. g. sample 1 is less concentrated than sample 2, but more than
sample 3). It is quite possible that the detection limits of this method of quantification
are too high for the amount of mtDNA contained in the Voegtly samples, indicating that a

more sensitive test would be required. A separate study is being conducted to determine

47

if an alternate method for quantification, real-time PCR, which amplifies the DNA and

thus has a lower detection limit, could prove useful for analyzing these samples.

DNA Amplification

Since quantification via QuantiBlot was unsuccessful, amplification procedures
were performed without knowing the DNA yield of the extracts, therefore, a more
empirical approach was needed to optimize the amount of DNA extract added to each
reaction. In early experiments, side-by—side PCR reactions were run, one using the DNA
extract undiluted and the other at a 1:20 dilution. During these experiments, it was
noticed that when amplification was successful with the undiluted DNA, amplification
also occurred with the 1:20 dilution, while the band intensity generally increased with the
diluted sample. On occasion PCR was successful only with the diluted DNA and not the
undiluted, and in some cases, an undiluted sample was inhibited, but inhibition was
overcome by dilution. Therefore, as experiments progressed, only a 1:20 DNA dilution
was used. The improvement in amplification at this concentration likely resulted from
the dilution of components within the DNA extract that either inhibited PCR or prevented
amplification of the target region. For example, with degraded samples, the many small
pieces of DNA could interfere with PCR because they contain the target sequence for the
primers and are preferentially amplified. Small pieces of DNA contained in the extract
may also act as random primers and result in non-specific amplification. In either case,
dilution of the extracts would lower the concentration of these components to a level at

which they do not interfere with the designed amplification.

48

In addition to the 1:20 DNA dilution, a PCR enhancer was incorporated into the
reaction in an effort to reduce PCR inhibition that remained after the dilution of the DNA
extract. The ingredients of the enhancer are proprietary, so one can only speculate on
what exactly is occurring with its addition to the reaction. One possible reason for
inhibition is the presence of compounds that bind to proteins. As a protein, T aq
polymerase would be a target for such inhibitors. Another protein, bovine serum albumin
(BSA) is often added to reduce PCR inhibition. BSA acts as a competitor for protein-
binding inhibitors; thus if enough BSA is added, T aq polymerase will be free to perform
its polymerase activity and PCR will no longer be inhibited. BSA and/or other similar
components may be present in the PCR enhancer. Although addition of the enhancer did
not improve all inhibited reactions, it improved some, and for that reason it was added as
a routine component to the PCR reaction.

The final addition made to the PCR reaction to optimize amplification results was
hot-start T aq polymerase. Standard T aq polymerase has an optimal activation
temperature of 72°C; however, some activity occurs at lower temperatures. Thus, if
primers, including the aforementioned small DNAs acting as primers, bind non-
specifically to the template DNA at lower temperatures, this early activity may lead to
spurious amplification of the template. These undesirable products have the potential to
monopolize the activity of the polymerase, leading to reduced amplification of the
targeted region. Initiating PCR at a higher temperature prevents such non-specific primer
binding and amplification. The hot-start T aq polymerase used in this study is coupled
with a temperature-dependent inhibitor which is released at 94°C, making the polymerase

inactive at temperatures below 60°C, thus reducing non-specific amplification due to low

49

temperature activity. When standard T aq polymerase was used in initial experiments,
only 1 of 31 samples produced a positive PCR product. Using hot-start T aq polymerase
improved the amplification results of those 31 samples six fold. Thus, the use of hot-start
T aq, together with a 1:20 DNA dilution, and a PCR enhancer greatly improved
amplification success.

The use of the Microcon-100 spin column was incorporated as part of DNA
preparation when a sample inhibited the PCR reaction. The pore size of the filters within
the Microcon-100 columns is small enough that double-stranded DNA molecules 125 bp
(300 bp single-stranded) and larger are retained while smaller molecules flow through
and are removed. During this process, smaller pieces of DNA, which, as mentioned
above, may compete with the targeted DNA for primer binding or act as random primers
during PCR, are removed. In addition, other, smaller water soluble inhibitors that are not
removed during the organic preparation would be eliminated. Use of the spin column
removed the inhibitory factors in 2/3 of the remaining inhibited reactions (Table 4). With
this success rate, Microcon-100 spin columns do provide an easy and quick way of

removing PCR inhibition, and thus should be used as a routine step when a sample

inhibits PCR.

DNA Quality

The quality of mtDNA in each sample was assessed by determining the largest
amplifiable fragment of DNA within the sample. Twenty-nine of the 89 samples
produced a 220 bp fragment, 21 produced a 107 bp amplicon, 5 samples produced the

329 bp fi'agment, while no samples amplified at the 402 bp size. The 29 samples which

50

did not produce a 220 bp amplicon, but were not inhibited, were not tested for
amplification of fragments smaller than 107 bp. The degraded state of the DNA likely
reflects the age and condition of the samples and is in line with the typical fragment
length isolated from ancient DNA (<300—500 bp) (O’Rourke etal., 2000). It is
interesting to note that three of the five bones that produced a 329 bp fragment were ribs,
while general amplification (regardless of size) was most successful with femurs
(discussed below). This could indicate that the DNA within ribs and femurs degrades on
different timelines due to the difference in bone structure. Flat bones such as ribs are
composed largely of spongy bone, with more accessible DNA. Also, there may be more
cells by weight in the spongy portion of a flat bone than in the compact portion of a long
bone. Both of these characteristics would lead to the extraction of higher quality DNA
from flat bones, although this relationship might only exist while the bones are well-
preserved. The larger distribution of spongy bone within flat bones also makes them
more susceptible to degradation and as time passes and weathering occurs, the structures
that protect the DNA will be destroyed. Bones that are composed largely of compact
bone, like femurs, degrade slower, and while there may be fewer cells within a section, or
DNA is less accessible, the microstructure stays intact longer, allowing for amplification
of DNA from long bones after it is no longer successful from flat bones. Similar results
have been found regarding quantity of DNA. Parsons and Weedn (1997) found that
fresh, spongy bones, such as ribs, have the potential to yield 10- to 20-fold more DNA

than compact bone; however such yields are less reliable with increased time since death.

51

DNA Sequencing

When amplification product was available for at least two bones from any one
individual, amplicons were sequenced to ensure that the DNA extracted from the
different bones originated from the same source. The limited amplicon size obtained
from these aged samples prevented the successful analysis of all 31 samples that were
processed for sequencing. Smaller amplicons proved to be the most problematic. This
was due to the poor quality of data at the 5’ end of every reaction, which can be attributed
to the large amount of unincorporated ddNTPs that pass through the capillary at the
beginning of the sequencing procedure. Deletion of this poor quality data before
sequence analysis rarely left enough quality data to be successfully analyzed from the
107 bp reactions; only two samples produced readable sequence, and only in the forward
direction. Therefore, sequencing of small amplicons using the cleanup methods
recommended by Beckman-Coulter is not advised. Better methods are needed for the
cleanup step to remove unincorporated ddNTPs. Sequence analysis of the 220 bp
fragments did provide sufficient data to show that the PCR results were from the bones
and not contamination (Table 6). In the eight individuals analyzed, the sequences being
compared were consistent within the individual, indicating the DNA extracted from
different bones originated from the same skeleton. Four individuals showed these
consistencies with no nucleotide differences (polymorphisms) from the Anderson
reference sequence. The remaining four had at least one polymorphism, (one individual
had two) within the sequenced interval. In no instance did the DNA isolated from bone
samples match the lab personnel involved in the extraction and analysis procedures and

all sequence polymorphisms were unique to the FBI mtDNA database.

52

Weathering Condition and Amplifiable DNA

One of the most important goals of this study was to determine if the visual
appearance of a skeleton or bone could be used to estimate the likelihood of obtaining a
PCR product. Originally, the anthropologists at The Smithsonian Institution
systematically graded the condition of each skeleton based on the guidelines of
Behrensmeyer (1978). To address the question of whether this overall classification was
predictive of DNA quality, bones were chosen with the intent of equally sampling five of
the six weathering stages (small sample size prevented the analysis of stage 0 skeletons).
It is interesting that as bones were being processed for DNA analysis, there did not seem
to be a correlation between how a skeleton was graded and the condition of a specific
bone sample. On the other hand, it was noticeable that the most distinguishing
differences occurred among bone types. Femurs tended to be in better condition than ribs
and pelves; they were rarely discolored beyond the outermost layer, and in no instance
was there a femur in several small pieces. In contrast, ribs and pelves were brown
throughout. In addition, the outermost layer of most pelves was missing in portions of
the samples or not present at all, leaving the spongy interior exposed and susceptible to
breakage. Similar flaking occurred in ribs, but to a lesser degree; however, ribs displayed
the greatest amount of breakage and in some cases, only tiny pieces of bone were present.
Some of this breakage may have resulted from the fragile nature of ribs, which would
affect how they break down under natural conditions or how well they stand up to
handling and transport. Comparing individual bone types among skeletal weathering

stages, it was not unusual to have a femur that came fi'om a more degraded skeleton (e. g.

53

stage four) that appeared to be in the same or better condition as one which came from a
less degraded skeleton (e. g. stage two). In comparison, the state of ribs and pelves
seemed to coincide more often with the skeletal ratings, although on occasion a rib or
pelvis was found that was in better or worse condition than the others of the same stage.
The difference between femurs and ribs or pelves in discoloration and condition could be
attributed to the structure of each bone type. Femurs and other long bones are composed
of compact bone from the cortex to the medullary cavity along the entire diaphysis (shaft)
of the bone and spongy bone beneath the cortex at the articular ends. In contrast, ribs and
pelves have spongy bone beneath the cortex along all portions of the bone, with compact
bone present only in the cortex. The compact nature of the femur diaphysis prevents
discoloring compounds from the soil to penetrate much beyond the cortical layer, while
the spongy nature of the interior of ribs and pelves would not prevent such action once
the cortical layer is penetrated or removed. In addition, denser, compact areas of bones
are more resistant to degradative processes than less dense, spongy areas, leaving femurs
in better condition than the rest of the skeleton.

ANOVA was used to examine the difference in amplification success among the
different categories of skeletal weathering, individual bone weathering, bone type, sex,
and age. No difference in amplification success was detected among the five skeletal
weathering stages. Interestingly, although stages one, two, four, and five were similar in
success (approximately 70%), bones from stage five skeletons amplified the most often,
while bones fiom stage three skeletons amplified the least often (45.45%) (Figure 5).
One way to account for this is chance; if no relationship exists between bone appearance

and DNA degradation, a random distribution of the data would allow for one category to

54

be the highest and another the lowest. A second possibility is that the DNA from more
degraded skeletons, such as those at stage five, is more accessible for extraction owing to
the fact that these bones are already broken down. Alternatively, all bones within a
skeleton do not degrade in the same manner with better preserved bone existing in an
otherwise highly degraded skeleton. The better preserved bones may contain more
amplifiable DNA, and the presence of these bones in the stage five skeletons increases
the amplification success of this category above that which would be found if only poor
bones were selected. If stage five skeletons provide the best source of DNA due to the
structure of a specific bone type rather that the condition of the skeleton as a whole, this
should be considered by the DNA analyst. Whatever the reason for the distribution of
amplification success across the stages of skeletal weathering, it is clear that the condition
of a skeleton is not indicative of the quality of DNA within the bones.

To incorporate the potential influence of specific bone types on amplification
success, each skeletal weathering stage was further divided by bone type. No significant
difference was detected for femurs, ribs, or pelves (Figure 6), further indicating that
staging based on skeletal weathering does not aid in the prediction of the quality of DNA
that can be extracted from the bone. Though skeletal rating is informative to the
anthropologist, it may not be a useful classification system for the DNA analyst.

A new four stage weathering classification system was created during the course
of this study and applied to each bone. Reclassification of the individual bones
confirmed that the degradation of a skeleton was not consistent throughout the skeleton
(Table 4). For example, in burial number 34, the skeleton was determined to be at stage

four under the original weathering system; however, when each bone was classified on its

55

own, the pelvis was classified as stage two, the rib as stage four, and the femur as stage
one. In general, all three bones within a skeleton were reclassified into different stages.
On occasion, a pelvis and rib from a single skeleton were reclassified as the same stage;
only rarely were all three bones reclassified as the same stage. If the level of degradation
is predictive of the quality of DNA within the bone, a bone rating system would reveal a
correlation that was masked when the whole skeleton rating system was used. However,
when the amplification success data were rearranged according to the bone staging
system, a significant difference among the stages was still not detectable (Figure 7). It
should be noted that as a consequence of reclassification, the sample sizes across the four
weathering stages was not equal. While there were close to 20 bones available for
analysis from each of the skeletal rating stages and all three bone types were represented
at each of the stages, stage one and stage three of the bone rating system had sample sizes
close to 20, while stage two had over 30 and stage four had only 10. The decrease in
sample size at stage four is due to the fact that no femurs or pelves were placed into this
category. Because of this unequal sampling, differences among the stages may be hard to
uncover, particularly among the most weathered bones. First, with a smaller sample size,
the bones at stage four may not be representative of all stage four bones, and indeed,
pelves and femurs are not represented. In addition, comparing the data fiom stage four to
those fiom stages with two or three times the number of samples would affect the
significance of any differences among the stages; the more equal the sample sizes are
among stages, the more significant a difference will be if it exists. Regardless, the data
available from this study do not appear to follow a trend and indicate that individual bone

weathering condition is not predictive of DNA quality.

56

The issue of unequal sample sizes among the stages of the bone staging system
continued when samples were further categorized by bone type (Figure 8); supporting the
idea that skeletal staging does not match bone staging. The difference in femur
amplification success among the three stages was not significant, although the small
sample size of stage three femurs, compared to the other two stages, may influence this
finding. Data from stage one and stage two femurs, which have larger sample sizes,
appeared to be very close in amplification success, suggesting that weathering condition
of femurs is not indicative of DNA quality. A larger sample size of stage three femurs in
addition to a representative sample of stage four femurs for comparison would solidify
this result.

Pelves were the only bone type for which amplification success appeared to
follow a trend across the bone weathering stages with a decrease from stage one to stage
three (no pelves being classified as stage four); however the difference among the three
stages was not significant. Again, these data may be skewed by the uneven distribution
of the pelves across the weathering stages. In this instance, stage one had the fewest
number of pelves. A larger sample size from stage one may confirm that there is no
difference in DNA amplification success among the different stages of pelves or may
reveal that stage one pelves contain the best quality DNA. In addition, having a set of
stage four femurs would help strengthen either conclusion, depending on the results.

Data from ribs did not appear to follow a downward trend; the lowest
amplification success occurred at stage one and the difference among the four stages was
not significant. This suggests that the degree of weathering of ribs also does not

influence the quality of the DNA within. Interestingly, considering only stages two

57

through four, amplification success decreased with increasing degradation; stage two and
stage three had nearly the same success while success at stage four dropped by 10%. The
sample size of stage one ribs was small and could influence the results found at this stage.
Alternatively, the small sample size of ribs at stage one may not have affected the results,
in which case, amplification success truly was worse with the most well-preserved bones,
perhaps due to the DNA being trapped in the better preserved material. DNA may be
released more easily from the more degraded ribs, with their more fragile microstructure.
A larger sample size of stage one ribs would confirm such a result or solidify that
amplification success was the same among ribs of different stages.

A commonality among femurs, pelves, and ribs when categorized by bone staging
was unequal sample sizes among the stages and small sample size at one or more stages,
as opposed to the equal sampling using skeletal staging. However, the reason for the
difference between the two staging methods varied depending on bone type. For
example, most femurs staged by the bone rating system were categorized as either stage
one or two, with only four rated as stage three and none as stage four. This indicates that
poorly preserved skeletons may still have well preserved long bones. In contrast, more
pelves were categorized as stage two or three by bone staging, with very few were
categorized as stage one. Thus, it appears that pelves of moderate degradation were part
of better preserved skeletons based on the skeletal rating system. The one problem with
this view is that in no instance was apelvis placed in stage four of the new system, while
pelves were found in association with stage five skeletons. This may be due in part to the
selection process of skeletal material; certainly only whole pieces of pelvis would be

collected from the burials, and skeletons without this desired bone type were less likely to

58

be sampled. Finally, only three ribs were rated as stage one of the bone rating system.
There are at least three possible explanations for this; the first is that ribs were generally
in a more degraded state compared to the rest of the skeleton due to their thinness and
overall fragile nature, and reclassifying them using the bone staging system revealed that.
Second, it is possible that ribs, due to their small size and spongy microstructure beyond
the cortex, were broken during transportation; thus ribs that appeared to be in better
condition at the excavation site or at The Smithsonian arrived at MSU in worse condition.
If this is true, then the number of stage four ribs would have increased as a result of
initially well-preserved ribs breaking during transportation. The microstructure of these
initially well-preserved ribs would still be intact and could increase the amplification
success of stage four ribs. The third possibility, as mentioned above, is that the small
sample size of ribs at stage one did not affect the results and amplification success was
the lowest with the most well-preserved bones due to the microstructure of the bone. It
would be interesting to base a study of DNA quality solely on the appearance of each
individual bone, equaling sample size among categories, to fully determine whether the
differences or lack thereof among bones of various weathering stages are real or not.
From the data generated here, it would be predicted that individual bone differences

would not influence DNA results.

Bone Type and Amplifiable DNA
The effect of bone type was analyzed on its own, showing a strong difference in
amplification success among pelves, femurs, and ribs. The strongest difference occurred

between femurs and pelves, followed by the difference between ribs and pelves, while

59

there was no statistical difference between ribs and femurs. This may be explained by the
structural differences between the different bone types. The structure of long bones gives
them less surface area per volume from which degradation can begin. In addition the
microstructure of compact bone may be conducive to preserving DNA more so than that
of spongy bone. With the osteocytes of compact bone embedded in spaces between
concentric layers of hydroxyapatite and collagen, the cells, and thus the DNA within,
may be well-protected. While the osteocytes of spongy bone are embedded in spaces
within the trabeculae, this level of protection is not further reinforced by multiple layers
of bone matrix, but instead is surrounded by open spaces; thus the less-protected
osteocytes of spongy bone may be more susceptible to degradation. With this in mind,
DNA from flat bones, such as ribs and pelves, will be better protected only in the cortex,
where the structure is compact. If this cortical layer is not present, the sample may not
provide useable DNA. In contrast, the compact microstructure extends beyond the
cortical layer in the diaphyses of long bones, such as femurs and fibulae, providing a
larger area of protected DNA. The difference in amplification success between ribs and
pelves may be attributed to the rib’s smaller, linear structure with a smaller surface area
to volume ratio. The pelvis is irregularly shaped with a large surface area, making the
cortical layer more susceptible to degradation. Gotherstrom et al. (2002) observed that
DNA preservation was connected to both hydroxyapatite and collagen preservation.
These two components were not studied at the microscopic level in this study; however,
at a macroscopic level, femurs were better preserved than either ribs or pelves. A

separate study has begun in Brazil to analyze the microscopic qualities of the Voegtly

samples.

60

Sex and Amplifiable DNA

The difference between males and females was significant, with samples from
males amplifying more often. There were insufficient data to compare male and female
samples among each of the stages of the skeletal staging system; many stages contained
no bones from one of the sexes. At each stage of the bone rating system, male bones
amplified more often than female bones; however, this difference was strongest at stage
three. The difference was not specific to any particular bone type as there was no
difference between the sexes when comparing bone type. Of the 36 individuals analyzed,
11 were female, 17 were male, 7 were not determined, and the “possible male” was also
sampled. Though the sampling between the sexes was fairly even, the traits available for
sexing the Voegtly samples were generally not ideal. The degraded state of the skeletons
lead to the estimation of sex from less sexually dimorphic qualities of the skeleton such
as the robustness of the skeleton, and not on the more reliable features of the pelvis or the
skull. On occasion, female sex was assigned if an earring was present within the burial,
and male sex was determined if remnants of a moustache were present. With this in
mind, it is possible that the skeletons were sexed incorrectly when judgment was based
solely on robustness and /or burial artifacts. Relying on robustness could be particularly
problematic for judging the sex of these individuals due to their ethnicity. The people
buried in the Voegtly Cemetery were known to be of Swiss-German heritage. In a
population that is more robust than average, there may not be enough sexual dimorphism
to rely on this trait for sex determination. In this case, more robust females would be
misclassified as male (as would more gracile males in a less robust population), changing

the distribution of amplification success. Again, the anthropological estimates of sex

61

must be confirmed to determine if these data are correct. A separate study has begun to
analyze nuclear DNA of the Voegtly samples to determine if the original sex
determinations were correct or not. Preliminary results indicate that on occasion, they
were not. If the anthropological estimates were generally correct, then male skeletons
amplified more often most likely because they were larger and more robust, and therefore

less susceptible to degradation.

Age and Amplifiable DNA

Age alone did not appear to affect amplification results in general. All aspects of
the data (weathering stages and bone type) in comparison to age were considered. Small
sample sizes may have been a factor in the analysis of these data however, as bones were
distributed across five age categories. Eight children (16 bones), 2 adolescents (5 bones),
5 young adults (13 bones), 19 adults (50 bones), and one older adult (3 bones) were
analyzed. Further dividing theses data based on weathering condition and bone type
generated categories containing only one or two samples, creating a problem of small
sample size. A significant difference was detected between amplification success and
age at stage one of the bone staging system, with bones fiom children and older adults
amplifying less often than those from adolescents, young adults, and adults. This may be
an artifact of the small sample size of the child and older adult categories (3 bones and 1
bone, respectively), or perhaps indicate a difference in the quality of preservation
between extreme and intermediate age categories. The latter idea supports Micozzi’s
(1991) description of a biphasic relationship between age and bone density, and thus

survivability. Beyond random chance, this may be explained by the structure of the

62

bones at those respective ages. For example, at very young ages, calcification has not
been completed and most of the skeleton is still constructed of cartilage. The structure of
the bones is more susceptible to degradation and may not provide a stable environment in
which DNA is protected. Such degradation would explain why of the over 300
individuals recorded in the Voegtly Church burial records of being less than one year old,
only 175 were recovered and of these, seven were available for DNA analysis. Similarly,
old-aged individuals were not well represented in this sample set largely due to the fact
that most individuals buried in the cemetery did not live long enough to be placed in the
older age category. The average life expectancy of individuals once adulthood was
reached was estimated to be in the early to mid-30$, and of the 230 individuals who lived
beyond the age of 20, only 80 individuals passed the age of 50 (Ubelaker etal., 2003).

At older ages the bones are well into the degenerative process, making them more prone
to degradation and possibly breakdown of DNA Interestingly, no significant difference
among the age categories was observed at any other weathering stage, and thus again the
biphasic finding for stage one bones could be artifactual. On the other hand, this biphasic
relationship might only be discemable in well-preserved bones, and not be detectable
once bones have progressed to a higher level of degradation, meaning that there is a
threshold of degradation at which there is a distinction among age categories. Finally,
dividing the samples among the various stages and ages results in very small sample sizes

for many of the categories, thus spurious results would not be unusual.

63

Conclusions

The results of this study indicate that data on age of an individual at death,
weathering condition of the whole skeleton, and weathering condition of an individual
bone are not useful for predicting the quality of DNA contained within the bones. Sex of
the individual may play a role in the preservation of DNA, with DNA from male
skeletons amplifying more often; however, the anthropological estimate of sex for the
skeletons used in this study would ideally be confirmed genetically before this conclusion
was reached. Bone type does appear to be useful in predicting the quality of DNA within
the bone. Reclassifying the bones by the condition of each led to small sample sizes in
several categories. Larger sample sizes based on individual bones would be useful to
determine if the difference in amplification success with increased bone degradation is
statistically significant; however, data from the stages with larger samples sizes, such as
stage one and two femurs and stage two and three pelves and ribs, suggest that bone
appearance is not indicative of DNA quality. Using bone type alone as a means of
predicting DNA quality, femurs were found to be the best bone from which to obtain
amplifiable DNA, and other long bones (fibulae, tibias, humeri, radii, and ulnas) may
show the similar results. Ribs were found to be the next best bone type, followed by
pelves. The difference in amplification success of both of femurs and ribs in comparison
to pelves was found to be strongly significant. This was true of bones of all age
categories and both sexes. Thus, with the knowledge of bone type, the forensic biologist
should be able to predict the quality of the DNA within the sample. This would allow the
scientist to choose the best available bone if more than one exists. In addition, guidelines

could be provided to the individuals collecting the bone samples, so those most likely to

64

contain usable DNA are not overlooked and are sent to the forensic biologist. All of

these factors have the potential to increase the efficiency of DNA analysis of bone.

65

References

Anderson, S., A.T. Bankier, B.G Barrell, M.H.L. deBruijn, A.R. Coulson, J. Drouin, I.C.
Eperon, D.P. Nierlich, B.A. Rose, F. Sanger, R.H. Schreier, A.J.H. Smith, R.
Staden, I.G. Young. 1981. Sequence and organization of the human mitochondrial
genome. Nature. 290:457-465.

Behrensmeyer, AK. 1978. Taphonomic and Ecologic Information from Bone
Weathering. Paleobiology. 4(2): 150-162.

Buikstra J .E. and DH. Ubelaker, editors, 1994. Standards for Data Collection from
Human Skeletal Remains, Proceedings of a Seminar at The Field Museum of
Natural History. Arkansas Archeological Survey Research Series No. 44.

F ayetteville: Arkansas Archeological Survey.

Boles, T., C. Snow, and E. Stover. 1995. Forensic DNA testing of skeletal remains from

mass graves: A pilot study in Guatemala. Journal of Forensic Sciences. 40:349-
355.

Gilbert, M.T.P., E. Willerslev, A.J. Hansen, 1. Barnes, L. Rudbeck, N. Lynnerup, and A.
Cooper. 2003. Distribution patterns of postmortem damage in human
mitochondrial DNA. American Journal of Human Genetics. 72:32-47.

Gotherstrom, A., M.J. Collins, A. Angerbjom, and K. Liden. 2002. Bone preservation
and DNA amplification. Archaeometry. 44(3):395-404.

Hagelberg, B, IO Gray, A.F. J effreys. 1991. Identification of the skeletal remains of a
murder victim by DNA analysis. Nature. 352:427-429.

Hall, T. 2004 July 4. BioEdit sequence alignment editor for Windows 95/98/NT/2K/XP.
Homepage. <http://www.mbio.ncsu.edu/BioEdit/bioedit.html> Accessed 2004
July 13.

Hedges, R.E.M. 2002. Bone diagenesis: an overview of process. Archaeometry.
44(3):319-328.

Hochmeister, M.N., B. Budowle, U.V. Borer, O. Rudin, M. Bohnert, R. Dimhofer. 1995.
Confirmation of the identity of human skeletal remains using multiplex PCR
amplification and typing kits. Journal of Forensic Sciences. 40:701-705.

Holland, M.M. and TJ. Parsons. 1999. Mitochondrial DNA sequence analysis—
validation and use for forensic casework. Forensic Science Review. 11:22-49.

Holland, M.M., C.A. Cave, C.A. Holland, and T.W.Bille. 2003. Development of a
quality, high throughput DNA analysis procedure for skeletal samples to assist

66

with the identification of victims from the World Trade Center attacks. Croatian
Medical Journal. 44(3):264-272.

Kaestle, F.A., and K.A. Horsburgh. 2002. Ancient DNA in anthropology: methods,
applications and ethics. Yearbook of Physical Anthropology. 45:92-130.

Kalmar, T., C.Z. Bachrati, A. Marcsik, and I. Rasko. 2000. A simple and efficient method

for PCR amplificable DNA extraction from ancient bones. Nucleic Acid Research.
28(12):e67.

McCutchen, L. 2003 Fall. Biology 2401: Human anatomy and physiology. Homepage.

<http://blcl .kilgore.cc.tx.us/kcapl/practical_2_photos.htm> Accessed 2004 July
28.

Micozzi, MS. 1991. Postmortem change in human and animal remains: A systematic
approach. Springfield: Charles C. Thomas.

Monson, K.L., K.W.P. Miller, M.R. Wilson, J .A. DiZinno, B. Budowle. 2002. The
mtDNA population database: an integrated software database resource for
forensic comparison. Forensic Science Communications. 4(2).
<http://www.fbi.gov/hq/lab/fsc/backissu/apri12002/miller1.htm> Accessed 2004
August 12.

O’Rourke D.H., G. Hayes, S.W. Carlyle. 2000. Ancient DNA Studies in Physical
Anthropology. Annual Review of Anthropology. 29:217-242.

Parsons, T.J. and V.W. Weedn. 1997. Preservation and recovery of DNA in postmortem
specimens and trace samples. In Forensic taphonomy: the postmortem fate of
human remains, ed. W.D. Haglund and M.H Sorg. Boca Raton2CRC.

Paabo, S., R.G. Higuchi, A.C. Wilson. 1989. Ancient DNA and the polymerase chain
reaction: The emerging field of molecular archaeology. Journal of Biological
Chemistry. 264:9709-9712.

Primorac, D., S. Andelinovic, M. Definis-Gojanovic, I. Drrnic, B. Rezic, M.M. Baden,
M.A. Kennedy, M.S. Schanfield, S.B. Skakel, H.C. Lee HC. 1996. Identification
of war victims from mass graves in Croatia, Bosnia, and Herzegovina by the use
of standard forensic methods and DNA testing. Journal of Forensic Sciences.
41:891-894.

Rogan, PK. and U. Salvo. 1990. Study of Nucleic acids isolated from ancient remains.
Yearbook of Physical Anthropology. 33: 1 95-214.

Shackleford, J .M. 1966. The ultrastructure of Mississippian and Archaic Indian bones

from various soil and drainage conditions. American Journal of Phys. Anthro.
25:291-98.

67

Solomon C .D. and N. Hasse. 1967. Histological and histochemical observations of

decalcified sections of ancient bones from excavations in Israel. Israel Journal of
Medical Science. 32747-54.

Smith, C.I. et al. 2003. The thermal history of human fossils and the likelihood of
successfirl DNA amplification. Journal of Human Evolution. 45:203-217.

Stout SD. and S.L. Teitelbaum. 1976. Histological analysis of undecalcified thin sections
of archaeological bone. American Journal of Physical Anthropology. 442263-69.

Tuross N. 1994. The biochemistry of ancient DNA in bone. Experentia. 50:530-535.
Ubelaker, D.H., E.B. Jones, Editors. Diane Beynon Landers, Associate Editor for
Archaeology. 2003. Human remains from Voegtly Cemetery, Pittsburgh,

Pennsylvania. Smithsonian Contributions to Anthropology Series, Number 46,
Smithsonian Institution Press, Washington, D. C.

68

I. ‘ l‘l‘llll‘lill'lollllilllr
I. r i -1.

."'llrl.|lll l -11

   

uitillliilliflll111111111