2°33 WNWIIIHIHIWIHHHIHIHIIHJIJIIIHIHWM

     

T—I
,I
_m

(2 2 / ‘H/I’

LIBRARIES
MICHIGAN STATE UNIVERSITY
EAST LANSING, MICH 48824-1048

This is to certify that the
thesis entitled

THE RECOVERY AND ANALYSIS OF MITOCHONDRIAL
DNA FROM EXPLODED PIPE BOMBS

presented by

MICHAEL EVERETT GEHRING

has been accepted towards fulfillment
of the requirements for the

MS. degree in Forensic Science

 

 

7 a”

Major Professor’s Signature
/ Z—// Z /C‘ 4/;
I l '

Date

MSU is an Affirmative Action/Equal Opportunity Institution

- .v--i-.-.-.~l..'.

 

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

 

FEB {1‘ 5 2008:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuepGS-sz

ABSTRACT

THE RECOVERY AND ANALYSIS OF MITOCHONDRIAL DNA FROM
EXPLODED PIPE BOMBS

By

Michael Everett Gehring

Explosives are involved in approximately 70% of terrorist incidents (Burke 2000).
Current methods of investigation have difficulty forming physical links between suspects
and explosive devices, and analysis of DNA left by the manufacturer may provide a
better alternative for their identification. It has been previously demonstrated that nuclear
DNA can be isolated from conflagrated pipe bombs (Bechaz et al. 2001; Esslinger et a].
2004), but successful recovery and profiling was quite limited. The goal of this research
was to determine the feasibility of recovering mitochondrial DNA (mtDNA) from
conflagrated pipe bombs. Eighteen subjects were recruited to handle a total of thirty eight
bombs. DNA was quantified using a quantitative PCR assay developed for this study,
amplified using semi-nested PCR in several short fragments, and then sequenced. Based
on the sequencing results, eighteen of the thirty eight samples could be assigned to a
single donor and seven more were assignable to a subset of three possible donors, while
12 could not be assigned. In addition, sequence information from sixteen of the eighteen
subjects was recovered from at least one of the two bombs handled. These results are
promising, and demonstrate the value of mtDNA analysis for suspect identification under

post-conflagration conditions.

ACKNOWLEDGEMENTS

I would like to thank my committee members for their time and patience in
helping me through the process of writing this. In particular, to my advisor, Dr. David
Foran: thank you for the endless hours of revisions and all of your help in working
through experimental hurdles. Also, to Lt. Shawn Stallworth: thank you for your time,
effort, and risk to life and limb in helping me to pursue this degree. I could not have done
this without you. Finally, Dr. Christina DeJong: thank you for giving your time and effort
to someone whom you had only just met.

I would like to acknowledge my 18 volunteers for their time and chaffed hands.
My appreciation goes out to Dr. Will Kopachik as well, who was gracious enough to
allow me the use of his laboratory.

I would also like to thank the Department of Homeland Security for their financial
support. In particular, I would like to recognize Laura Marty at Oak Ridge Associated
Universities, for her patience in getting me through the endless governmental red tape.

And finally, my deepest gratitude to my family and friends, in particular Alicia
Smith, who helped me through even the roughest of spots. Thank you for your patience

and unwavering support.

iii

 

 

TABLE OF CONTENTS

LIST OF TABLES ......... - ...................................................................... v
LIST OF FIGURES .............................................................................. vi
INTRODUCTION ............................................................................... 1
Advances in Forensic Applications of DNA Analysis ......................... 1
Mitochondrial DNA Background .................................................... 2
DNA Quantification ....................................................................... 3
Quantitative PCR ...................................................................... 4
PCR Inhibition ....................................................................... 5
Nested PCR ............................................................................... 6
Study Aims ............................................................................... 6
MATERIALS AND METHODS ............................................................. 8
Investigators ............................................................................... 8
Sample Decontamination and Preparation ............................................ 8
Bomb Fragmentation Analysis ................................................... 10
DNA Isolation and Purification .................................................... 10
QPCR Sample Quantification ............................ . ....................... 11
Sample Amplification and. Sequencing '. ......................................... 13
Statistical Analyses ..................................................................... 15
RESULTS ....................................................................................... 16
QPCR Sample Quantification ............................................................. 16
Bomb Fragmentation ..................................................................... 17
Correlation of Subject Traits and DNA Yield ................................. 18
QPCR Quantification of Inhibition ................................................... 18
Amplification of Samples Using Nested PCR ................................. 20
Correlation of QPCR Quantification and Nested PCR Amplification ......20
Sequencing and Donor Assignation ................................................... 22
DISCUSSION ........................................................................................ 24
MtDNA Recovery, Sequencing, and Contamination ................................. 24
QPCR Assay Performance ............................................................ 28
Considerations and Complications ................................................... 32
CONCLUSIONS .............................................................................. 34
APPENDICES .............................................................................. 35
REFERENCES .............................................................................. 39

iv

LIST OF TABLES

Summary of Sample Assignations and Contamination . . . .. .......................... 22

LIST OF FIGURES

Cloning of the QPCR Assay Internal Control .................................................... 12
DNA Concentrations Reported by QPCR Assay .......................................... 16
Bomb Fragmentation .............................................................................. 17
Quantity of Recovered DNA Across Bomb Fragmentation Levels ........................ 18
The Effects of Inhibitors Present in a Sample Assayed by QPCR ........................ 19
The Amplification of Samples Using Semi-Nested PCR ................................. 20
The Correlation of QPCR Quantification and Standard PCR Amplification ............... 21

vi

 

INTRODUCTION

In recent years, terrorism has become a priority in the American awareness. In
over 70% of terrorist incidents, explosives are used, most commonly in the form of
improvised explosive devices (Burke 2000). Of these, approximately 43% are pipe
bombs (AENRB Report 2004). These types of explosive devices are easy to construct out
of readily available materials, and can be quite effective. Pipe bombs have been involved
in a wide range of incidents, including use in the Olympic Park bombings in Atlanta and
by the Unabomber, among others (Schweitzer and Dorsch 1998; Burke 2000).

One of the major problems facing law enforcement agencies in addressing these
incidents is that most explosive devices are left on a time delay and leave behind limited
forensic evidence. It is therefore very difficult to determine possible suspects or link
individuals to the device based on the physical evidence left at the scene. Currently, in
order to provide a physical link between suspects and explosive devices, fragments of the
devices are only examined for fingerprints, though this rarely yields any evidence

(Stallworth personal communication).

Advances in Forensic Applications for DNA Analysis

With the advances made in forensic biology, analysis of DNA may offer a better
approach to analyzing this kind of physical evidence. Brief contact between a person and
object is enough to leave behind sufficient DNA for analysis and identification of that
individual (Bechaz 2001; Migron et a1. 1998; Van Oorschot and Jones 1997). Further,

there have been at least two studies into the use of short tandem repeats (STR3) in the

 

analysis of improvised explosive device fragments (Bechaz 2001; Esslinger et al. 2004).
In both studies, STRs were examined and had partial success, though the studies had
limited sample sizes (5 and 20 samples, respectively). Out of these, only a single full
nuclear DNA (nDNA) profile was generated, and several partial profiles. Two factors are
likely to be primarily responsible for the degree of success in DNA recovery and analysis
— the quantity of DNA on the object after conflagration, and the extent of resultant DNA
degradation. The conflagration of an explosive device generates large amounts of heat.
This can be both from the burning of the fuel, as well as from metal fatigue if a metal
container is used. This heat could have a detrimental impact on the integrity of any DNA
present. It is therefore likely that any DNA recovered would be highly fragmented. In
addition, only a small number of cells are likely to be transferred from the limited contact

between a person and the explosive device.

Mitochondrial DNA Background

Mitochondrial DNA (mtDNA) analysis offers a promising alternative to
examining nDNA markers. Because of its higher copy number, from hundreds to
thousands of copies per cell, and higher resistance to degradation than nDNA (reviewed
by Holland and Parsons 1999), mtDNA analysis may be a better approach for evidence
exposed to such harsh conditions. For example, mtDNA can be extracted from hair,
fingernails, or even ancient bone when nDNA cannot (reviewed in Merriwether et a1.
1994). In addition, because mtDNA is maternally inherited, a reference sample can be
obtained from not only the suspect, but also any maternally related family member. The

disadvantage of mtDNA analysis is the loss in individualizing power, though this may be

offset by any gain in the reliability of obtaining results. The problem of DNA degradation
that is likely to occur under the conditions of a bomb blast can also be addressed using
mtDNA. Because it is analyzed using sequencing, small overlapping fragments can be

used to construct a complete sequence.

DNA Quantification

Once a sample of DNA has been obtained from a piece of evidence, it is
important to know how much DNA is available for analysis and whether any inhibitors
are present that may interfere with the amplification process. Traditional forensic
methods for quantifying DNA involve slot blots, relying on a probe hybridizing to the
DNA and using a subsequent enzymatic reaction to estimate the quantity of DNA present.
Such techniques are time consuming and have lower detection limits of around 60pg
(Technical Bulletin http://www.promega.com). Quantitative polymerase chain reaction
(QPCR), also known as real-time PCR, offers an alternative approach and is a convenient
means to answer both questions. There have been numerous studies using QPCR to
quantify mtDNA (Alonso et al. 2003; Chui et al. 2003; Gourlain et al. 2003; Richar et al.
2003; Steuerwald et a1. 2000; von Wurmb-Schwark et al. 2002). In all cases, QPCR
assays showed a wide linear range (minimum of 5 orders of magnitude) of quantification
and high sensitivity, often on the single copy level. There are several advantages to using
QPCR, including the ability to target quantification to specific regions of DNA based on
research needs, less processing time and space required, and objective measures of

concentration values.

Quantitative PCR

QPCR works by detecting increases in amplified DNA product (the arnplicon)
during the amplification process. In standard PCR, the amount of DNA produced bears
little relationship to the starting amount of DNA, as a plateau exists when one or more
reagents become limiting and amplification ceases. At lower concentrations of DNA, the
rate of amplification proceeds in an exponential manner. During the exponential grth
phase of amplification, samples pass a preset threshold of amplification rate. This value is
taken as a sample’s critical threshold (C!) value, which is directly related to its starting
DNA concentration. In order to assess the relationship between C, and starting
concentration, DNA standards, samples of known concentration, are amplified at the
same time as unknown samples. A standard curve is then derived from a dilution series of
these standards by plotting the concentration of the standards versus their C. values.

Detection of the amount of amplicon in a reaction can be done in one of two
ways. Double stranded DNA in the reaction can do detected with a dye, such as Syber
Green, which intercalates between the strands of DNA, causing the dye to fluoresce when
excited. This fluorescence will increase in a linear relationship to the concentration of
double stranded DNA in a sample. The second method of detection is a probe that can
hybridize to a specific sequence in the amplicon. Depending on the nature of the probe, it
may fluoresce only when hybridized to the amplicon or after degradation of the probe.
This study used a Taqman probe for detection, which is a short length of DNA, typically
between 18—27 base pairs (hp), with a fluorescent molecule at the 5’ end (a “reporter”)
and a “quencher” molecule at the 3’ end. The reporter is excited throughout the

amplification process. However, when the probe is intact, the reporter is in close

' n.

 

proximity to the quencher, which absorbs any emissions from the reporter. When the
amplicon is present, the probe hybridizes to it and is digested by the 5’ exonuclease
activity of the polymerase, freeing both the reporter and quencher. This allows the
fluorescence of the reporter to be detected by the QPCR instrument. Since the target level
of fluorescence increases in a linear relationship to the concentration of the target

amplicon, the starting concentration of DNA can be calculated using a standard curve.

PCR Inhibition

In many samples, especially those derived from forensic sources, contamination
from inhibitors, such as blood and soil (Akane et al. 1994; Watson and Blackwell 2000),
can be a major hindrance to the amplification process. Inhibition is especially
problematic in QPCR, where different reaction kinetics between pristine standards and
inhibited samples can lead to inaccurate quantification (Tichopad et al. 2000; Wilson
1997). Deacetylated bovine serum albumin (BSA) has been shown to relieve PCR
inhibition resulting from a wide variety of inhibitors, both known and unknown (Kreader
1996) and may offer a solution to any inhibitors co-isolated with DNA. However, the
problem remains of determining whether a lack of amplicon production is due to
inhibition or insufficient template DNA. In some studies, this problem has been
addressed using an internal control in a QPCR assay (Raggam et a1. 2002; Siebert and
Larrick 1993). Internal controls are fragments of DNA added to each reaction, co-
amplified with the sample, and detected using an alternate probe providing a positive

control within a given reaction environment.

Nested PCR

Amplification of DNA from degraded samples for sequence or STR analysis can
be extremely difficult (Bender et al. 2004). However, it has been shown that DNA which
is degraded or has low copy number can be successfully amplified using nested PCR
(Strom and Rechitsky 1998), a technique in which two successive rounds of PCR are
used to amplify small amounts of DNA. In the first round, DNA is amplified under
standard conditions. In the second round, the reaction is supplemented with product from
the first round reaction, providing it with a template. This subsequent round utilizes
primers that sit internal to the first round primers, meaning that only fragments amplified
in the first round that contain the second round primer sites are amplified. This allows for
a greater number of PCR cycles while eliminating non-specific amplification products,
and serves as a powerful technique for amplifying very small copy number samples with
relatively high fidelity. A related technique is semi-nested PCR, in which the second
round of PCR uses one of the primers from the first round and a second primer internal to

the first set.

Study Aims

The feasibility of using mtDNA sequencing in the identification of suspects from
exploded pipe bomb fragments was examined in this study. Subjects handled pipe bomb
components, which were subsequently assembled and detonated. DNA was isolated from
swabs of the bomb fragments and quantified by QPCR assay. MtDNA was then amplified

in several small fragments via semi-nested PCR and the sequence determined.

Comparisons between these sequences and subject reference sequences were used to

assign bombs to subjects.

MATERIALS AND METHODS

Investigators

This study was carried out by three investigators, Michael Gehring (investigator
l) and Amy Barber (investigator 3) of Michigan State University, and Lt. Shawn
Stallworth (investigator 2) of the Michigan State Police bomb squad. Except where noted,

investigator 1 performed all actions.

Sample decontamination and preparation

Each pipe bomb was assembled using a l-foot zinc-coated, galvanized steel pipe
nipple of l-inch diameter, two l-inch diameter zinc-coated, galvanized steel end caps
obtained locally, a length of safety fuse (2 or 4 feet), and a Thermolite Connector ignition
cap (Ensign Bickford; Simsbury, CT). In order to sterilize them, all pipe components
were soaked in a 10% bleach solution for 1 hour, rinsed with distilled water, and placed
in a Spectrolinker XL-1500 UV Crosslinker (Spectronics Corporation; Westbury, NY)
for 10 minutes, turning halfway through for complete exposure. The fuse and Thermolite
Connector ignition cap were wiped down with a 10% bleach solution. Fuses were affixed
via a hole drilled in one end cap and secured with hot glue. The internal portion of the
fuse was stripped of its protective coating subsequent to gluing. Samples were sealed in
paper bags and randomly selected by subjects. Eighteen subjects participated in this
study, and a total of 51 bombs were prepared and detonated. For the samples used in the
study, each subject handled 2 bombs, except for 1 subject who handled 3 and another

who handled l, for a total of 38 pipe bombs. These were designated samples 9, 11—14,

l7, and 20—51 and were used for all subsequent processing and analyses. Samples 1—8,
10, 15—16, and 18—19 were used for protocol optimization (Appendix A). Subjects
removed the bomb components and handled them by screwing and unscrewing the end
caps for a total of 30 seconds, as well as giving buccal swabs for DNA reference.
Subjects also filled out a questionnaire regarding the condition of their hands, use of
moisturizers, and time since washing. Procedures involving human subjects were
approved by UCRIHS prior to recruitment and subjects signed consent forms before
participation. Different sets of identifying numbers for samples and subjects were
randomly assigned and the remainder of the experiment was performed blind.
Immediately before detonation, investigator 2 filled samples 3/4 full with SR 4756 (IMR
Powders; Shawnee Mission, KS), a single-base, smokeless gunpowder. Samples were
detonated at the Lansing, MI Fire Fighter Training Facility smoke room. Fragments were
collected by investigators l and 3 after each detonation and stored in paper bags at room
temperature. No face protection was used during the first half of collections, though it
was deemed advisable for health reasons to use particulate filter masks during the second
half of collections. All visible fragments were collected in a sweep of the room to avoid
cross—contamination between detonations. Although some fragments may have persisted
in the room and been collected afler a subsequent detonation, it is unlikely that they
would have been of sufficient size to be swabbed during processing. Gloves were worn

for all procedures.

Bomb fragmentation analysis

Each bomb was assigned a fragmentation .score based on how intact it was. Scores
included 4 levels, “low” for pipes with at least 75% of the pipe length intact, “medium”
for 50—75% of the length intact, “high” for 25—50% intact, and “complete” for anything

with less than 25% of the pipe length intact.

DNA isolation and purification

External bomb fragment surfaces still retaining their zinc coating were swabbed
first with a sterile cotton swab moistened with lOOuL of digestion buffer (20mM Tris,
SOmM EDTA, 0.1% SDS, pH 7.5), immediately followed by a dry cotton swab as per
Sweet et al.’s (1997) double swab method. Bomb fragments were handled only by non-.
swabbed sides. The ends of both swabs were then broken off and placed into a microfuge
tube with 400uL of digestion buffer and 6uL of Proteinase K. Tubes were vortexed,
centrifuged briefly, and incubated overnight at 56°C. Swab tips were then placed in spin
baskets, centrifuged at 14,000 rpm for 5 minutes and discarded. Liquid spun out of the
swab and any remaining from the incubation tube was combined and extracted using an
equal volume of phenolzchloroformzisoamyl alcohol (25:24: 1 ), followed by an equal
volume of chloroform. The aqueous layers were then transferred to Microcon YM-100
spin columns (Millipore; Billerica, MA) and centrifuged at 500 x g for 20 minutes. This
was followed by 2 washes of 300uL TE (lOmM Tris, lmM EDTA, pH 7.5). Samples
were resuspended in 20uL of TE and stored at -20°C. Gloves and protective clothing

were worn during processing. Halfway through processing, contamination became a

10

concern and for the second half of samples processed, surgical masks were also used.
Subject buccal swabs and sample DNAs were processed at this and all subsequent steps

separately to prevent cross-contamination.

QPCR sample quantification

Samples were quantified with a novel QPCR assay. A 118bp region of mtDNA
was amplified using the primers F16400 (5’-ACC-ATC-CTC-CGT-GAA-ATC-AA-3’)
and H16498 (Shields and Kocher 1991) at 900nM each. H16498 is a universal mtDNA
primer for vertebrates that anneals between the hyper-variable regions in the human
mitochondrial genome. Reactions were carried outwith Taqman Universal PCR Master
Mix (Roche; Branchburg, NJ) at a 1x concentration and deacetylated BSA at a final
concentration of 0.5mg/ml. The amplification program consisted of a 10 minute uracil-
DNA glycosylase denaturation step at 70°C, a 10 minute 95°C polymerase activation
step, a 2 minute 95°C DNA denaturation step, followed by 50 cycles of a 15 second 95°C
denaturation step and a 1 minute 60°C annealing and extension step. Quantification was
accomplished via a PAM-labeled Taqman probe, mtDNA-PAM (5’-FAM-CCT-CGC-
TCC-GGG-CCC-ATA-AC-TAMRA-3’) at 250uM. The assay also included an internal
control which used the same primers to amplify a 118bp fragment of lambda phage DNA
template and a lambda-specific probe for detection (see below for internal control
construction). The internal control probe, mtDNA-1C (5’-VlC-CGA-GCG-TGT-TTA-
TCG-GCT-ACA-TCG-TAMRA-B’) was at 250uM concentration and the internal control

template was at 1.67fM (~1000 copies). All reactions contained 1 uL of template in a

ll

25 uL total volume and were run in triplicate on an ABI 7700 alongside standards (see
below). All samples were analyzed using ABI Prism SDS 2.1 software.

The internal control was created by PCR amplifying a fragment of lambda DNA using
the primers mtDNA-IC-F (5’-GGA-TTC-ACC-ATC-CTC-CGT-GAA-ATC-AAC-GCC-
GGA-CTA-AGT-AGC-AAT-C-3’) and mtDNA-lC-R (5’-GAA-TTC-ACC-CTG-AAG-
TAG-GAA-CCA-GAA-GCG-AAC-CAA-TCG-AGT-CAG-T-3’), which have sites
complimentary to F16400 and H16498 and restriction enzyme sites in addition to the

lambda phage complimentary sequences (Figure 1). This amplicon was purified on a

Figure l. Cloning of the QPCR Assay Internal Control.

    

 

TTIIITIIIIIIIllllllllllllllllllllllllllllllllllllllll

pUC18

Iv

/

 

DilllllllllllllllllHllllllllHlllllllllllllllllllllllllllllillill

A 78bp stretch of lambda DNA was amplified using 46bp primers which included lambda-complimentary
sequence (black), F16400 or R16498 sequence (blue) and a restriction enzyme site (red). The resulting
amplified product was restriction digested and ligated into pUC18. This figure is presented in color.

Microcon YM-30 spin column, and cleaved with EcoRI and HindIII for 2 hours at 37°C,
followed by enzyme denaturation for 10 minute at 70°C. A pUC18 plasmid was digested
with the same restriction enzymes, and both the plasmid and amplicon were added to a
ligation reaction containing T4 1i gase and its corresponding buffer and incubated
overnight at 4°C. HBlOl chemically competent cells (Promega; Madison, WI) were
transformed with the ligation reaction, plated with X-Gal and IPTC on LB-Ampicillin

agar plates (SOug/ml Ampicillin), and grown overnight at 37°C. White colonies were

12

selected and screened for the insert by PCR using F 16400 and H16498. One positive
colony was selected and sequenced to determine the fidelity of the insertion.

Standards for the assay were created by amplifying an 864bp fragment of human
mtDNA using the Armed Forces DNA Identification Laboratory (AF DIL;
http://www.afip.org/Departments/oafme/dna) primers F15989 and R285. This amplicon
was purified with a Microcon YM-100 spin column as described above, and quantified
with a DU 520 UV/V is Spectrophotometer (Beckman Coulter; Fullerton, CA). Dilutions
of the amplicon were made with concentrations ranging from 8.784pg/uL to 0.8784ag/uL
at 1 order of magnitude increments (corresponding to approximately 107 to 10 copies/uL,
respectively). Standards were run without the internal control during optimization to
assess the linear range of the assay.

A series of lO-fold dilutions, from 100 to 105, of one of the samples of DNA
isolated from a pipe bomb was also analyzed with QPCR to examine the effects of
inhibition on C, values of the internal control. Samples were run in triplicate and

parameters were as described above except no BSA was added to the reactions.

Sample Amplification and Sequencing

Three regions of each sample were amplified by semi-nested PCR. All primer
sequences were obtained from AFDIL. The first half of hyper-variable region 1 (HV 1 ),
denoted HVl-l, was amplified with primers F15989 and R16322, and used R16281 for
semi-nested PCR. The second half of HVl (HVl-2) was amplified with F16144 and
R16410, and used F 16190 for semi-nested PCR. Hyper-variable region 2 (HV2) was

amplified using F82 and R484, and used F155 for semi-nested PCR.

13

Amplifications were performed with a 2 minute 94°C denaturation, followed by
38 cycles consisting of a 30 second 94°C denaturation step, a 1 minute 60°C annealing
step, and a 1 minute 72°C extension step. Amplifications involving R484 used the above
parameters, except substituting a 45°C annealing temperature. Reactions were carried out
in 20uL volumes, using 0.5 uL of template, primers at 2uM, 0.5mg/ml deacetylated BSA,
2.5mM MgClz, 200uM dNTPs, and Taq polymerase at lU/reaction. SuL of the reaction
was run on a 1.5% agarose gel and assigned a number, 0—3, based on band intensity, with
0 representing no band and 3 representing a high intensity band. A semi-nested reaction
was carried out for any sample failing to amplify to a mid-intensity (denoted 2) band.
Nested PCR was performed as above, using luL of the initial PCR reaction for template
and 24 cycles. Reference samples were amplified with F15989 and R569, using the above
parameters for 32 cycles. All PCR products were cleaned using Microcon-100 spin
columns as described above and then sequenced using a CEQ DTCS Quick Start Kit
(Beckman Coulter; Fullerton, CA) at a reaction volume of lOuL. The primers used for
semi-nested PCR were also used for sequencing samples, while reference samples were
sequenced with F15989, R16410, F15, and R569. PCR reagent blanks were run with each
reaction. Sequencing reactions were purified according to the manufacturer’s protocol
and analyzed on a CEQ 8000 Genetic Analyzer, as per the instrument protocol. Sample
sequences for HVl used a capillary separation time of 45 minutes, while sample
sequences for HV2 and all reference sequences were separated for 60 minutes. Sequences
were aligned and proofread using BioEdit Sequence Alignment Editor (Hall 1999).
Subjects’ complete mitochondrial sequence (haplotype) frequencies obtained from the

buccal swabs were determined using the FBI mtDNA database (Monson et al. 2002).

14

Haplotypes were compared to the forensic portion of the database, and included all
sequences covering at least the same window of bases, though due to differences in the
number of bases sequenced for subjects, the number of individuals compared against in

the database could vary.

Statistical Analyses

Statistical analyses were performed using R 1.9.1 statistical analysis software
(httpzllx-x'ww.r-proicctore’) and all used a=0.05. The data were checked for normality and
the quantities obtained from the QPCR assay were log transformed. The relationship of
DNA quantity to bomb fragmentation was evaluated using a type II ANOVA because of
the random error in both dependent and independent variables. A nested ANCOVAR was
run to determine if hand washing time and moisturizer use influenced the quantity of
DNA recovered. The analysis was restricted to non-contaminated samples, and an
interaction term was not included in the model. Quantification of inhibition was
evaluated with a linear regression model. Data from QPCR were averaged across
replicates and dilution factors were log] 0 transformed. A logistic regression model was
applied to the relationship between QPCR sample quantities, region amplified, and
amplicon band intensity. Three models were applied to examine the probability reaching
three different band intensities, given a quantity of DNA and region amplified. No
interaction terms were included in these models. For a review of these statistical

techniques, see Scheiner and Gurevitch (2001).

15

RESULTS

QPCR sample quantification

Sample replicates ranged in quantity from 184 copies/uL to 7770 copies/uL for
samples with detectable levels of mtDNA, though 19 replicates yielded no detectable
levels of mtDNA (Figure 2, Appendices B and C). The average quantity detected was
1287 copies/uL. Taking the average across replicates, samples ranged from 204
copies/uL to 6443 copies/uL, with 6 samples giving no detectable signal. The internal
control amplified consistently in the samples that yielded no detectable mtDNA
(Appendices 2B and 38). However, in 24 replicates, the internal control was not detected.
The average quantity of mtDNA in these samples was 2835 copies/uL, and ranged from
577 to 7770 copies/uL. The assay showed a linear range of detection of 7 orders of
magnitude when the internal control was not present but only 5. orders of magnitude when
the internal control was included. In these cases, standards containing 102 or fewer
copies/uL consistently failed to amplify but the internal control did amplify. In addition,
the internal control consistently failed to amplify in standards containing 104 or more
copies/uL.

Figure 2. DNA Concentrations Reported by QPCR Assay

 

 

 

 

 

 

 

 

8
"i
3’ 3‘3
§ 2
U —
93 -
.. ... i Mi Illnim'n
I I I fl
0 2000 4WD 5000 8000
Copies/UL

Each bin represents quantities spanning a 100 copy/uL range.

16

Bomb fragmentation

Fragmentation of the pipe bombs ranged from little fragmentation of the end caps
to complete destruction of the pipe (Figure 3). Fragment sizes were collected as small as
2 millimeters. Higher levels of fragmentation resulted in less useable surface area due to
abraded edges which could not be swabbed. Although the room was closed during
detonation, in some instances the door or ceiling vent were blown open and fragments
were propelled out of the confines of the room. These were discarded to prevent possible
cross-contamination. The relationship between the level of bomb fragmentation and
quantity of DNA recovered was examined, and though the overall model was not
significant (p=0.06l76), individual levels were. Both “high” and “complete” levels of
fragmentation were significantly different from low fragmentation in terms of quantity of
DNA recovered, with effect sizes of -0.6235 (p=0.0117) and -0.5269 (p=0.0448),
respectively (Figure 4). These results supported a trend of higher bomb fragmentation

corresponding to lower recovered DNA amounts.

Figure 3. Bomb Fragmentation

Low Medium High Complete

 

The fragmentation pattern of the bombs ranged widely, from very few fragments and the main body of the
bomb mostly intact to completely fragmented. The scale is indicated by the checkered marker. with each
square measuring 1 cm2. This figure is presented in color.

Figure 4. Quantity of Recovered DNA Across Bomb Fragmentation Levels.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

as | .._._
i2- , 1 ~ - I l
r. e l - -
33-4

Bomeragmcntation

Boxes enclose the 1“ through the 3rd quartiles, with a line intersecting the box at the mean. Brackets
indicate the 95% confidence interval of each grouping.

Correlation of subject traits and DNA yield

The time since washing of hands and the use of moisturizer had no statistically
significant effect on the quantity of mtDNA isolated, as measured by QPCR. The interval
since washing had an effect size of -0.097027 with a p = 0.2222, while the frequency of
moisturizer use had an effect size of -0.l34809 with a p = 0.4188. No interaction between

the terms was observed.

QPCR quantification of inhibition

Replicates of an undiluted bomb DNA sample failed to generate a signal when
amplified without BSA, both for mtDNA and the internal control, despite previously
having yielded a signal in the presence of BSA. Dilutions of the sample from 1:10 to
1:105 resulted in internal control signal, with decreasing control C. values associated with
an increasing dilution factor (Figure 5A) but no mtDNA signal. A linear regression

showed a statistically significant correlation between the log of the dilution factor of the

18

sample and the internal control C, value, with an effect size of-1.116 and p=0.0214

(Figure 5B). The model had an r-squared of 0.8189.

Figure 5. The Effects of Inhibitors Present in a Sample Assayed by QPCR.

1000 EH

 
 
 
 
   
    

1.000

1.1X10 El

2?

1.000

 

 

 

 

 

1.000
1.11!)
Cycle
('1—
A Q
8
a
G 3‘
O
5 s-
Q_ o
m I I I I U f f
1.0 1.5 2.0 2.5 3.0 3.5 4.0

logiddintion fiactor)

A: Lines show the internal control Ct values for a sample without BSA amplified using sample dilution
factors from 1 to 105 at 1:10 increments (green, yellow, red, and purple plots, respectively). The red line
indicates the detection threshold. B: A regression plot of the log of dilution versus the CI value of the
internal control.

Amplification of samples using nested PCR

Initial PCR amplification yielded sufficient product for sequencing from 15 of 38
samples for HVl -1, 5 of 38 samples for HVl-2, and insufficient product for sequencing
in all samples for HV2. Semi-nested PCR of all samples with insufficient first-round

amplicon produced sufficient product for sequencing in all cases (Figure 6).

Figure 6. The Amplification of Samples Using Semi-Nested PCR.

1 st Round

1
I
.r
A

2nd Round

34 35 36 37 38 39 40 41 42 43 44

 

All first round products were assigned a value, 0-3, based on the appearance of band intensity, with a 0
being no band (e.g. sample 34), a 1 being a barely perceptible band (e.g. 42), a 2 being a mid-intensity band
(e.g. sample 35) and a 3 being a strong band (eg. sample 33). Only samples classified as a 0 or 1 were
amplified in a second round of semi-nested PCR.

Correlation of QPCR quantification and nested PCR amplification

There was a statistically significant correlation between the quantity of DNA as
determined by QPCR and the probability of seeing bands of differing intensities from
PCR amplification from the first round of nested PCR (Figure 7). The first correlation
modeled the probability of seeing any band (high, medium, or low intensity) on a gel as a

function of initial mtDNA concentration and had an r—squared of 0.2762. The effect size

20

Figure 7. The Correlation of QPCR Quantification and Standard PCR

 

 

 

 

 

Amplification.
ea 0 o o o e a o HOE"
3'
S «a-
' O
a 4* m
D.
s 34 5 54
b V... a:
E ° _S' .5“
3 ~ .
E "1..
O
2
O
o“ . . . . “i
1.5 2.0 2.5 3.0 3.5

logio(c0pies/uL)

Data points are plotted by copy number and band intensity. Black points represent amplification products
from region HVl-l while red points represent products from HVl-2. The black and red lines are the
logistic regression curves of this analysis for HVl-l and HVl-2, respectively. Solid lines represent the
probability of seeing any type of band as a function of copy number (note the log transform). Dashed lines
represent the probability of seeing a mid or high intensity band versus seeing a light or no band as a
function of copy number. The dotted lines represent the probability of seeing a high intensity band versus
seeing anything else as a function of copy number. This figure is presented in color.

of concentration was -2.7090 (p= 0.0001 17), and the region amplified had an effect size
of -0.4785 (p= 0.400452). The second correlation modeled the probability of seeing a
medium or high intensity band on a gel as a function of initial DNA concentration. The r-
squared of this model was 0.2513, with concentration having an effect size of -2.7660 (p=
0.00130). In this model, the region amplified also had a statistically significant effect,
with an effect size of 1.3675 (p= 0.02716). The final model was based on the probability
of seeing only a high intensity band as a function of DNA starting concentration. The r-
squared was 0.2397, and both concentration and region amplified were statistically

significant with effect sizes of -2.6947 (p= 0.0141) and 1.8021 (p= 0.0431), respectively.

21

Interestingly, although the region HVl-l covered a span of 333bp, 67 bases longer than
HVl-2, mid and high intensity bands were seen at lower concentrations of DNA (Figure

7), though there was no statistical difference for low intensity bands.

Sequencing and donor assignation

Eleven of the 18 subjects had haplotypes not present in the FBI mtDNA database,
which had approximately 1900 sequences covering the same regions. Non-unique
haplotypes ranged in frequency from 0.10—1.20%, with an average of 0.67% in the
complete database. Frequencies were much higher when restricted to the subject’s racial
group, ranging from 0.49%—7.84%, with an average of 4.27%. Of the 3 investigators, 2
had unique haplotypes, while the third had a frequency of 0.05% in the database (0.13%
within racial group). Sequences were obtained for all bomb DNA samples, with an
average read length of 256 bases for HVl (n=3 8) and 283 bases for HV2. (n=22)
(Appendix D). Eighteen of the 38 samples were correctly assigned to a single donor, 7
were narrowed to groups of three donors due to shared haplotypes, and 12 were unable to

be assigned due to contamination (Table 1).

Table 1. Summary of sample assignations and contamination.

 

 

 

 

Donor Number of Contamination Number of
Assignation samples Percent Occurred samples Percent
assignable to single donor 18 47.37% prior to detonation 2 5.26%
assignable to subset of donors 7 18.42% during collection 2 5.26%
not assignable 12 31.58% not assignable 14 36.84%
incorrectly assigned 1 2.63%

 

 

 

 

All percentages are based on a total of 38 samples. In all cases where an assignation was made to a subset

of donors, only 3 donors matched the sample.

22

 

A single sample was assigned to the wrong subject. This sample showed contamination
from investigator 1 and had substitutions and insertions consistent with a subject other
than the donor, possibly indicating cross-contamination. Sequence information was
recovered for 16 of the 18 subjects off of at least one bomb and 7 had sequence
information recovered from both of their samples. Out of 38 samples, 18 showed some
level of contamination (Appendix D). Fourteen of these samples showed a mixture of
haplotypes, while 4 of them showed a single haplotype. Of these, 6 were assignable based
on the clear presence of an investigator haplotype and a haplotype consistent with one of
the possible donors. Of the 4 single haplotypes, 3 were consistent with one of the
investigator haplotypes. In the majority of cases, the contamination could be attributed to

one of the investigators, though in 5 cases, a unique haplotypes was observed.

23

DISCUSSION

MtDNA recovery, sequencing, and contamination

This study demonstrated that usable DNA could be recovered from pipe bomb
fragments, even after being subjected to the extreme conditions of low-yield explosive
conflagration. With an individualizing success rate of 47.37%, the use of mtDNA showed
a much higher potential than the 5% success rate shown in an earlier analysis of STRs for
the identificatioii of suspects (Esslinger et al 2004). These results are more compelling
when considering the fact that 5 of the subjects had non-unique haplotypes within the
group of subjects. The increase in success rate stems from at least three separate issues.
The first, as discussed in the introduction, is the higher relative copy number of mtDNA
than single copy STRs used in forensic analysis. Secondly, differential degradation may
be increasing the relative copy number of mtDNA to nDNA. It is assumed that shed
epithelial cells are being deposited on the bomb, and it has been shown for a wide variety
of dead biologic material, including bone, fingernails, and hair, that mtDNA is more
likely to be recovered from these materials (reviewed in Holland and Parsons 1999).
Perhaps the more important difference is the ability of the investigator to elicit
information from mtDNA by changing a variety of parameters that are not possible with
STRs. Government and private laboratories currently rely on commercial kits for the
analysis of STRs. This leaves the investigator little room for adjustment based on the
nature of the sample. In degraded samples, larger loci are typically lost (Bender et al.

2004). Often, the only parameters that can be changed to compensate for this are the

starting concentration of DNA used and sample injection time into the DNA analysis

24

instrument. In the sequencing of mtDNA, nested PCR and small overlapping fragment
amplification for complete sequence construction can be used for degraded samples. The
power of nested, or even semi-nested PCR, as was used in this study, was demonstrated
by the clean amplification of samples which were originally unable to be seen on an
agarose gel after 38 PCR cycles (Figure 6). Likewise, the power of amplifying shorter
overlapping DNA fragments was demonstrated by the much higher success rate for
sequencing similar total read lengths of HV 1 (256bp, n=3 8), which was amplified as two
separate fragments, compared to HV2 (283bp, n=22), amplified as a single fragment.
Taken together, these strategies illustrate the power of mtDNA analysis for samples taken
from post-conflagration materials.

Unfortunately, the power of discrimination is limited for mtDNA compared to
nDNA, making this approach less robust for forensic applications. Probabilities of non-
uniqueness for STRs can be put into the one in several trillion range because each locus
assorts independently, meaning that one can multiply the probability of having a
particular allele at each unlinked locus across all loci to determine the probability of
having all of those alleles together. For example, if someone had alleles at two
independent loci, each with a frequency of 10% in a given population or ethnic group,
then the probability of randomly finding another person with both would be the product
of the allele frequencies, 1%. The same is not true for mtDNA. Unique features in
mtDNA are linked and therefore a haplotype is equivalent to a single, highly polymorphic
locus. In addition, because mtDNA is directly inherited, all individuals related through
maternal lineage will have identical haplotypes. This means that the probability of

finding an identical haplotype is functionally determined by its frequency in a database

25

and can only be used to implicate a group of people sharing that haplotype (e. g. siblings),
rather than an individual. Currently, the F BI’s mtDNA forensic database has less than
5,000 entries, which limits statements about frequency to about 1 in 5000, or 0.02%. It
should be noted that even when sharing a haplotype, subjects varied in frequency in the
database. This is due to the fact that they had different ranges of bases sequenced,
causing more or fewer sequences to be included for comparison from the database,
altering their frequency. In this study, 7 samples were assignable to one of two groups of
3 donors, 4 samples fiom one group and 3 samples from the other (Appendix D). All of
these samples originated from 5 subjects, 3 shared one haplotype, and 2 shared another.
One other subject had a single substitution difference from the group of 2 subjects
sharing a haplotype which fell outside of the range of sequenced bases in samples. Had
there been no contamination, 12 of the 38 samples could have been narrowed down to
one of two groups of 3 donors each, but as it was, 5 of the samples originating from these
6 subjects were not assignable due to contamination.

The second drawback to using mtDNA and approaches to dealing with low copy
number samples is the high sensitivity and chance for contamination. In this study, 47%
of the samples showed some type of contamination. Materials were decontaminated
before subject contact, and investigators wore gloves for each procedure in the study.
About halfway through bomb detonations, investigators began wearing particulate filter
masks during collection and surgical masks for bomb fragment processing. Despite these
precautions, samples still became contaminated from a variety of sources. There did not
appear to be a difference in the amount of contamination observed between samples

processed with and without masks. The most prevalent contamination (50% of

26

contamination, n=9) was from investigator 1, who decontaminated materials, collected
and swabbed fragments, and ran all subsequent analyses. Unfortunately, it is impossible
to determine at which step the contamination occurred. However, during all amplification
steps, no contamination was detected in reagent blanks. This suggests that contamination
occurred during the bomb preparation stage, bomb fragment collection, swabbing, and/or
DNA purification phases. Contamination occurred during bomb setup and fragment
collection, as investigator 3’s contact with materials was limited to this period and
accounted for 11% (n=2) of contaminated samples. More surprising was contamination
from investigator 2 (11% of contamination, n=2), who only had contact with materials
pre-conflagration when filling them with smokeless powder, placing them and lighting
fuses, a period of time much less than either of the other investigators. In 28% (n=5) of
contamination cases, no source for the contamination could be determined. This could
indicate a failure of decontamination methods, transfer of contaminants at some point
during construction, detonation, or processing, or some combination thereof. Possible
sources include DNA present in the room where detonations were carried out, other bomb
squad personnel on the scene, or persons that entered the lab in which samples were
processed.

This approach is very feasible in regards to detecting individuals coming into
contact with explosives prior to detonation. In 2 of the 38 bombs, extremely limited
contact while wearing latex gloves pre-conflagration was sufficient to deposit enough
DNA to obtain a haplotype. However, attention must be paid to the care put into the
handling of evidence and subsequent processing. Gloves alone are insufficient barriers to

prevent contamination, as seen by contamination from investigator 2. As such, it would

27

be advisable to use facemasks and even disposable sleeves during collection, and all

subsequent processing should be performed in a sterile isolation chamber.

QPCR assay performance

The QPRC assay for sample quantification performed well and was predictive of
amplification success. As can be seen in Figure 6, QPCR concentrations correlated with
the band visualization results from standard PCR. R-squared values were significant, but
only accounted for about a quarter of the variation seen for gel band visualization. This
could be due to a variety of factors, including pipetting error while setting up the PCR
reactions, the subjective nature of the band intensity scoring, and the unknown
relationship of PCR target size and amount of areas spanning that fragment remaining
after degradation. Band intensity is a relative measure, not quantitative, and can be
affected by a variety of factors, including staining intensity and sample loading variation,
among others.

Likewise, the QPCR assay, while yielding a value for mtDNA concentration, is
really measuring the number of fragments that span the target 118bp sequence. While
there is a correlation between amplification success and QPCR concentrations, it is
impossible to make statements concerning the number of fragments spanning size regions
other than 118bp. The assay measured a particular region of 118bp, which is not
indicative of the concentration of other fragment sizes. However, in the course of this
study all samples were subjected to equivalent environmental conditions, and while no
statement is made concerning the actual concentrations of samples, it is assumed that

degradation was fairly consistent among samples. Therefore the values obtained from the

28

QPCR assay can be used as relative measures of DNA concentration, independent of
fragment size. In addition, QPCR is variable in its results, necessitating the need for
replicates. Despite these confounding effects, the relationship between the QPCR and
band visualization data was significant and consistent across regions, as would be
expected for different estimations of the same sample concentration. It should also be
noted that despite having sample concentrations, equal volumes of sample DNA were
used in the nested PCR reactions. This was done to facilitate comparison between QPCR
values and standard PCR results, as well as contributing to the validation of the QPCR
assay. To this end, the QPCR assay did show a significant correlation to amplification
success (Figure 7).

Another interesting observation from this study was the relationship between the
region of mtDNA amplified and the concentration needed for successful amplification.
Although the region HVl-l covered a bigger region than HV1-2, mid and high intensity
bands were seen at lower concentrations of DNA for HVl-l (Figure 6). These differences
were statistically significant, though there was no significant difference between the two
regions in the probability of at least a faint intensity band being present. This may be
indicative of differing efficiencies of primer sets, in other words the different affinity of
primers to bind to the template DNA under a particular set of conditions. These
efficiencies can be altered because of base composition, mismatched bases, and PCR
annealing temperature. Parameters for these amplifications were likely closer to the
optimal conditions for the primers amplifying HV 1-1 than HV1-2, resulting in the better
amplification success. It should be noted, however, that amplicon size is also a critical

factor in the probability of success (Bender et al. 2004). No samples had visible bands for

29

the amplification of HV 2, the largest fragment amplified, during the first round of
amplification, despite several showing strong bands for HVl-l and HVl-2. It is doubtful
that primer efficiencies alone would have been responsible for that disparity in
amplification success. This situation is more likely due to a combination of amplicon size
and primer efficiency differences.

The incorporation of the internal control into the QPCR assay was effective in
establishing that sample amplification failure was not due to inhibition of the reaction.
This information is useful both for establishing confidence in the concentration reported
by the assay for a sample and the effectiveness of sample purification procedures.
Unfortunately, the internal control introduced some complications into the assay. Because
the internal control utilized the same primers as those used for mtDNA amplification, it is
likely that whichever DNA was at the higher concentration out-competed the lower
concentration DNA in primer binding, and therefore amplified preferentially when large
differences in copy number between the internal control and target sequence existed. This
could be seen in a number of ways. The internal control consistently failed to amplify in
standards containing an order of magnitude or more target copies than the internal
control, as it also failed to amplify in 24 of the pipe bomb samples containing higher
concentrations of mtDNA. In addition, standards containing a tenth or less copies than
the internal control failed to amplify. The minimum concentration of mtDNA reported by
the assay was 180 copies/uL, a value close to the order of magnitude difference at which
standards failed to amplify. Taken together, these data suggest that an order of magnitude
difference between 2 target molecules is sufficient to suppress amplification of the less

concentrated DNA.

30

The internal control did, however, prove useful in quantifying inhibition.
Amplification of one sample without BSA showed a linear relationship between the log
of the dilution factor (and therefore the relative concentration) of inhibitors and the Ct
value of the internal control from the QPCR assay. This means that it may be possible to
quantify the level of inhibition and calculate actual concentrations given the level of
inhibition and concentration reported by a QPCR assay. While this method could not be
utilized simultaneously with sample quantification in this study because of the
competitive inhibition problem, retooling of the assay to amplify the internal control and
sample with different primers would make this approach feasible. This opens up several
possible avenues of study for the future. Currently, DNA concentrations fi'om
amplification-based techniques do not take into account the influence of inhibition in the
reaction. Even when an inhibition-relieving compound like BSA is used to lessen PCR
inhibition, it is unknown what the reaction efficiency is, and therefore any measure of
concentration is only an “effective” concentration, or equivalent concentration of DNA
which, if amplified at full efficiency, would show the same characteristics as the sample
in question. This is useful when the aim is determining concentration for use in a
reaction, but effective concentration is much less useful for studies looking at total DNA
recovered. The goal in a recovery study is to determine how much DNA can be extracted
from a particular source. Effective concentrations only provide a measure of how much
DNA can be amplified as influenced by the other materials co-isolated with the DNA,
and are not reflective of the true concentration of DNA. The second generation assay
could therefore be used in determining the effectiveness and limitations of various

inhibition relievers, such as BSA or commercial enhancers, examining the relationship

31

between sample source and level of inhibition, or obtaining true concentrations, rather
than effective concentrations, of samples and relating those values to sample type or

source.

Considerations and complications

There were several factors explored in this study that could have affected the
quantity of DNA recovered from bomb fragments, including characteristics of the
subjects and of the bombs themselves. No relationship was found between the time since
hand washing or use of moisturizer by subjects and the quantity of mtDNA reported by
the QPCR assay. Intuitively, both of these factors should affect the number of epithelial
cells that could be shed during contact with an object, but neither had a significant effect.
There are at least two confounding factors in this analysis. The first was that it has been
shown that some people are much more likely to deposit cells by casual contact than
others (Alessandrini et al. 2003). It would be useful in future studies to assess the status
of each subject to determine their likelihood of shedding cells and relate that to the
quantities of DNA isolated from samples, however, this study did not take this variable
into account. The second problem was the unbalanced nature of the data. Because
subjects reported these parameters, certain conditions had large numbers of observations,
such as daily use of moisturizer, while others had very few, such as occasional use of
moisturizer. This greatly reduced the power of the analysis. Because this analysis was
secondary to the main focus of the study, it was not deemed crucial to control these

parameters.

32

The influence of the level of bomb fragmentation levels on mtDNA quantity
recovered was also examined. Although the overall relationship of fragmentation to DNA
recovered was not significant, the data showed a clear trend of increasing quantity of
DNA recovered with decreasing bomb fragmentation (Figure 3). This trend is likely due
to the quality of fragment surfaces. In highly fragmented bombs, the zinc coating was less
likely to remain intact and fragments were more likely to have been abraded by contact
with the detonation room walls. This resulted in less surface area for swabbing for more
fragmented bombs. In addition, with smaller fragments, more of the surface was closer to
an edge, which tended to be ragged and cracked, making it more difficult to swab the

complete surface.

33

CONCLUSIONS

The results of this study demonstrate the value of mtDNA analysis as a source of
information under post-conflagration conditions. Not only were almost half the samples
assignable to a single donor, but the vast majority of subjects left recoverable DNA, even
after conflagration. This is a marked improvement over using nDNA markers. In
addition, the QPCR assay can provide valuable information for sample quantification and
inhibition, as well as be predictive of amplification success in standard PCR. Even with
high levels of degradation, sequence analysis is possible with mtDNA utilizing nested
PCR and sequence construction from multiple overlapping fragments. It is clear that this
approach is extremely sensitive, not only for picking up sequence information, but also
contamination. As such, extensive precautions to guard against the introduction of
additional DNA into the evidence should be taken in future applications. Despite these
complications the exclusion of large numbers of suspects, or even probable identification

from a limited pool of suspects can be achieved using this mtDNA analysis approach.

34

APPENDIX A

Fifty one bombs were handled, assembled and detonated. Samples from bombs l—
8, 10, 15—16, and 18—19 were used to test and optimize the procedures described in this
study. Samples 1 and 2 were non-handled controls used to evaluate the decontamination
procedure. Along with samples 6—8, they were processed as described in the methods and
then purified with Wizard DNA Cleanup System (Promega; Madison, WI). These
samples were then amplified (standard, not nested, PCR) and sequenced as described
above. Samples 3—5 were spotted with Lambda DNA and used to evaluate the recovery
potential of the swabbing technique for a known amount of starting DNA. These were
processed as above and analyzed by gel electrophoresis using a 1% agarose gel. Samples
10, 15—16, and 18—19 were processed as above. Sample 19 was supplemented with
positive control DNA and PCR amplified to determine the effects of sample inhibition.
Samples 15—16, and 18 were amplified with varying concentrations of BSA. Sample 10
was used to examine the effects of amplicon size on amplification success. Samples 10,
15—16, and 18—19 were then amplified by nested PCR to optimize primer combinations,

and, where successful, sequenced as described above.

35

APPENDIX B

1.000E-1

 

1000

lOOOE-I

 

é lOOOE-Z
<
10005-3
10005-4

1000E-5

 

 

1.0008-1

1000

lOOOE-l

 

ai 1000 E-2
<
10005.:

1000154

 

1.000 E-5

 

 

LWE‘I .. . ....'_. 1'..1. .4.*.‘l;;::..:1'41._...’ti;.*i;:..:.;.

 

 

 

 

 

 

 

 

1.052 1015'] IOEH 1.0E05 10E46 IOEt'l

QumlrtylcopresluL)
QPCR quantification data for samples 9- 38 A. Amplification plot of mtDNA. The red line indicates the
detection threshold B. Amplification plot of internal control C: Amplification plot of standards from 10 —
107 copies/u] (right to left). D: Standard curve generated from standards. Samples are indicated by a red
“X” while standards are denoted by blue squares. This appendix is presented in color.

36

APPENDIX C

1.000EOI

 

1000

1.000E-1

 

 

.5 1.000 5.2
Q
10005.:

1.00054 , :,

 

1.000 E-5
0

 

 

I'mOEfl .:.:::;:i:,:.:::::::""

 

 

 

 

 

 

 

 

 

 

 

 

 

17

 

IOEH 1052 105+; IOEM 105+: 105% 10547
Qimtrtyicoprw‘uL)
QPCR quantification data for samples 39- 51. A. Amplification plot of mtDNA. The red line indicates the
detection threshold. B: Amplification plot of internal control. C: Amplification plot of standards from 10 —
107 copies/u] (right to left). D: Standard curve generated from standards. Samples are indicated by a red
“X” while standards are denoted by blue squares. This appendix is presented in color.

37

APPENDIX D

 

was
narrowed down to. Each was assigned a unique color, which was also used for sample assignation.
Contamination indicates if any non—donor sequence was detected and, when possible, the number of the
investigator it was assigned to. “U” indicates a unique sequence of unknown origin. Sequence ranges are
numbered using the Anderson convention (Anderson et al.. 1981). This appendix is presented in color.

38

REFERENCES

Akane A, Matsubara K, Nakamura H, Takahashi S, Kimura K.~ 1994. Identification of the
heme compound copurified with deoxyribonucleic acid (DNA) from bloodstains, a

major inhibitor of polymerase chain reaction (PCR) amplification. J Forensic Sci.
39:362-72.

Alessandrini F, Cecati M, Pesaresi M, Turchi C, Carle F, Tagliabracci A. 2003.
Fingerprints as evidence for a genetic profile: morphological study on fingerprints

and analysis of exogenous and individual factors affecting DNA typing. J For Sci. 48:
586-92.

Alonso A, Martin P, Albarran C, Garcia P, Garcia 0, Fernandez de Simon L, Garcia-
Hirschfeld J, Sancho M, de la Rua C, Femandez-Piqueras J. 2003. Real-time PCR '

designs to estimate nuclear and mitochondrial DNA copy number in forensic and
ancient DNA studies. For Sci Int. 139:141-149.

Anderson S, Bankier AT, Barrell BG, de Bruijn MH, Coulson AR, Drouin J, Eperon IC,
Nierlich DP, Roe BA, Sanger F, Schreier PH, Smith AJ, Staden R, Young IG. 1981.
Sequence and organization of the human mitochondrial genome. Nature. 290:457-65.

Arson and Explosives National Repository Branch Report. 2004. 2004 Third Quarter‘
Statistical Report. 05-02:1-1 1.

Bechaz H, and The Australian Federal Police. 2001 The profiling of DNA of fired
cartridge cases and pipe bomb fragments [thesis]. Canberra, Australia: Univ. of
Canberra.

Bender K, Farfan MJ, Schneider PM. 2004. Preparation of degraded human DNA under
controlled conditions. Forensic Sci Int. 139:135-40.

Burke R. 2000. Counter-terrorism for emergency responders. CRC Press LLC. Boca
Raton, FL. Pp. 115-138.

Chui RWK, Chan LYS, Lam NYL, Tsui NBY, Ng EKO, Ranier TH, Lo YMD. 2003.
Quantitative analysis of circulating mitochondrial DNA in plasma. Clin Chem.
492719-26.

Esslinger KJ, Siegel JA, Spillane H, Stallworth S. 2004. Using STR analysis to detect
human DNA from exploded pipe bomb devices. J Forensic Sci. 49:481-4.

Gourlain K, Amellal B, Ait Arkoub Z, Dupin N, Katlama C, Calvez V. 2003.

Quantitative analysis of human mitochondrial DNA using a real-time PCR assay. HIV
Med. 4:287-92.

39

Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41 :95-98.

Holland MM, Parsons TJ. 1999. Mitochondrial DNA sequence analysis — validation and
use for forensic casework. For Sci Rev. 1 1: 21-51.

Kreader, CA. 1996. Relief of amplification inhibition in PCR with bovine serum albumin
or T4 gene 32 protein. Appl Environ Microbiol. 62:1102-6.

Merriwether DA, Rothhammer F, Ferrell RE. 1994. Genetic variation in the new world:
ancient teeth, bone, and tissue as sources of DNA. Experientia. 50:592.

Migron Y, Hocherrnan G, Springer E, Almog J, Mandler, D. 1998. Visualization of
sebaceous fingerprints of fired cartridge cases: a laboratory study. J Forensic Sci. 43:
543-548.

Monson KL, Miller KWP, Wilson MR, DiZinno JA, Budowle B. 2002. The
mitochondrial DNA population database: an integrated software and database
resource for forensic comparison. Forensic Sci Commun.4.

Raggam RB, Leitner E, Milhlbauer G, Berg J, Stocher M, Grisold AJ, Marth E, Kessler
HH. 2002. Qualitative detection of Legionella species in bronchoalveolar lavages and

induced sputa by automated DNA extraction and real-time polymerase chain reaction.
Med Microbiol Immunol. 191 :1 19-25.

Richard ML, Frappier RH, Newman J C. 2003. Developmental validation of a real-time
quantitative PCR assay for automated quantification of human DNA. J For Sci.
48 : 1 04 1 -6.

Scheiner SM, Gurevitch J. Design and analysis of ecological experiments. Oxford
University Press. New York, NY. Pp. 134-157, 197-216.

Schweitzer GE, Dorsch CC. 1998. Modern mutations of global terrorism. The Bridge.
28:1-9.

Shields F, Kocher T. 1991. Phylogenetic relationships of North American ursids based on
analysis of mtDNA. Evolution. 45:218-221.

Siebert PD, Larrick J W. 1993. PCR MIMICS: competitive DNA fragments for use as
internal standards in quantitative PCR. Biotechniques. 14:244-9.

Stallworth, F/Lt. Shawn E. Bomb squad commander, Michigan State Police. Aug. 20,
2004.

40

Steuerwald N, Barritt JA, Adler R, Malter H, Schimmel T, Cohen J, Brenner CA. 2000.
Quantification of mitochondrial DNA in single oocytes, polar bodies and subcellular
components by real-time rapid cycle fluorescence monitored PCR. Zygote. 82209-15.

Strom CM, Rechitsky S. 1998. Use of nested PCR to identify charred human remains and
minute amounts of blood. J Forensic Sci. 43:696-700.

Sweet D, Lorente M, Lorente JA, Valenzuela A, Villanueva E. 1997. An improved
method to recover saliva from human skin: the double swab technique. J Forensic
Sci. 42:320-2.

Tichopad A, Didier A, Pfaffl MW. 2004. Inhibition of real-time RT-PCR quantification
due to tissue specific contaminants. Mol Cell Probes. 18:45-50.

Van Oorschot RAH, Jones MK.1997. DNA fingerprints from Fingerprints. Nature. 387:
767.

Von Wurmb-Schwak N, Higuchi R, Fenesh AP, Elfstroem C, Meissner C, Oehmichen M,
Cortopassi GA. 2002. Quantification of human mitochondrial DNA in a real time
PCR. For Sci Int. 126:34-39.

Watson RJ, Blackwell B. 2000. Purification and characterization of a common soil
component which inhibits the polymerase chain reaction. Can J Microbiol. 46:63 3-42.

Wilson, IG. 1997. Inhibition and facilitation of Nucleic Acid Amplification. Appl
Environ Microbiol. 63 :374 1 -5 1 .

41

   

iiiiiiiiiiii