USF Libraries
USF Digital Collections

A phylogenetic approach for the study of variation and determination of population affiliation of indigent human skeleta...


Material Information

A phylogenetic approach for the study of variation and determination of population affiliation of indigent human skeletal remains
Physical Description:
Wetherington, Hattie Bea
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:


Subjects / Keywords:
Mitochondrial dna
Molecular forensics
Discriminant analysis
Dissertations, Academic -- Biology -- Masters -- USF   ( lcsh )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


ABSTRACT: Mitochondrial DNA (mtDNA) has played a major role in human population studies over the past decade due to its maternal inheritance and negligible recombination (Macaulay, 1999). The mtDNA control region has been the focus of these studies due to the highly polymorphic nature of this non-coding region. Forensic scientists also use mtDNA to help determine the identity of missing individuals when nuclear DNA is not present. However, when skeletal remains are unclaimed, identification becomes near impossible. Therefore, mtDNA can play a valuable role in identification in terms of population affiliation, especially in conjunction with morphological analysis. The goals of this research were two-fold: 1) to determine population affiliation of unknown skeletal samples using phylogenetics and 2) to find a method of extraction that leaves a majority of the remains intact.This research depended on the donation of samples from sixteen skeletal remains from the Hillsborough County Medical Examiners office. Mitochondrial DNA from ten of these cases were extracted, amplified, and sequenced in order to determine population affiliation via phylogenetic analysis of hypervariable region I (HVR I). These sequences were aligned and compared to that of sequences in a pre-existing mtDNA control region database (Handt, 1998). The crania of the skeletal remains were measured and subsequently analyzed by the forensic anthropology program FORDISC 2.0 to morphologically determine population affiliation. A secondary morphological analysis included input of the measurements into SPSS, a statistical program package, as a separate discriminant function assessment. This analysis was dependent on a database of craniometrics from known individuals (Jantz, and Moore-Jansen, 2000).
Thesis (M.S.)--University of South Florida, 2005.
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Hattie Bea Wetherington.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 86 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001680954
oclc - 62407241
usfldc doi - E14-SFE0001167
usfldc handle - e14.1167
System ID:

This item is only available as the following downloads:

Full Text


A Phylogenetic Approach for the Study of Va riation and Determination of Population Affiliation of Indigent Human Skeletal Remains by Hattie Bea Wetherington A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology College of Arts and Sciences University of South Florida Major Professor: James R. Garey, Ph.D. Lorena Madrigal, Ph.D. Bruce J. Cochrane, Ph.D. Date of Approval: March 28, 2005 Keywords: mitochondrial dna, race, craniometr ics, molecular forensics, discriminant analysis Copyright 2005 Hattie Bea Wetherington


i TABLE OF CONTENTS List of Tables iii iv List of Figures Abstract v Chapter 1: Introduction 1 Human Variation in the Medicolegal Pr ofession 3 Role of Mitochondrial DNA 4 Mitochondrial DNA D-loop use for Forensic Assessment of Race 7 Role of Phylogenetics 8 Chapter 2: Materials and Methods 10 Sample Collection 10 IRB Approval 10 Sample Extraction 11 Decalcification 11 DNA Extraction 11 Molecular Analysis 12 Amplification 12 Purification 13 Sequencing 14 Database Sequences 15 Quality Control 15 Morphological Assessments 16 Sample Source 16 Cranial Measurements 16 FORDISC 2.0 Analysis 18 Statistical Analysis 18 Selection of Dataset 19 Correlations 19 Discriminant Analysis : Database 19 Discriminant Analysis: Unknowns 20 Chapter 3: Results and Discussion 21 Molecular Data 21 DNA Extraction from Bone 21


ii Sequences 21 Phylogenetic Analysis 23 Case Identification 25 Morphological Data 27 FORDISC 2.0 27 Case # 82-0013 29 Case # 82-0607 29 Case # 82-1005 30 Case # 88-0220 31 Case # 96-0486 32 Case # 98-1909 34 SPSS Analysis 34 Molecular vs. Morphological Assessments 39 Conclusions 43 References 47 Appendices Appendix A: HVRBASE Sequence Designations 51 Appendix B: Quality Control Alignments 52 Appendix C: Bootstrapped Neighbor-Joining Tree 1 55 Appendix D: Bootstrapped Neighbor-Joining Tree 2 64 Appendix E: FORDISC 2.0 Output s for Case 82-0607 72 Appendix F: FORDSIC 2.0 Output s for Case 82-1005 74 Appendix G: FORDISC 2.0 Outputs for Case 88-0220 75 Appendix H: FORDISC 2.0 Outputs for Case 96-0486 76


iii LIST OF TABLES Table 1 Beckman Protocol for Cycle Sequencing Reactions 14 Table 2 Cranial Measurements: Abbreviati ons and Definitions 17 Table 3 Matrix of Craniometric Data fr om Six Unknown Cases 28 Table 4 SPSS Eigenvalues for Forensic Data Bank 36 Table 5 Summary Table: Mole cular vs. Morphological Assessments 42


iv LIST OF FIGURES Figure 1. Mitochondrial DNA Showing the D-loop region 6 Figure 2. Anderson Sequence of HVR I Showing Primer Region 13 Figure 3. 0.9% Agarose Gel Visualizing PCR Amplified Bone Product 21 Figure 4. Neighbor Joining Tree showi ng lab personal 22 Figure 5. Partial neighbor joini ng tree using Tajima and Nei Distance a) European Clade b) African Clade 24 Figure 6 FORDISC 2.0 Multigroup Classi fication for 82-0013 29 Figure 7 FORDISC 2.0 Multigroup Classi fication for 82-0607 30 Figure 8 FORDISC 2.0 Multigroup Classi fication for 82-1005 31 Figure 9 FORDISC 2.0 Multigroup Classi fication for 88-0220 32 Figure 10 FORDISC 2.0 Multigroup Classification and Two Group Discriminant Function analysis for 96-0486 33 Figure 11 FORDISC 2.0 Multigroup Classi fication for 98-1909 34 Figure 12 SPSS Canonical Plot showing group distributions for Forensic Data Bank 35 Figure 13 SPSS Standardized Canonical Discriminant Function Coefficients 36 Figure 14 SPSS Canonical plot showi ng unknown groups 38 Figure 15 Classification Results for ungrouped and grouped cases 38, 39


v A Phylogenetic Approach for the Study of Va riation and Determination of Population Affiliation of Indigent Human Skeletal Remains Hattie Bea Wetherington ABSTRACT Mitochondrial DNA (mtDNA) has played a major role in human population studies over the past decade due to its maternal inheritance and negligible recombination (Macaulay, 1999). The mtDNA control region has been the focus of these studies due to the highly polymorphic nature of this non-co ding region. Forensic scientists also use mtDNA to help determine the identity of missing individuals when nuclear DNA is not present. However, when skeletal remains are unclaimed, identification becomes near impossible. Therefore, mtDNA can play a valu able role in identification in terms of population affiliation, especia lly in conjunction with morphological analysis. The goals of this research were two-fold: 1) to determine population affiliation of unknown skeletal samples using phylogenetics a nd 2) to find a method of extraction that leaves a majority of the remains intact. This research depended on the donation of samples from sixteen skeletal remains from the Hillsborough County Medical Examiners office. Mitochondrial DNA from ten of th ese cases were extracted, amplified, and sequenced in order to determine populati on affiliation via phylogenetic analysis of hypervariable region I (HVR I). These sequences were aligned and compared to that of sequences in a pre-existing mtDNA control re gion database (Handt, 1998). The crania of


vi the skeletal remains were measured and subsequently analyzed by the forensic anthropology program FORDISC 2.0 to morphologically determine population affiliation. A secondary morphological analysis included input of the measurements into SPSS, a statistical program package, as a sepa rate discriminant function assessment. This analysis was dependent on a database of cr aniometrics from known individuals (Jantz, and Moore-Jansen, 2000). The results were co mpared to those of FORDISC 2.0 as well as that of the phylogenetic tree constructed using molecular techniques. This study will determine whether phylogenetic analysis is a legitimate way to determine population affiliation of unknown individuals, thereby benefiting future forensic studies.


1 CHAPTER 1 Introduction During the last century, forensic science has come to the forefront of the legal field as a legitimate means for determining gu ilt. This has been due to the continuous advances in science over the last two decades, and the ability to incorporate different subdisciplines of the field. Forensic anthropology remains one of the more important and intriguing disciplines of this medico-legal field due in pa rt to the vast work done by William Bass and other influential figures over the past forty years. Forensic anthropology is the study of human remains to determine the circumstances of death (R hine, 1998). Thomas Dwight t ook careful osteological measurements to accurately determine ag e and sex of skeletons (Rhine, 1998). Few advances were made during th is time until the advent of World War II, which caused an increase in awareness in the field of fore nsic anthropology. Soldiers remains provided the means for a large anthropological study fo r the purpose of iden tification and proper burial by family members. Today, forensic anthropology is a growing field, widely recognized as an important tool in forensics. Molecular work in the field of forensic s has a more recent history. Advances in biotechnology in the past twenty years have revolutionized the field. The invention of polymerase chain reaction in 1983 by Kary Mu llis allowed large am ounts of targeted copies of DNA to be used in molecula r analysis, decreasing time and increasing


2 reliability (Rabinow, 1996). In 1985, English geneticist, Alec Jeffreys, described the process of DNA fingerprinting or DNA typing (profiling) (Bu tler, 2001). This was based on his discovery of tandem repeats present in DNA sequences. These variable number of tandem repeats (VNTRs) differed from indi vidual to individual and allowed for the ability to perform identity tests through restriction enzyme digests specific for each fragment (RFLP) (Butler, 2001). This tech nique provided the highe st discriminatory power for identification at the time. Pr eviously, ABO blood groups were used for discriminatory analysis of individuals. This analysis was quick and useful for exclusion of individuals, but limited as an informative and inclusionary tool. Currently, the highest power of discrimination and rapid analysis in molecular forensics is combined in multiplex short tandem repeat analysis (STR). Because the tandem repeat is short, multiple STRs can be analyzed in the same DNA test, or multiplexed (Butler, 2001). Automation of these techniques has also prov ided an important benefit as the use for molecular testing increases. Highly publicized cases have increased the demand for DNA analysis in forensics. Molecular techniques allow for accurate identification of a decedent with only a sample of DNA. Forensic anthropology is dependent on morphological assessment of human remains, which can be biased toward issues such as race. Without complete human remains, morphological assessment beco mes difficult, if not impossible. The combination of morphological and molecu lar assessments allows more certain identification of human skeletal remains.


3 Human Variation in the Me dicolegal Profession A large degree of emphasis is placed on cate gorization of indivi duals in terms of racial groups in law enforcement. While this type of categorization seems necessary in a criminological context, it is not necessarily appropriate. There is a consensus among anthropologists that racial categorization is a human construct and that races do not exist. The concept of race has been rooted in typology (Brace 1982; Gould 1983). This idea was dominant for the latter half of the nineteenth and early twentieth century, and failed to view the human species for its diversity instead grouping indi viduals as a set of stereotypes (Johnston and Little 2000). Johann Blumenbach is responsible for the early racial categorization of huma ns into the following five cl asses: Caucasoid, Mongoloid, American, Ethiopian, and Malayan (M olnar, 2002). He depended on physical measurements of the human cranium to classify individuals into each of these groups (anthropometrics). Ceseare Lombroso, an It alian army physician who became interested in the motivation behind crime, proposed that criminal tendencies had to do with an individuals race. He based this on obs ervations of commonalities between prison inmates and concluded that criminals manifest primitive animal like characteristics based on the facial structure and body type of th e individual (Smith, 1968) This remains a common view in law enforcement, and is incorrectly used to explain why certain individuals commit crime. Human variation exists, but not in clearly delineable states. Categorizing human variation into clearly define d groups or races is assuming variation is not a complex interaction between genes an d environment. The distribu tion of human variation is gradual across the geographic landscape, and is primarily distributed as clines or


4 gradients (Brace, 2000; Cavalli-Sforza, 1976). This is also true for the variation of mitochondrial DNA across continents. Human population studies using molecular data have shown that the variation of populations is greater within population groups rather than between population groups (Lewonton, 1974; Brown and Armelagos, 2001; Romualdi, 2002). Racial categorization is a reality in law enforcement. A forensic scientist can dismiss the issue of race, e xplaining it as non-existent co nstruct, but law enforcement officials often prefer to work with colleagues who share th eir view on race. Therefore forensic scientists need to ta ke into account the need to id entify race in the medico-legal profession. Most people visualize the variation in human populations in terms of phenotypic appearance, and differe nces in cranial features alone allow more than 75% separation between individuals of different population gr oups (Gill, 1990). Forensic scientists need to be pragmatic in accomm odating the needs of law enforcement officials in determining the identity of human remains. Role of Mitochondrial DNA Large amounts of DNA are not usually pres ent in skeletal re mains because of degradation due to environm ental conditions. Mitochondrial DNA molecules are present in hundreds to thousands of copies per cell, compared to the nuclear compliment of two copies per cell (Fisher, 2005; Melton, 2002) When DNA is extracted from bone, the majority of the extracted DNA is mitochondrial. This allows for molecular analysis of degraded samples when nuclear DNA cannot be recovered. With a reference sample of mtDNA from the mother or other siblings, iden tification of a decedent can be made. This type of analysis has specifically been used to identify the remains of missing victims,


5 such as those from the World Trade Center attacks of 9/11. (Tang, 2003). The use of mtDNA in this forensic application is simpler than using nuclear DNA because recombination is negligible in mt DNA (Macaulay, 1999). Mitochondrial DNA is maternally inherited because the male game te does not contribute mitochondria to the zygote. Assuming the clonal nature of mtDNA, it cannot uniquely iden tify individuals as well as nuclear DNA (Melton, 2002). Mitochondrial DNA contains two variable segments in which selection plays little role in maintaining. These polymorphic regi ons are useful for human identification. Genetic drift affects these non-coding regions and provides the basis for studying variation in human populations (Kimura, 1979; Tajima, 1989). Two individuals that are related by maternal descent should have iden tical mtDNA sequences. The more similarity between the two sequences in this region th e more closely relate d the individuals. By comparing a large amount of sequences, popu lation structure should begin to emerge. Different populations would remain isolated by language or by geographic boundaries, so mutations would remain similar for in dividuals in closer geographical/cultural locations. Gene flow does occur between normally isolated population groups, making distinct identification of an individuals popula tion affiliation more difficult. It is feasible that mtDNA analysis may be able to distingu ish between populations that are genetically distinct from one another due to geographic isolation. Hypervariable region I (HVRI) and hype rvariable region II (HVRII) are the two non-coding regions of mitochondrial DNA that have been the focus of human population studies. The control region or displacement l oop (d-loop) is responsible for the regulation of mtDNA replication (Ande rson, 1981). The rest of the 16,569 bases of the human


mtDNA genome consists of coding region a nd is responsible for the production of various proteins and structur al RNAs involved in the process of energy production in the cell (Isenberg and Moore, 1999). Unlike th e linear nuclear chromosome, mtDNA is circular and highly economical in organization. There ar e very few n on-coding bases between genes except for the D-loop region where little coding function has been determined (Anderson, 1981). This makes the D-loop highly attractive for the study of variation in populations. The numbering of nucleotides was assigned according to Anderson, et al. (1981), with position number one falling in the middle of the non-coding D-loop, and HVR I to the left of position one (16,000-16,430), and HVR II to the right of position one (40-440) (Butler and Levin, 1998) (figure 1). Figure 1. Mitochondrial DNA showing the d-loop region. Figure adapted from Hartl and Jones, 1998. 6


7 Mitochondrial DNA d-loop database use fo r forensic assessment of race Due to the growing use of mitochondrial DNA analysis in forensics and population studies, the need to incorporate sequences into a retrievable format for colleagues to access became a necessity. Da tabases containing t housands of mtDNA hypervariable region sequences have been compiled for use in the scientific community. These sequences have traditionally been used in human population st udies to reconstruct migration patterns, analyze population varia tion, and establish possible evolutionary histories. The forensic science community has realized the potentia l use of mitochondrial DNA not just for individual identification, bu t also for possible popul ation affiliation. A recent paper by Wittig et al. discussed the use of the D-Loop-BASE database in order to analyze 850 sequences from Ge rmany, Switzerland, and Austria. German sequences were subdivided into northern, eas tern, western, southe rn, and south-western populations (2000). The analysis included pair-wise comparison of sequences to establish mean sequence differences in each populat ion/subpopulation group, but did not include phylogenetic methods. The purpose of this anal ysis was an attempt to establish the potential of mitochondrial DNA for use in popul ation differentiation of individuals. The authors found that the mean sequence diffe rence (MSD) between the three populations was high, indicating a good potential for usi ng mtDNA as a discrimina ting tool. In order to determine whether the high MSD was caused by diverse sub-populations and was measurable, the MSD and genetic distances were calculated for all population pairs. No significant differences could be detected between the three populations. The authors concluded that this was due to the small nu mber of sequences available for analysis.


8 Mitochondrial DNA databases provide a useful tool for studying human population structure. The most comprehens ive database, HvrBase, was compiled by Handt et al. (1998) and currently cont ains 4079 HVR I and 969 HVR II sequences representing a number of population groups. Other databases have been compiled to study mtDNA diversity between individuals of different populations, these include the mtDNA Concordance (Miller and Dawson, 1998) and MitoMap (Brandon, 2004). More accurate determinations of population affiliation may result with the addition of more sequences into these databases by the scientific community. Role of Phylogenetics Phylogenetic analysis would be an appr opriate means for determining possible population affiliation from an unknown case samp le. Phylogenetic analysis attempts to determine relationships between different gr oups or taxa based on their evolutionary history. Since mitochondrial DNA recombinat ion in normal humans is negligible, phylogenetic analysis would be justified b ecause coalescence would be less likely to occur (Macaulay, 1999). The work described here incorporates phylogenetics to determine the population affiliation of a set of ten unidentified sk eletal remains from the Hillsborough County Medical Examiners Office. The term populati on affiliation used subsequently in this paper refers to the self-identified geographi c ancestry of an individual of a particular human population as reported in the sequence databases for human mtDNA variants. For example, the terms European (Caucasian), Af rican, and Asian refer to individuals with ancestry in populations from these geogra phic locations. When discussing population affiliation in medico-legal context, individuals are referred to by their phenotypic


9 appearance of white or black. The term Hispanic is also used to describe the ancestry of an individual in forensic anthropo logical studies. It is used in appropriately in this context, as the term Hispanic refers to the ethnic ity or cultural background of individuals from Latin America, not necessarily the ancestry of the individual, which could be European or indigenous. It is important to keep the contex t of these terms in mind when analyzing the results from each analysis. Molecular analysis was carried out using mitochondrial DNA HVR I sequences from the unidentified individuals These sequences were analy zed with a subs et of known HVR I sequences and a phylogenetic tree constructed. A morphological analysis was also performed. Craniometric da ta from known individuals were compared to that of the unknown cases using discriminant function analysis. This analysis was then compared to the results of a craniometric analysis by FORDISC 2.0, a forensic anthropology computer program (Ousley and Jantz, 1996). The molecular and morphological analyses were then compared to one another in order to de termine if the two were in agreement. Another goal of this re search was to develop a less destructive means of extracting DNA from bone samples. Conven tional methods of DNA extraction result in the destruction of large portions of th e bone (Hochmiester, 1991; Prado, 1997). This technique utilizes a less invasi ve drill method leaving the skeletal remains intact for further morphological study. The project also serves to de monstrate that molecular techniques combined with morphological pr actices will furthe r solidify forensic identification of unknown individuals.


10 CHAPTER 2 Materials and Methods Sample Collection Fifteen skeletons were provided by th e Hillsborough County Medical Examiners Office for use in this research. The human skel etal remains were of indigent decedents in which identification has not been positively assigned. Remains were assigned case numbers according to the date the remains were discovered. Eight of these cases have been examined by forensic anthropologist, Dr. Curtis Weinker at the University of South Florida. Population affiliation has been tentatively assigned using FORDISC 2.0. IRB Approval: Permission to collect biological samples from lab personnel and skeletal remains was obtained through the Universi ty of South Floridas Institutional Review Board, IRB number 100150b. No identifying information was kept in association with the samples from lab personnel. Minimum risk consent form s were signed by participants and kept in locked filing drawer in the laboratory. Sample Extraction: Out of the fifteen cases, bone powder was ex tracted from femurs and pelvises of ten cases. The method included using a 1/8 drill bit to core bone material from the remains in order to obviate destruction of skeletal mate rial for further morphological analysis. The extraction of bone material took place inside the autopsy bay of the medical


11 examiners office. The drill bits were cleaned with a DNA inhibitor, DNA Away (Molecular BioProducts, San Diego) between each extraction of bone material. The typically brown clumps and bone powder (a pprox 100 mg of material ) was placed in a 1.5 ml Eppendorf tube and decalcified using Hochmeisters protocol for extraction of bone samples (1991). Decalcification: Five hundred microliters of EDTA at pH 8 was added to the 1.5 ml tubes and agitated on a rotor at room temperature for 24 hours. The supernatant was discarded and an additional 500 l of EDTA was added. This process was repeated until the calcium ions were extracted from the bone into solu tion; a process which was monitored by the addition of ammonium sulfate which precipitates in the presence of calcium ions. This process typically took three to five days. The accumulated ions were discarded by washing the pellet in nanopure water and centrifuging to discard the supernatant. This washing was repeated two additional times. DNA Extraction: DNA was then extracted using a modi fication of Hempsteads standard DNA extraction protocol. 300 l of pre-warmed extraction buffer (3.5% SDS, 1M Tris pH8, 100mM EDTA) and 50 l of proteinase K was added to the mixture in the 1.5 ml Eppendorf tube and incubated for two hours at 60 o C with agitation and then for an additional ten hours without agitation. An equa l volume of phenol was then added to the mixture (350 l) and gently mixed for ten minutes on a rotor at room temperature. The solution was then centrifuged using Thermo El ectron Corporation microcentrifuge at high


12 speed for five minutes. The aqueous layer was th en transferred to another 1.5 ml tube and the phenol extraction was repeated one additional time. Phenol/chloroform and chloroform extractions followed using the same technique. One-tenth volume of sodium acetate (NaAc) and two volumes of 100% ethanol were added to the aque ous layer and stored at -20 o C overnight or at -80 o C for one hour. The sample was centrifuged for fifteen minutes, and 500 l of 70% ethanol was used to wash the DNA pellet. The sample was centrifuged for ten minutes and the supernatant discarded. The pellet was dried using the speed vacuum (Savant) and resuspended in 20 l of nanopure water. The resuspended DNA was further purif ied using UltraClean Soil DNA Isolation Kit (MoBio Laboratories, Carlsbad). The k it uses beaded soluti on in conjunction with detergents, heat, and mechanic al force to lyse cellular co mponents in the solution. The DNA is then bound on a silica spin filter afte r centrifugation and washed in nanopure water to recover the DNA (MoBio Laboratories, Carlsbad). This particular kit was used in order to remove any soil or humic resi dues that may have been left after DNA extraction of the bone samples. Molecular Analysis Amplification: A 572 bp segment of the HVR I of th e mtDNA control region was amplified using primers H16498 (5-CCTGAA GTAGGAACCAGATG-3) and L15926 (5TCAAAGCTTACACCAGTCTTGTAAA CC-3) (figure 2). The 100 l reactions were set at a denaturing temperature of 94 o C for 15 seconds, annealing temperature of 50 o C for 30 seconds, and an extension temperature of 72 o C for 30 seconds. This cycle was


repeated 39 times with a 72 o C hold for ten minutes and a 4 o C hold at the end of the 40th cycle. The DNA was then visualized by gel electrophoresis by loadi ng the sample onto a 0.9% agarose gel alongside a DNA 100 base pair ladder. L 1 59 2 6 AAACTAAT AC ACCAGTCTTG TAAACC GGAG ATGAAAACCT TTTTCCAAGG ACAAATCAGA 15960 GAAAAAGTCT TTAACTCCAC CATTAGCACC CAAAGCTAAG ATTCTAATTT AAACTATTCT 16020 CTGTTCTTTC ATGGGGAAGC AGATTTGGGT ACCACCCAAG TATTGACTCA CCCATCAACA 16080 ACCGCTATGT ATTTCGTACA TTACTGCCAG CCACCATGAA TATTGTACGG TACCATAAAT 16140 ACTTGACCAC CTGTAGTACA TAAAAACCCA ATCCACATCA AAACCCCCTC CCCATGCTTA 16200 CAAGCAAGTA CAGCAATCAA CCCTCAACTA TCACACATCA ACTGCAACTC CAAAGCCACC 16260 CCTCACCCAC TAGGATACCA ACAAACCTAC CCACCCTTAA CAGTACATAG TACATAAAGC 16320 CATTTACCGT ACATAGCACA TTACAGTCAA ATCCCTTCTC GTCCCCATGG ATGACCCCCC 16380 TCAGATAGGG GTCCCTTGAC CACCATCCTC CGTGAAATCA ATATCCCGCA CAAGAGTGCT 16440 ACTCTCCTCG CTCCGGGCCC ATAACACTTG GGGGTAGCTA AAGTGAACTG TATCCGA CAT 16500 CTGGTTCCTA CTTCAGG GTC ATAAAGCCTA AATAGCCCAC ACGTTCCCCT TAAATAAGAC 16560 H 1 6 4 98 Figure 2. Section of the Anderson (1981) mitochondrial DNA d-loop used to amplify a 571 bp portion of the hypervariable I region (Genbank). Purification: The PCR amplified products were purifie d using Ultrafree-MC centrifuge cups (Millipore, Billerica). Three hundred microliters of nanopure water was added to each cup. Approximately 30 l of PCR product was added to each cup and the mixture centrifuged at room temperatur e at 2000 g for approximately five minutes. This leaves approximately 5-20 l of retentate in the filter cup. Th e fraction collected in the 1.5 ml tube was discarded and another 250 l of nanopure water was added to the filter cup. The 13


mixture was centrifuged at room temperature for approximately five minutes at 2000 g until 20 l retentate is remaining. Th e fraction in the 1.5 ml tube was again discarded and the process was repeated one additional time. The retentate was pipetted up several times over the membrane, taken up, and transferred to a new 1.5 ml tube. Another 20 l of nanopure water was added to the membrane and pipetted up over the membrane several times to remove any remaining DNA from the membrane. Five microliters of the purified product was run alongside the DNA logical ladder in order to quantify the concentration of the template. Sequencing: The concentration of product was then us ed to determine how much template to use in the sequencing reaction. Beckman sequenc ing protocols were set up according to Table 1. Reactants Full Reaction Half Reaction Quarter Reaction DNA Template 0.5-10.0 l 0.5-5.0 l 0.5-7.5 l Primer 1.6 pMol/ul 2.0 l 1.0 l 0.5 l DTCS Quick Start Master Mix 8.0 l 4.0 l 2.0 l Sterile water (adjust to total volume) X l X l X l Reaction Volume 20.0 l 10.0 l 10.0 l TABLE 1. Beckman protocol for cycle sequenci ng rections (Beckman Corp, Foster City). 14 Reactions were cycle sequenced as follows: 96 o C for 20 seconds, 50 o C for 20 seconds, and 60 o C for 4 minutes. This is repeated for 30 cycles and followed by a 4 o C hold. Once cycle sequencing was complete, ethanol precip itation of the template was carried out.


15 Five microliters of a stop solution premix (2/5 volume 3M Sodium Acetate pH 5.2; 2/5 volume 100mM EDTA pH 8.0; 1/5 volume glycogen) and 60 l of cold 95% ethanol was added to the reaction and mixed. The solution was centrifuged at 14,000 rpm at 4 o C for 15 minutes. The supernatant was pipetted off and 200 l of cold 70% ethanol added. Solution was again centrifuged at 14,000 rpm at 4 o C for two minutes. The supernatant was discarded and another wash of 200 l of 70% ethanol took place. The solution was centrifuged for 2 minutes at 14,000 rpm and the s upernatant pipetted off. The pellets were allowed to dry at room temperature for 10-15 minutes or vacuum dried on low heat setting. The pellet was then re-suspended in 40 l of sample loading solution (SLS) provided in the Beckman sequencing kit and reading for loading into the sequencer. Database Sequences: Representative sequences were taken fr om 308 individuals of five population groups: European descent, Af rican descent, Asian descen t, Hispanic descent, and Amerindian descent (Appendix A). HVR 1 sequ ences were obtained from the HvrBase, which presents data in an aligned, search able format according to population and/or continent affiliation (Handt, 1998). The unknown sequences obtained from the previous molecular techniques were added to the databa se in a text file and aligned using the alignment program, Clustal. The alignment file was input into the phylogeny program, MEGA, which allowed the individual sequences to be assigned to one of the five general population groups. Quality Control Several precautions were taken to ensure quality control of unknown sequences. Buccal swabs were taken from members of the laboratory in which the bone samples


16 were analyzed. These control sequences were then compared to that of the bone samples ruling out possible contamination from laboratory personnel. DNA extraction and amplification took place in a fume hood th at was thoroughly wiped with DNA Away. The fume hood was also sterilized by exposi ng the area to UV radiation overnight prior to use. Morphological Assessments Sample Source: Morphological analysis consis ted of measuring the crania of the case samples that were used in the molecular analysis. Out of the ten samples, six had skulls intact for craniometric study. The cases numbere d the following: 98-1909, 82-1005, 82-0607, 880220, 96-0486, and 82-0013. The measurements from the unidentified case samples were added to case measurements reported in the Forensic Data Bank located on the Inter-University Consortium of Political and Social Research (ICPSR) website. This database contains decedent information on thousands of id entified skeletal remains, including craniometrics, and was compiled by Dr. Richard Jantz and colleagues at the University of Tennessee (Jantz, 2000). Cranial Measurements: Twenty-one measurements were taken for input into FORDISC 2.0. Larger measurements were taken with a pair of spreading calipers, while smaller measurements were obtained using digital sliding calipers. Each measurement was repeated three times and the mean reported. The measurements are described in Table 2.


Cranial Abbreviation Cranial Definition GOL Glabello-Occipital Length BNL Basion-Nasion Length BBH Basion-Bregma Height XCB Maximum Cranial Breadth WFB Minimum Frontal Breadth ZYB Zygomatic Breadth AUB Biauricular Breadth BPL Basion-Prosthion Length NLH Nasal Height NLB Nasal Width MAB Palate Breadth MAL Palate Length OBH Orbital Height OBB Orbital Breadth DKB Interorbital Breadth EKB Biorbital Breadth FRC Frontal Chord PAC Parietal Chord OCC Occipital Chord FOL Foramen Magnum Length UFHT Upper Facial Height Table 2. Cranial measurements: Abbreviations and Definitions. Used for morphological assessment of case samples. 17


18 These measurements were chosen based on co mpleteness of data in the Forensic Data Bank. If data was missing for certain measur ements in the database, they were not included in the morphological analysis. FORDISC 2.0 Analysis: The data collected from the six skulls were input into the forensic anthropology computer program called FORDISC 2.0 Pers onal Computer Forensic Discriminant Functions (Ousley and Jantz, 1996 ) in order to assess race based on craniometric data. The program analyzed the data by comparing the unknown measurements with known samples already based in the software repres enting African Americans, American whites, Chinese, Vietnamese, Hispanic, and Native American populations. In addition, Howells groups (1973, 1989) represent archeological sa mples examined. They reflect older and more geographically diverse samples (Ubelaker, 2000). The software uses a ready-made discriminant function analysis based on Giles and Elliot (1962; 1963) or Jantz and Moore-Jansen (1988). Statistical Analysis: The Forensic Data Bank (Jantz and Moore-Jansen, 2000) was used in a statistical analysis in order to determ ine if individuals grouped into populations. The statistical program SPSS (SPSS Inc., Chicago) was used for discriminant analysis of the data. Two separate analyses were run using 159 random ly chosen male and 159 randomly chosen female cases representing each of the five identified racial groups. The unknown cranial measurements were then input into the above analysis to determine population affiliation.


19 Selection of dataset: The female and male cases used in the di scriminant analysis were chosen out of 1400 individual cases in the Forensic Data Bank. The cases were chosen at random after several rounds of discrimination. First, all cases that contained missing information were removed from the dataset. From the remain ing cases, approximately thirty individuals were chosen at random as representatives from each racial category totaling one hundred and fifty nine individual cases. Correlations: All twenty-one measurements were run through a correlation an alysis in SPSS in order to determine which variables were strong ly associated with one another. A bivariate correlation analysis was chosen from the Analyze option in the toolbar of SPSS. The Pearson coefficient was chosen to distinguish between variables with high proportionality to one another. One variable from each correl ated group with Pearson coefficients greater than 0.5 was chosen. The twelve independent variables input into the discriminant analysis were the following: GOL, BBH, X CB, AUB, NLB, MAB, OBH, OBB, DKB, FRC, FOL, UFHT. Discriminant Analysis: Database The discriminant analysis was first executed using the measurements from the one hundred and fifty nine chosen cases from th e FDB. The discriminant function option is located in the classify section of the SPSS toolbar. The grouping variable for the analysis was population groups 1-5 (African, Caucasian, Hispanic, Asian, Amerindian). The twelve independent variables described above were input in the analysis as the independent variables.


20 Discriminant Analysis: Unknowns The unknown cases were input into the know n dataset and another discriminant analysis was run. The same twelve independent variables were used as well as the same grouping variables. Unknowns cases were not designated a grouping variable, therefore SPSS classified them as ungrouped cases in the output file.


CHAPTER 3 Results and Discussion Molecular Data DNA Extraction from Bone: Mitochondrial DNA was successfully extracted and amplified from eight of ten cases using a drilling method of extracti on. Out of these eight, seven cases were successfully sequenced. The success rate appeared to be dependent on many factors, including age of remains, degree of e nvironmental degradation, and degree of contamination. Sequences: 21 ands r amplified DNA from PCR appeared DNA sequences of the HVR I region were obtained by direct sequencing of PCR products. Bevealing to contain ample DNA for sequencing (figure 3). Figure 3. Gel electrophoresis showing HVR I amplified bone samples against a 100bp marker.


22 equences were compared to that of la boratory personnel in order to determine whethe S r samples were contaminated by indivi duals working in close proximity to the samples. Appendix B shows aligned sequences of laboratory personnel with those of each unknown sample. Figure 4 shows a neighbor-j oining tree containing unknown sequences and lab personnel sequences. igure 4. Neighbor-Joining tree showing laboratory personnel in relation to unknown sequences. F


23 hylogenetic Analysis: the assumption that population affiliation can be determined phylog P In order to test enetically, a neighbor-joining (NJ) tree containing 308 known sequences was constructed using the Tajima and Nei distance parameter model, which takes into account the unique base composition of mitochond rial DNA (Tajima and Nei, 1984). This correction method assumes that substitutions o ccur with equal probability at any site along the sequence, and that substitutions o ccur at equal rates (Nei, 2000). The NJ tree showed a detailed topology of the shallow nodes, but very little reso lution in the deeper nodes. This tree revealed five clades of four groups (see Appendix C). Two Latin American clades, one African clade, and one European clade were distinguishable. The out-group for the analysis consisted of three Neanderthal sequences and one Pan troglodytes (chimpanzee) sequence. Bootstrap values were low for all constructed trees (see Appendix C for node values ). The first Latin American clade contained 26.6% of the Latin American sequences, which consis ted mostly of Columbian and Chilean individuals. The larger Latin American clade (II) contained 64.2% of the Latin American sequences included in the analysis. This clade was more varied in sequences representing different Latin American populations. It included mostly Mexican and Brazilian sequences, as well as the remaining Chilean and Columbian sequences It also contained the Amerindian sequences from Costa Ri ca. The European clade contained 88.2% European sequences, and the African clade contained 81.7% of the African sequences. The Latin American Group II grouped basa lly in this particular analysis.


a. 24 24


25 b. Figure 5. Partial Neighbor-Joining tree constructed using the Tajima-Nei parameter model showing the European (a) and African (b) clades. Unidentified case numbers are in red, known reference samples in blue.


26 26 in subsequent analyses, but all revealed ighbor-joining tree. Upon the removal of Pan above analysis changed slightly. erged: two Latin American clades, one European ade grouped basally in th e analysis. Also, the was higher in all but the African American can sequences as opposed to the 81.7% of e contained 92.6% European sequences. Latin nces; and the Latin American clade II In this analysis, Latin American clade I contained whereas the remaining Latin American unidentified cases were grouped within the same clades. 98-3587 fell in the same group as clade (figure 5A). The Ande rson reference sequences was included in the analysis and also grouped within the European clade. Cases 88-0220 and for PCR analysis (HeLa mtDNA) fell into By looking at the molecular tree analysis, it is reasonable to assume that human variation in the m were able to separate five clades of th many individuals of different populations. Despite only three population groups being Other relevant distance models were used the same topology in each constructed ne troglodytes as an out-group, the topology for the (Appendix D). The same five groups em clade, and one African clade. The African cl percent of individuals classified correctly clade. The African clade contained 74.7% Afri the previous analysis. The European clad American I clade contained 30.2% Latin seque contained 60.5% Latin sequences mostly Brazilian and Mexican sequences, sequences were contained in the larger Latin Clade II. Case Identification: In both tree analyses, the Cases 82-0013, 82-1005, CC94-0011, 96-0468, and individuals in the European 98-1909, as well as the positive control used the African clade (figure 5B). itochondrial DNA can be distinguished phylogenetically. The NJ trees ree population groups using 308 taxa representing


27 represe ulation studies around the world, by their self-ident ified population affiliation. Morphology and enetic ntification in law enforcement, and to say with some statistic nted, the unknown sequences did cluster within gr oups of taxa with similar population background. Case 98-1909, for example, was later positively identified as an African-American male. This case did group w ith individuals of Af rican descent in both molecular tree analyses (figure 5B). A mtDN A sequence from a HeLa cell line was also included. This cell line comes from the cervi cal cancer cells of an African American individual named Henrietta Lacks. As seen in figure 5b, the HeLa sequences also grouped in the African clade. The Anderson reference se quence was extracted mtDNA from human placenta of obtained from normal or caesarean section deliveries (Drouin, 1980). The maternal descent of these sample s is not stated. The reference sequence grouped in the European clade, more speci fically near taxa of German descent. Molecular results using mtDNA reflect only maternal lineages. Individual sequences in Handt database were compiled from pop and are referenced gs can be different for the same individual, depending on what aspect of the DNA is being investigated. In the case of mtDNA, if an individual appeared black phenotypically, but had a maternal ancestor th at was of European ancestry, the mtDNA would presumably be European, despite self -identification as African. This compounds the issue of population identification in the realm of forensics. Typically, phenotype is the determining factor of ide al significance that an individual had European ancestry does not help matters if the individual differs phenotypically.


28 usley and Jantz, 1996). Typicality probabilities represent how likely the unknown belongs to any particular group. Typicality probabilities vary depending on the combina tion of groups involved in the analysis. Due to program constraints, this analysis used only 308 sequences from the Handt database. The accuracy of population affiliati on is greater than bootstrap analysis suggests. Perhaps a new way of determining statistical significance would be useful. Morphological Data Two morphological assessments were carr ied out. One involved the use of SPSS, a statistical program package in which the morphological measurements were analyzed using discriminant function analysis. The other involved the use of FORDISC 2.0, a forensic anthropology program that allows ready-made discriminant function analyses traditionally used in forensic anthropology. FORDISC 2.0: Craniometric data from each unknown case was input into FORDISC 2.0 for racial assessment. Table 3 shows a matrix of obtained skull measurements from the six unknown cases. Output for each unknown case in FORDISC includes the percentage of correctly classified individuals into the appropriate group, classi fication of the unknown individual in a multi-group analysis, posteri or and typicality probabilities of the unknown into each group, comparison of craniometric group means for each variable in the analysis with the unknown variables, canonical plot for multi-group analysis, and histogram plots for two group analysis. Posterio r and typicality probabilities are used in reporting statistical significance of the results in FORDISC. Po sterior probabilities reflect the probability of group membership under the assumption that th e unknown belongs to one of the groups in the function (O


29 measurements). Highlighted measurements are abnormal values for the appropriate distance. Table 3. Matrix of craniometric data from six unknown cases (each representing the mean of three separate 82-0013 82-0607 82-1005 88-0220 96-0486 98-1909 GOL 213 mm 214 mm 172 mm 177 mm 201 mm 188 mm BBH 134 mm 171 mm 132 mm 144 mm 135 mm 135 mm XCB 130 mm 145 mm 135 mm 147 mm 133 mm 137 mm BNL 96 mm 139 mm 96 mm 106 mm 99 mm 102 mm WFB 92 mm 96 mm 96 mm 105 mm 96 mm 99 mm ZYB 117 mm 129 mm 120 mm 129 mm 125 mm 134 mm AUB 115 mm 120 mm 114 mm 129 mm 117 mm 119 mm BPL 82 mm 98 mm 90 mm 106 mm 96 mm 110 mm NLH 50 mm 53 mm 52 mm 48 mm 50 mm 51 mm N LB 19 mm 26 mm 24 mm 27 mm 23 mm 28 mm MAB 51 mm 57 mm 52 mm 61 mm 55 mm 59 mm MAL 39 mm 48 mm 40 mm 50 mm 47 mm 54 mm OBH 32 mm 35 mm 36 mm 32 mm 34 mm 36 mm OBB 34 mm 40 mm 39 mm 37 mm 36 mm 39 mm DKB 22 mm 26 mm 20 mm 27 mm 24 mm 26 mm EKB 89 mm 95 mm 94 mm 98 mm 99 mm 108 mm FRC 105 mm 115 mm 115 mm 117 mm 111 mm 116 mm PAC 112 mm 109 mm 115 mm 111 mm 117 mm 123 mm OCC 99 mm 97 mm 98 mm 92 mm 96 mm 94 mm FOL 36 mm 35 mm 36 mm 33 mm 37 mm 39 mm UFHT 50 mm 66 mm 64 mm 70 mm 62 mm 75 mm


30 ase #82-0013: Meats fr 82-0e com o all t in DISare g Hgroum thsis, FORDISC classified the unknown as white fema le based on a posterior probability of 0.799 (figure 6). The typicality pr obabilities were all zero percent indicating the unknown reme atyp ny ofwn ps. rouation3 C suremen om skull 013 wer p ared t populat ions of both sexes presen the FOR C softw exclud in owells ps. Fro is analy measu nts were ical of a the kno opulation Multig p Classific of 82-001 ----------------------------------------------------------Glassifitanceiliti roup C ed Dis Probab es into from Posterior Typicality ------------------------------------------------------WM 184.7 .201 .000 WF ** WF *9 .000 181. .799 BM 8 .000 209. .000 BF 4 .000 205. .000 AM 228.0 .000 .000 AF .000 207.1 .000 JM 207.0 .000 .000 JF .000 205.8 .000 HM 202.9 .000 .000 12.8 .000 CHM 2 .000 VM 4 .000 235. .000 ------------------------------------------------------is closest to WF s Figure probab 6. F 2.0 o owing Mahalanobis distance, po steribility, icality ility fonown c to eaction gro Case number 82-0607 was initially compared to White males, White females, males, Black females, and Hispanic ales. FORDISC classified the unknown as a male based on the posterior probability being the highesl five (see dix Eicalitybilitiezero for all groups. Because thlity robabilities were low, another analysis wa s run including all popu lation groups for all ach group ssignment (see Appendix E) decreased with the increase in groups included in the ORDISC r the unk utput s h ompared or proba and typ h popula up. Case #82-0607: Black m white t for al groups Appen ). Typ proba s were e typica p s exes, again excluding Howells groups. The percent correct classification of e a


31 .6% to 58.9%. The posterior a nd typicality probabil ities also varied. FORDI analysis from 77 SC classified the unknown individual as Vietnamese male when compared to eleven population groups in this analysis. Posterior probability for the classification of the unknown as a Vietnamese male was 0.263. The typicality probability was 0.461the highest out of the eleven typi cality probabilities (figure 7). Multigroup Classification of 82-0607 --------------------------------------------------------------------------------------------------------------------Group Classif Probabilities ied Distance r Typicality into from Posterio ------------------------------------------------------------------------------------------------------------------WM .415 14.5 .192 WF .270 16.7 .062 BM .143 19.6 .015 BF .082 21.8 .005 AM .061 22 .9 .003 AF .183 18.6 .025 JM .384 14.9 .154 JF .412 14.5 .188 HM .234 17 .4 .044 CHM .247 17.2 .050 VM ** V .461 M ** 13.8 .263 ----------------------------------------------------------------------------------------------------------is closest to VMs Figure 7. FORDISC 2.0 output showing the unknown classification into Vietnamese males. Case # 82-1005: Discriminant function results from FORD ISC classified skull 82-1005 as a white female based on a posterior probability of 0.991 and typicality of 0.191 (figure 8). Percent of correct classification of individua ls was 81% (Appendix F). This analysis was conducted using white males, white females, black males, black females, and Hispanic males as the comparison groups.


32 Multigroup Classification of 82-1005 ----------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ----------------------------------------------------------WM 37.7 .002 .010 WF ** WF ** 25.3 .991 .191 BM 43.1 .000 .002 BF 37.2 .003 .011 HM 36.3 .004 .014 ---------------------------------------------------------is closest to WFs Figu Cas Classifon ocase 88220 waani white males, white females, black males feales, anispanales. Correct classification of cases osrior prlity he unknown measurementsling nto theHispan latio ypicality ted with all eleven groups. FORDISC again cl known as Hispanic male. Correct ro to 0.018 bability increased to 0.761 for the unknown compared to the Hispanic male po re 8. Multigroup classification showing case number 82-1005 classification as white female. e #88-0220: icatif -0s Hispc male based on a group analysis of blackmd Hic m was 80.9% and the pteobabiof t fali ic male popun was 0.729. T probabilities were zero (see Appendix G). Another analysis was conduc assified the un classification dropped to 65.2%, but typicali ty probability increased from ze and posterior pro pulation (figure 9).


33 Multigroup Classification of 88-0220 ---------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ---------------------------------------------------------WM 35.8 .084 .005 WF 38.2 .024 .002 BM 37.3 .039 .003 BF 37.6 .034 .003 A M 46.9 .000 .000 AF .001 42.5 .003 JM 41.4 .005 .001 JF 40.0 .010 .001 HM ** HM ** 31.4 .761 .018 CHM 41.3 .005 .001 VM 37.6 .033 .003 ----------------------------------------------------------is closest to HMs Figure 9. FORDISC 2.0 output showing the unknown classification as Hispanic male. Case #96-0486: Discriminant function results for cas e 96-0486 were varied depending on which groups were included in the analysis. Prel iminary analysis classified the unknown as white female in a multigroup analysis including white males, white females, black males, black females, and Hispanic males (Appendix H). Typicality probabilities for all groups were zero. Posterior probability for the lik elihood of the unknown being white female was 0.639. An eleven-grou p analysis was conducted (Appendix H). FORDISC then classified 96-0486 as white male. Posterior probability of this classification was lower than the previous classification as white female -0.418. Typicality probabilities were still zero. A third analysis us ing only two groups, white male and white female, classified the unknown as white female. Posterior probability was 0.747. Typicality probabilities were zero (figure 10) Discrimi nant function score of th e unknown in this two-group analysis was .082. The class means for wh ite male was 3.565 and for white females as .565 (Appendix F). w


34 Multigroup Classification of 96-0486 From t Into Group Total Percen Group Correct WM WF Counts ----------------------------------------------------------WM 93.9% 107 7 114 WF 88.9% 9 72 81 --------------------------------------------------------Totals: 91.8% 179 195 Two 486 Group Discriminant Function Results of 96-0 --------------------------------------------------------Group erior Typicality Classified Distance Post Into ilities from Probab ----------------------------------------------------------WM .253 .000 72.8 WF **WF** 70.6 .747 .000 ----------------------------------------------------------up classification and two group discriminant function results for a two group analysis male) of case number 96-0486. Figure 10. Multigro e (white male/white f Classification of case 96-0487 was difficult because half the m easurements for the variables in the analysis were equally close to white females as they were to white males. Twenty-one variables were used in the analysis, which is more than necessary to make a racial assessment of an unknow n individual. The maximum cranial length (GOL) for this particular analysis seemed to be larger than what is typical for any population group. FORDISC brought up a message window bringing this fact to light. This variable was not removed from this analysis, but did not weigh heavily when results of a two-group discriminant function coefficient analysis we re observed (see Appendix F). The variables that were useful in determining identification were bizygomatic breadth (ZYB), basionbregma height (BBH), and basion-nasion length (BNL) with relative weights over 10%.


35 ase #98-1909: Measurem the black male population. FORD 37 for the analysis that included white males, white females, black males, black females, and Hispanic males. Typicality pr enroup analysis was run. Posterior probability was still high -0.914 for this analysis (figure 11). C ents from skull 98-1909 were most closely related to those of ISC 2.0 gave posterior probability of 0.9 obability was 0.002. Typicality proba bility rose to 0.284 when the elev g Multigroup Classification of 98-1909 ----------------------------------------------------------Group C Pil lassifiedDistance robabities into from Posterior Typicality ----------------------------------------------------------WM 39.1 .000 .002 WF 4 2.9 .000 .000 BM ** BM ** 19.8 .914 .284 BF 25.9 .043 .076 AM 35.5 .000 .005 AF 33.6 .001 .009 JM 30.1 .005 .025 JF 31.4 .003 .018 HM 30.2 .005 .025 CHM 26.8 .028 .061 VM 35.2 .000 .006 ----------------------------------------------------------is closest to BMs Figure 11. FORDISC 2.0 output showing the unknown classification as black male. SPSS Analysis: Discriminant function analysis using SPSS gave varied results when compared to FORDISC 2.0. First, the known male referen ce samples from the Forensic Data Bank were ran in the discriminant analysis wit hout an unknown in order to determine whether discrim ination between groups w ould be possible. The female dataset was not used in the e groups available from the Forensic Data Bank. Correlations produced the following significant measurements for the discriminant SPSS analysis due to lack variable referenc


36 : GOL, BBH, XCB, AUB, NL B, MAB, OBH, OBB, DKB, FRC, FOL, and UF d, with Amerindian population ce ntroid (group 5) falling slightly farther ay than the Asian and Hispanic centroids. The percent or variance for each of the four functions in the analysi nted for 52.4% of the variance. Figure 12. Canonical Discriminant Function Plot showing group distributions of 159 known individuals function analysis HT. Groupings were visualized in the ca nonical plot from the analysis (figure 12). African American (group 2) and Caucasian (group 1) populati ons seem to be the most separated groups. Group centroids for Caucasians for function one and two are .298 and -0.569. For African Americans the group centroid is 2.245 and .741, respectively. The group centroids for Asian (group 4) and Hi spanic (group 3) popu lations are the most closely space aw s u is listed in table 4. Functio n one acco from craniometric data in the Forensic Data Bank.


Eigenvalues Function Eigenvalue % of Variance Cumulative % Canonical Correlation 1 1.387(a) 52.4 52.4 .762 2 .960(a) 36.2 88.6 .700 3 .238(a) 9.0 97.6 .438 4 .064(a) 2.4 100.0 .245 analysis. The highest discriminatory measurements used in the analysis were the following: GOL, BBH, MAB, and DKB. These were based on th e standardized canonical discriminant function coefficients for function one or tw o in figure 13. Coefficients 0.5 and higher were consi Table 4. Table showing the SPSS eigenvalues and the percent variance of each function of the discriminant der ed significant. Standardized Canonical Discrimi nant Function Coefficients Function 1 2 3 4 GOL -.215 -.564 -.146 .154 BBH -.631 .122 -.059 .118 XCB -.038 .047 .359 .376 AUB -.288 .525 .218 -.102 NLB .281 .404 -.466 .508 MAB .721 .081 .250 -.074 OBH .267 .104 .088 -.469 OBB -.043 -.022 .557 -.036 DKB .316 -.659 .425 .204 FRC .164 .224 -.297 .108 FOL -.139 -.270 -.030 -.126 UFHT .059 .106 -.332 -.333 F igure 13. Discriminant function coefficients showing which craniometric variable contributed most to the lassification analysis. In the second discriminant function analysis, the unknown data was added to the atrix in SPSS. Missing values were replaced with the average for the appropriate value. f originally grouped cases were classified c m Classification results indicat ed 71.1% o 37


38 correctly. SPSS classified five out of s ped/unknown cases as Caucasian, one case African Am visualization each un case lation to known groups. Cases 82-0013 and 960486 were oiers in fromis disc riminalysis (e 14). On this twonction plot, these cases fell into the lower left quadrant, away from all other groups in sis. They were clas sified Caucasian due to thei r location on the canonical plot. function tw o equaling -0.5. Case 88-0220 grouped closest to the Asian and ber 98-1909 was closest to the centroid for the African American population in the plot. The classific res figu show numb individuals that the discriminant functiolysiuped he apiate groupings. Group 1, which represents Caucasians had and 83.1% correct cl assification rate. Fortynine of fifty-nine individuals identified as Cau n werssifiedh e appropriate predicted group. Group 2, representirican cent ped 189 indi Group three had a m individuals classified correctly. Only 31% Hispanic cases were classified in the appropriate group membership. Group four n populat ion and had 62% correct classification rate. Group five, Ameri-Indian population, had a eve n ungrou erican, and one case as Asia n (figure 15). The canonical plot allowed of grouped e in r utl the plot th nant a figur f u the analy Cases 82-0607, 82-1005, and 95-6189 grouped most closely with the group 1 centroid, indicating similarities with the Caucasian population. Case 82-1005 grouped most cl osely to Caucasian population measurements with function one equaling -0.7 and Hispanic centroid. Canonical function for case num ation ults in re 15 the er of n ana s gro in to t propr casia e cla in t ng Af de s redict of 1 viduals grouped correctly. uch lower percentage of represented the Asia 75 % correct classification.


Figure 14. Canonical Discriminant plot showing individual ungrouped cases in relation to known individuals from each population grouping. Classification Results (A,B) A Grouping Predicted Group Membership Total 1 2 3 4 5 Original Count 1 49 1 5 3 1 59 39 2 0 18 1 0 0 19 3 3 3 6 3 4 19 4 5 0 8 31 6 50 5 0 0 1 2 9 12 Ungrouped 5 1 0 1 0 7 cases


40 a. 71.1% of original gro uped cases correctly classified Figure 15. Classification results for grouped and ungrouped cases. The two tables show predicted group membership in terms of individual totals (a) and percentages (b). Both morphological results relied on disc riminant function analysis in order to predict group membership of unknown cases. FORDISC utilizes a tailored discriminant function analysis specific for forensic an thropological race identi fication (Giles and Elliot, 1962). SPSS was also able to discriminate between groups and classify unknown cases into population ca tegories. Both analyses resulted in similar classification groups. Results for cases 88-0220 (FORDISC = HM, SPSS = A/H), and 98-1909 (FORDISC = BM, SPSS = W) were the same in both mor phological analyses. FO RDISC originally classified case 82-0607 as white male, but a la rger group analysis resulted in change of 6-0486 were outliers near the Caucasian grouping in the SPSS analysis. FORDISC classified th ales. This result is not unexpected considering the fact t the daset for c le cranial measurements. Female craniometrics are typically smaller across population groups than their male counterparts. Molecular vs. Morphological Assessments: The results for cases th were luded both mlecular and morphological assessments agreed in four out of six cases: cases 82-0013, 82-1005, 96-0486, and 98classification to Vietnamese male. SPSS classified this case as white. Cases 82-0013 and 9 ese cases as white fem hat ta SPSS in luded only know n ma at inc in o B Grouping Predicted Group Membership % Total 1 2 3 4 5 Original Count 1 83.1 1.7 8.5 5.1 1.7 100.0 2 .0 94.7 5.3 .0 .0 100.0 3 15.8 15.8 31.6 15.8 21.1 100.0 4 10.0 .0 16.0 62.0 12.0 100.0 5 .0 .0 8.3 16.7 75.0 100.0 Ungrouped cases 71.4 14.3 .0 14.3 .0 100.0


41 Tatched the mologicsse ss 909 alsoas classified in the same group in both analyses as an individual of African scent. Te fifth ce, 880, was assified Hisp male in the FORDISC ancestry in the molecular analysis. This is ount of taxa in the phylogenetic ican descent, a majority of these countries rather than individuals of Spanish/European ancestry (Appendix A). Case 88confounded the mtDNA results. Cases CC94-0011 and 98-3587 were included in the molecular analysis, but did not contain complete crania for morphological assessment 0607 and 95-6189 did not amplify. 1909 (Table 5). Cases 82-0013, 82-1005, and 96-0486 grouped with taxa of European descent his m rpho al a ment of Caucasian ancestry. Case 981 w de h as 220 cl as anic assessment, but grouped with taxa of African not surprising due to the limitations of usi ng a large am analysis. Despite the inclusion of taxa of La tin Amer individuals represented indigenous Latin Am erican population groups of their respective 0220 could have had a distant ma ternal ancestor of African descent, which would have using all twenty-one cranial measurements described in Table 3. DNA from cases 82


42number. Cases in red matched in both molecular and morphological results. Cases in blue differed in both Table 5. Summary table showing the molecular vs. morphological assessments for each unidentified case analyses. Case Number. Molecular Results FORDISC 2.0 Results SPSS results Dr. Weinkers Assessment 82-0013 European White Fe male Outlier/Caucasian No assessment 82-1005 European White Fema le Caucasian White Female 96-0486 European White Female Outlier/Caucasian White Female 98-1909 African Black Male African Black Male 88-0220 African Hispanic Male Asian/Hispanic Asian Male 82-0607 No assessment Vietnamese Male Caucasian No assessment CC94-0011 European No assessment No assessment No assessment 95-6189 No assessment No assessment Caucasian Black Male 98-3587 European No assessment No assessment Black Male


43 CHAPTER 4 Conclusions The primary purpose of this research wa s to test molecular phylogenetics as a means of identification in forensics, and to compare molecular results to the traditional means of identifying populati on affiliation in forensic an thropology using morphology. The secondary goal was to develop a new method of skeletal extraction of DNA to limit destruction of bone for further morphological an alysis while still a llowing ample DNA to be recovered for mo lecular analysis. The new drilling method of bone extr action allowed a 70% success rate in direct sequencing of samples. It provided a practical means of ex traction of skeletal material without destroying the bone for further morphological analysis. This was performed using less than 100 mg of bone powde r while leaving only a 1/8 drill hole in the bone. Population affiliation in this study was used to describe the geographic origin of an individual of a particular human populat ion. The terms European, African, and Asian reflect this terminology. The use of the term Hispanic, which is an traditionally used to describe ethnicity or cultural affiliation, is us ed in the medico-legal profession to define individuals originating from Latin America, without necessarily taking into account the potential for European/Spanish or indigenous ancestry. This type of ambiguity leads to the rejection of races in th e anthropological and biological arena, due to the lack of


44 clearly delineable variation in the distribu tion of human populations. Since it is necessary to take into account phenotypic differences in human populations for law enforcement purposes, forensic scientists have to work within this system of classification. Molecular analysis revealed that it is possible to compare mitochondrial DNA HVR I sequences of many individuals from different population groups phylogenetically. The neighbor-joining trees constructed from the analysis revealed separate clades representing three population groups; European descent, African descent, and Latin American descent. The sequences representing individuals of Latin American descent were split into two separate clades. The first clade included individua ls with the potential of having European/Spanish descent. The s econd Latin American clade contained more taxa representing a varied geographic an cestry. This group included indigenous populations from Latin America. The first clad e appears to consist of Latin American individuals with a southern European/Spa nish mitochondrial lineage while the second clade may represent Latin American indivi duals with a native American mitochondrial lineage. Only 81% of the individuals from the da tabases that reported African ancestry were included in the African-descent clad e while most of the remaining 19% of individuals fell within the European clade. Similarly, only 88% of the individuals from the databases that reported European ancestr y were included in the European-descent clade while the remaining 12% fell into a va riety of other clades This suggests that nearly 19% of individuals w ho self-identify with an Afri can descent have a European mitochondrial genome and approximately 12% of individuals who se lf-identify with a European descent have non-European mitochondrial genomes.


45 In the molecular analyses, five cases of seven grouped with individuals of European descent. Two cases grouped within the African-descent clade. When compared to the results from FORDISC 2.0, four out of the five cases matched the predicted grouping. Individuals who were positively identified (HeLa, 98-1909) also grouped appropriately. The SPSS analysis allowed for another means of analyzing the morphological data independently from FORDISC. This analysis was conducted using a dataset of metrics from individuals of known popula tion background. The discriminant function analysis allowed decent separation between population groups. The unidentified cases were added to the analysis and population affiliation identified by locality in the canonical plot. Both morphological analyses reflected the same results in all cases but one. Several factors may preclude the use of molecular phylogenetics for forensics. Mitochondrial DNA reflects only the maternal lineage, which allows reconstruction of genealogy due to lack of recombination. This may not reflect phenotype, which is the most important factor from the medico-legal standpoint. Self-identification of population affiliation may not necessarily reflect the mol ecular classification of an individual. The presence of a maternal ancestor different from the self-identif ied population of the individual would reveal a separate mito chondrial DNA ancestry. Despite the limitations of mtDNA analysis, it is another means of determining id entifying information about a decedent when lack of comparison material is present for forensic analysis. From the small dataset analyzed in this study, mitochondrial ancestry appeared to match phenotype in approximately 80% of the cases. This is based on results from both the unidentified


46 remains and from the sequences taken from the databases. The comparison of the morphometrics and molecular phylogenetics, ma y together provide a better means of identification of the population affiliation of a descendent based on skeletal material.


47 REFERENCES 1. Anderson, S., A.T. Bankier, B.G. Barrel, MHL Dubruihn, A.R. Coulson, J. Drouin, & Eperon, I.C. 1981. Sequen ce and Organization of the Human Mitochondrial Genome. Nature. 290: 457-465. 2. Brace, C.L. (1982). The roots of the race concept in American physical anthropology. In F Spencer (ed.): A History of American Physical Anthropology : 1930-1980. New York: Academic Press pp. 11-29. 3. Brace, C.L. (2000). A four-letter word called race. In L. Reynolds and L. Lieberman (eds.): Race and Other Miscalculations, and Mismeasures: Papers in Honor of Ashley Montagu. Dix Hills, N.Y.: General Hall, Publishers 4. Brandon, M. C., Lott, M. T., Nguyen, K. C., Spolim, S., Navathe, S. B., Baldi, P. & Wallace, D. C. (2004). MITOMAP: a human mitochondrial genome database. Nucleic Acids Research 33 (Database Issue):D611-613, 2005. URL: 4. Brown, R., & Armelagos, G. 2001. Apportionm ent of Racial Diversity: A Review. Evolutionary Anthropology. 10:34-40. 5. Brown, W.M. (1985). The mitochondrial geno me of animals. In R.J. MacIntyre (ed.): Molecular Evolutionary Genetics. Plenum, N.Y. pp. 95-130. 6. Butler, J.M, & Levin, B.C. (1998). Forensic Applications of Mitochondrial DNA. Trends in Biotechnology. 16(4): 158-162. 7. Butler, J. (2001) Forensic DNA Typing Biol ogy and Technology Behind STR Markers. London: Academic Press. pp. 1-10. 8. Cavalli-Sforza, L.L. & Bodmer, W.F. (1976). Genetics, Evolution, and Man. San Francisco: WH Freeman and Co. 9. Cooper, A. & Poinar, H. (2000). Ancient DNA: Do It Right or Not at All. Science. 289 (5482): 1139-1140.


48 10. Fisher, C.L., Isenberg, A.R., Stewart J ., Miller, K., Theisen, C., Wilson, M. DiZinno, J., & Budowle, B. (1/2005) Mitochondrial DNA: Today and Tomorrow. Laboratory Division, Fede ral Bureau of Investigation. 11. Giles, E. & Elliot, O. (1962). Race Identi fication from Cranial Measurements. Journal of Forensic Sciences. 7: 147-157. 12. Giles, E. & Elliot, O. (1963). Sex De termination by Discriminant Function Analysis of Crania. American Journal of Physical Anthropology. 21: 5368. 13. Gill, G.W. & Rhine, S. Editors. (1990). Skeletal Attribution of Race. Anthropological Papers No.4. Albuquer que: University of New Mexico. 14. Gould, S.J. (1983). The Mismeasure of Man. New York: WW Norton. 15. Handt, O., Meyer, S., & Von Haeseler, A. (1998). Compilation of human mtDNA control region sequences. Nucleic Acids Research. 26(1): 126129. 16. Hartl., D. & Jones, E.W. (1998). Genetics: Principles and Analysis Boston: Jones and Bartlett Publishers. 17. Hochmeister, M.N., Budowle, B., Bore r, U.V., Eggmann, U., Comey, C.T., & Dirnhofer. R. (1991). Typing of Deoxyr ibonucleic Acid (DNA) Extracted from Compact Bone from Human Remains. Journal of Forensic Sciences. 36 (6): 1649-1661. 18. Howells, W.W. (1989). Skull Shapes and the Map. Cambridge: Harvard University Press. 19. Isenberg, A.R. & Moore, J.M. (1999). M itochondrial DNA Analysis at the FBI Laboratory. Forensic Science International. 1(2). 20. Jantz, R.J. & Moore-Jansen, P.H. (2000) Database for Forensic Anthropology in the United States, 1962-1991. Inter-university Consortium for Political and Social Research (no. 2581). 21. Johnston, F. & Little, M. (2000). Histo ry of Human Biology in the United States. In Human Biology: An Evolutionar y and Biocultu ral Perspective. New York: John Wiley and Sons, Inc. pp. 27-42. 22. Kimura, M, (1979). The neutral th eory of molecular evolution. Scientific American. 241(5):98-126.


49 23. Kittles, R., Bergen, A., Urbanek, M., Vi rkkunen, M., Linnoila, M., Goldman, D., & Long, J. (1999). Autosomal, Mito chondrial, and Y chromosome DNA Variation in Finland: Evidence for a Male-specific Bottleneck. American Journal of Physical Anthropology. 108: 381-399. 24. Lewontin, R.C. (1974). The Genetic Basis of Evolutionary Change. New York: Columbia University Press. 152-157. 25. Macaulay, V. Richards, M. & Syke s, B. (1999) Mitochondrial DNA recombination no need to panic. Proceedings of the Royal Society of London. 226: 2037-2039. 26. Miller, K., Dawson, J.L., & Hagelber g, E. (1996). A Concordance of Nucleotide Substitutions in the First and Second Hypervariable Segments of the Human mtDNA Control Region. International Journal of Legal Medicine. 109: 107-113. 27. Molnar, S. (2002). Human Variation: Races, Types, and Ethnic Groups, Fifth Edition Englewood Cliffs, New Jersey: Prentice Hall. 28. Nei, M. & Kumar, S. (2000). Molecular Evolution and Phylogenetics. New York: Oxford Press. 29. Ousley, S.D. & Jantz, R.L. 1996. FORDISC 2.0: Personal Computer Forensic Discriminant Fuctions. The University of Tennessee, Knoxville. 30. Prado, V.F., Castro, A.K.F., Oliveira, C.L ., Souza, K.T., & Pena, S.D.J. (1997). Extraction of DNA from Human Skeletal Remains: Practical Applications in Forensic Sciences. Genetic Analysis: Biomolecular Engineering. 14: 41-44. 31. Rabinow, R. (1996). Making PCR: A Story of Biotechnology. London: University of Chicago Press. 32. Rhine, S. (1998). Bone Voyage: A Journey in Forensic Anthropology. Albuquerque: University of New Mexico Press. 33. Romualdi, C., Balding, D. Nasidze, I., Risch, G., Robichaux, M., Sherry, S., Stoneking, M., Batzer, M., & Barbujan i, G. 2002. Patterns of Human Diversity, within and among continen ts, inferred from biallelic DNA polymorphisms. Genome Research 12: 602-612.


50 34. Lombroso, C. (1968). Crime: Its Causes and Remedies. Montclair, NJ: Legal Classics Library. 35. Tajima, F. (1989). Statistical me thod for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585-595. 36. Tajima, F. & Nei, M. (1984). Estimati on of Evolutionary Distance Between Nucleotide Sequences. Molecular Biology and Evolution. 1(3): 269-285. 37. Tang, J. (2003) New York Academy of Sciences Examines DNA Forensic Techniques Emerging From WTC Tragedy. 38. Ubelaker, R. (2002) Application of Forensic Discriminant Functions to a Spanish Cranial Sample. Forensic Science Communications 4(3). 39. Wittig, H., Augustin, C., Baasner, A., Bulnheim, U., Dimo-Simonin, N., Edelmann, J., Hering, S., Jung, S., Lutz, S., Michael, M., Parson, W., Poetsch, M., Schneider, P.M., We ichhold, G., & Krause, D.(2000). Mitochondrial DNA in the central European population Human identification with the help of the forensic mtDNA D-loop Data base. Forensic Science International. 113: 113-118.




Appendix A: HVRBASE sequence designations. The following designations are representations of the sequences used in the phylogenetic analysis: European Ancestry African Ancestry Latin American Ancestry Asian Ancestry Amerindian Ancestry Designations R1-R22 BR1-BR15 German 1-26 AFAM, AFAMO 1106 Brazilian (BRO), Columbian 1-20, Mexican 1-20 Chilean 1-20 China Huetar 4-20; Boruca 1-3; Mayan 4, 23, 34,42 Population European American, European African American Brazilian, Columbian, Mexican Chilean Chinese Costa Rican, Mexican Source Citation Kittles et al. 1999 Am. J. Phys. Anthropology 108: 381-399 Brown et al. 1998. Am. J. Hum. Genetics 63: 1852-1861. Parsons et al. Armed Forces Institute of Pathology. Alves-Silva et al. 1999 Hum Biol. 71: 245-259 Torroni et al. 1993 Am. J. Hum Gen. 53: 563-590 Horai et al. MBE 10: 23-47, 1993. Jorde et al. 1995. Hum Genet. 57: 523-538 Torroni et al. 1993 Am. J. Hum Gen. 53: 563-590 Santos et al. 1994. Hum Biol. 6:963977 51






54 98-1909 AGCACATTACAGTCAAATCCCTTC--TCGTCCCCATGGATGACCCCCCTCAGATAGGAGT 88-0220 AGCACATTACAGTCAAATCCCTTC--TTGTCCCCATGGATGACCCCCCTCAGATAGGAGT *********************** * ********************* cheek3 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAG-------------Cheek2 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG--CTACTCTCCT 82-0013 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG-----------Cheek8 CCCTTGACCACCATCCTCCGTGAAATCAATAT-CCCGCAACAAGAGTG--CTACTCTCCT 82-1005 CCCTTGACCACCATCCTCCGTGAAATCAATATATCCCGCACAAG---------------MITOMAPREF -----------------------------------------------------------Cheek7 CCCTTGACCACCATCCTCCGTGAAATCAATATCCC------------------------96-0486 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG--CTACTCTCCT Cheek5 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG--CTACTCTCCT 98-3587 CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG--CTACTCTCCT HeLa CCCTTGACCACCATCCTCCGTGAAATCAATAT--CCCGCACAAGAGTG--CTACTCTCCT Cheek4 CCCTTGACCACCATCCTCCGTGAAATCAATAT-CCCGCA-CAAGAGTG--CTACTCTCCT CC94-0011 CCCTTGGCCACCATCCTCCGTGAAATCAATAT-CCCGCAACAAGAGTG--CTACTCTCCT 98-1909 CCCTTGACCACCATCCTCCGTGAAATCAATAT-CCCGCAACAAGAGTG--CTACTCTCCT 88-0220 CCCTTGACCACCATCCTCCGTGAAATCAATAT-CCCGCAACAAGAGTGTGCTACTCTCCT cheek3 -----------------------------------------------------------Cheek2 CGCTCCGGGCCCATAACACTTGGGGGTAGCTAAAGTGAACTGTATCCGACATCTGGTNCC 82-0013 -----------------------------------------------------------Cheek8 CGCTCC-----------------------------------------------------82-1005 -----------------------------------------------------------MITOMAPREF -----------------------------------------------------------Cheek7 -----------------------------------------------------------96-0486 CGCT-------------------------------------------------------Cheek5 CGCTCCGGGCCCATAACACTTGGGGGTAGCTAAAGTGAACTGTATCCGACATCTGGTTCC 98-3587 CGCTCCGGGCCCATAACACTTGGGGGTAG------------------------------HeLa CGCTCCGGGCCCATAACACTTGGGGGTAGCTAAAG------------------------Cheek4 CGCTCCGGGCCCATAACACTTGGGGGTAG------------------------------CC94-0011 CGCTCCGGGCCCATAACACTTGGGGGTAGCTAAAG------------------------98-1909 CGCTCCGGGCCCATAACACTTGGGGGTAGGCTAAAG-----------------------88-0220 CGCTCCGGGCCCATAACACTTGGGGGTAGCTAAAA------------------------cheek3 -------------------------------Cheek2 TACTACA------------------------82-0013 -------------------------------Cheek8 -------------------------------82-1005 -------------------------------MITOMAPREF -------------------------------Cheek7 -------------------------------96-0486 -------------------------------Cheek5 TACTACAGGGAATCCAAGGAAAACCGAGGAAA 98-3587 -------------------------------HeLa -------------------------------Cheek4 -------------------------------CC94-0011 -------------------------------98-1909 -------------------------------88-0220 -------------------------------


Appendix C: Bootstrapped Neighbor Joining Tree 1 Neighbor-joining tree utilizing Tajima and Ne i distance parameter. Outgroup includes Pan troglodytes and Homo neandertalensis. 55


Appendix C: Continued 56


Appendix C: Continued 57


Appendix C: Continued 58


Appendix C: Continued 59


Appendix C: Continued 60


Appendix C: Continued 61


Appendix C: Continued 62


Appendix C: Continued 63


Appendix D: Bootstrapped Neighbor-Joining Tree 2 NJ tree utilizing Tajima-Nei Distance Parameter. Excludes Pan troglodytes as outgroup. 64


Appendix D: Continued 65


Appendix D: Continued 66


Appendix D: Continued 67


Appendix D: Continued 68


Appendix D: Continued 69


Appendix D: Continued 70


Appendix D: Continued 71


72 Appendix E: FORDISC Ou tputs for Case 82-0607 Multigroup Classification of 82-0607 ---------------------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ----------------------------------------------------------------------WM **WM** 94.0 .811 .000 WF 102.1 .014 .000 BM 97.1 .171 .000 BF 106.1 .002 .000 HM 107.3 .001 .000 -----------------------------------------------------------------------is closest to WMs Discriminant function results using 14 variables: GOL XCB ZYB BPL MAB AUB UFHT WFB NLH NLB OBB OBH FRC PAC OCC FOL ---------------------------------------------------------------------------Group Total Into Group Percent Number WM WF BM BF HM Correct ---------------------------------------------------------------------------WM 115 88 13 6 0 8 76.5% WF 81 10 67 0 0 4 82.7% BM 88 4 0 67 13 4 76.1% BF 82 1 4 9 65 3 79.3% HM 32 5 1 2 2 2 68.8% -----------------------------------------------------------------------------Total: 398 Correct: 309 77.6%


73 Appendix E: Continued Discriminant function results using 14 variables: XCB ZYB BPL MAB AUB UFHT WFB NLH NLB OBB OBH FRC PAC OCC -------------------------------------------------------------------------------------------------------------Group Total Into Group Percent Number WM WF BM BF AM AF JM JF HM CHM VM Correct -------------------------------------------------------------------------------------------------------------WM 168 110 17 5 0 3 2 5 0 8 5 13 65.5% WF 133 12 102 0 5 0 1 0 6 5 0 2 76.7% BM 125 7 1 60 14 1 1 10 3 10 12 6 48.0% BF 107 1 4 9 73 0 3 0 11 4 0 2 68.2% AM 46 2 0 0 0 32 3 2 2 0 4 1 69.6% AF 28 0 1 1 0 5 15 1 2 1 1 1 53.6% JM 100 2 3 8 0 9 3 43 10 7 11 4 43.0% JF 100 1 6 2 14 0 5 9 55 4 1 3 55.0% HM 37 4 0 3 3 1 3 2 1 18 1 1 48.6% CHM 79 5 0 9 0 1 8 4 1 5 40 6 50.6% VM 51 4 2 1 0 1 2 3 3 4 5 26 51.0% ----------------------------------------------------------------------------------------------------------------Total: 974 Correct: 574 58.9%


74 Appendix F: FORDISC Ou tputs for Case 82-1005: FORDISC 2.0 Analysis of 82-1005 Discriminant function results using 20 variables: GOL XCB ZYB BBH BNL BPL MAB AUB UFHT WFB NLH NLB OBB OBH EKB DKB FRC PAC OCC FOL ------------------------------------------------------Group Total Into Group Percent Number WM WF BM BF HM Correct ------------------------------------------------------WM 115 94 11 4 0 6 81.7 % WF 81 7 68 0 2 4 84.0 % BM 89 3 0 68 13 5 76.4 % BF 82 1 4 8 67 2 81.7 % HM 32 2 1 1 2 26 81.3 % ------------------------------------------------------Total: 399 Correct: 323 81.0 % Multigroup Classification of 82-1005 ----------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ----------------------------------------------------------WM 37.7 .002 .010 WF ** WF ** 25.3 .991 .191 BM 43.1 .000 .002 BF 37.2 .003 .011 HM 36.3 .004 .014 ----------------------------------------------------------is closest to WFs


75 Appendix G: FORDISC Outputs for Case 88-0220: FORDISC 2.0 Analysis of 88-0220 Discriminant function results using 21 variables: GOL XCB ZYB BBH BNL BPL MAB MAL AUB UFHT WFB NLH NLB OBB OBH EKB DKB FRC PAC OCC FOL ------------------------------------------------------Group Total Into Group Percent Number WM WF BM BF HM Correct ------------------------------------------------------WM 114 92 10 4 0 8 80.7 % WF 81 5 69 0 2 5 85.2 % BM 88 3 0 67 13 5 76.1 % BF 82 1 4 9 67 1 81.7 % HM 32 2 1 1 2 26 81.3 % ------------------------------------------------------Total: 397 Correct: 321 80.9 % Multigroup Classification of 88-0220 ----------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ----------------------------------------------------------WM 60.9 .028 .000 WF 64.5 .005 .000 BM 60.8 .030 .000 BF 56.9 .209 .000 HM ** HM ** 54.4 .729 .000 ----------------------------------------------------------is closest to HMs


76 Appendix H: FORDISC Outputs for Case 96-0486: The following is the discriminant function and multigroup classification results for the preliminary five group analysis for case 96-0486: FORDISC 2.0 Analysis of 96-0486 Discriminant function results using 21 variables: GOL XCB ZYB BBH BNL BPL MAB MAL AUB UFHT WFB NLH NLB OBB OBH EKB DKB FRC PAC OCC FOL Group Total Into Group Percent Number WM WF BM BF HM Correct WM 114 92 10 4 0 8 80.7 % WF 81 5 69 0 2 5 85.2 % BM 88 3 0 67 13 5 76.1 % BF 82 1 4 9 67 1 81.7 % HM 32 2 1 1 2 26 81.3 % Total: 397 Correct: 321 80.9 % Multigroup Classification of 96-0486 Group Classified Distance Probabilities into from Posterior Typicality WM 75.5 .210 .000 WF ** WF ** 73.3 .639 .000 BM 76.9 .107 .000 BF 78.8 .042 .000 HM 84.5 .002 .000 is closest to WFs


77 Appendix H: Continued The following is the eleven group analysis results: FORDISC 2.0 Analysis of 96-0486 Discriminant function results using 17 variables: GOL XCB ZYB BBH BNL BPL MAB AUB UFHT WFB NLH NLB OBB OBH FRC PAC OCC ---------------------------------------------------------------------------------------------------------------Group Total Into Group Percent Corr ect Number WM WF BM BF AM AF JM JF HM CHM VM -----------------------------------------------------------------------------------------------------------------WM 166 120 17 2 0 3 1 4 0 11 8 0 72.3 % WF 132 11 104 0 3 0 1 0 6 6 0 1 78.8 % BM 125 7 1 70 17 2 1 8 3 10 5 1 56.0 % BF 107 1 4 10 77 0 2 0 9 4 0 0 72.0 % AM 46 0 0 0 0 32 5 1 2 0 5 1 69.6 % AF 28 0 1 1 0 4 16 0 2 2 1 1 57.1 % JM 100 1 0 8 1 10 1 47 10 7 11 4 47.0 % JF 100 1 5 2 9 0 4 5 59 5 4 6 59.0 % HM 37 4 0 3 3 1 1 2 0 19 3 1 51.4 % CHM 79 1 0 6 0 0 6 7 2 5 48 4 60.8 % VM 51 0 0 0 0 0 0 2 2 1 5 41 80.4 % ------------------------------------------------------------------------------------------Total: 971 Correct: 633 65.2 %


78 Appendix H: Continued Multigroup Classification of 96-0486 ----------------------------------------------------------Group Classified Distance Probabilities into from Posterior Typicality ----------------------------------------------------------WM ** WM ** 45.2 .412 .000 WF 45.3 .401 .000 BM 48.9 .065 .000 BF 49.2 .057 .000 AM 66.0 .000 .000 AF 54.6 .004 .000 JM 52.7 .010 .000 JF 52.4 .012 .000 HM 50.1 .035 .000 CHM 54.1 .005 .000 VM 69.1 .000 .000 ----------------------------------------------------------is closest to WMs

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001680954
003 fts
005 20060215071026.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 051202s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001167
QH307.2 (Online)
1 100
Wetherington, Hattie Bea.
2 245
A phylogenetic approach for the study of variation and determination of population affiliation of indigent human skeletal remains
h [electronic resource] /
by Hattie Bea Wetherington.
[Tampa, Fla.] :
b University of South Florida,
Thesis (M.S.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 86 pages.
ABSTRACT: Mitochondrial DNA (mtDNA) has played a major role in human population studies over the past decade due to its maternal inheritance and negligible recombination (Macaulay, 1999). The mtDNA control region has been the focus of these studies due to the highly polymorphic nature of this non-coding region. Forensic scientists also use mtDNA to help determine the identity of missing individuals when nuclear DNA is not present. However, when skeletal remains are unclaimed, identification becomes near impossible. Therefore, mtDNA can play a valuable role in identification in terms of population affiliation, especially in conjunction with morphological analysis. The goals of this research were two-fold: 1) to determine population affiliation of unknown skeletal samples using phylogenetics and 2) to find a method of extraction that leaves a majority of the remains intact.This research depended on the donation of samples from sixteen skeletal remains from the Hillsborough County Medical Examiners office. Mitochondrial DNA from ten of these cases were extracted, amplified, and sequenced in order to determine population affiliation via phylogenetic analysis of hypervariable region I (HVR I). These sequences were aligned and compared to that of sequences in a pre-existing mtDNA control region database (Handt, 1998). The crania of the skeletal remains were measured and subsequently analyzed by the forensic anthropology program FORDISC 2.0 to morphologically determine population affiliation. A secondary morphological analysis included input of the measurements into SPSS, a statistical program package, as a separate discriminant function assessment. This analysis was dependent on a database of craniometrics from known individuals (Jantz, and Moore-Jansen, 2000).
Adviser: Dr. James R. Garey.
Mitochondrial dna.
Molecular forensics.
Discriminant analysis.
0 690
Dissertations, Academic
x Biology
t USF Electronic Theses and Dissertations.
4 856