UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Allelic frequencies of two DNA restriction fragment length polymorphism systems collected from an undefined… Bonnycastle, Lori L.C. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_bonnycastle_lori_l_c.pdf [ 3.13MB ]
JSON: 831-1.0086674.json
JSON-LD: 831-1.0086674-ld.json
RDF/XML (Pretty): 831-1.0086674-rdf.xml
RDF/JSON: 831-1.0086674-rdf.json
Turtle: 831-1.0086674-turtle.txt
N-Triples: 831-1.0086674-rdf-ntriples.txt
Original Record: 831-1.0086674-source.json
Full Text

Full Text

ALLELIC FREQUENCIES OF TWO DNA RESTRICTION FRAGMENTLENGTH POLYMORPHISM SYSTEMS COLLECTED FROM ANUNDEFINED POPULATION : APPLICATION TO DNA PROFILINGbyLORI L. C. (NEE CHEUNG) BONNYCASTLEB.Sc. (Microbiology), University of British Columbia, 1984M.Sc. (Med. Microbiology), University of British Columbia, 1987A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF PATHOLOGYWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMarch 1992© Lori L.C. BonnycastleSignature(s) removed to protect privacyIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)________________________________Department of-The University of British ColumbiaVancouver, CanadaDateDE-6 (2/88)Signature(s) removed to protect privacyVTABLE OF CONTENTSHLi ocGENERAL IN1RODUCTION 1THESIS HYPOTHESIS 17MATERIALS 20CHAPTER 1: VARIABLE NUMBER OF TANDEM REPEAT (VNTR) ANALYSISIntroduction 24Methods 26Section I (data base generation) 38Section II (forensic case sample analysis) 107Conclusions 128CHAPTER 2: DIMORPHIC RESTRICTION ENDONUCLEASE RECOGNITIONSITE ANALYSISIntroduction 130Methods 134Section I (data base generation and reliability studies) 142Section II (discriminating power studies) 170Conclusions 177GENERAL CONCLUSIONS 179SUMMARY CONCLUSIONS 183ADDENDUM: EFFECTS OF ENVIRONMENT ON DNA STABILITYIntroduction 186Methods 188Results and Discussion 19211ABSTRACTRestriction fragment length polymorphisms (RFLPs) within the deoxyribonucleic acidmolecule were analysed to determine the extent of their ability to provide discriminating humanDNA profiles for forensic identification. The two types of RFLPs studied were variable number oftandem repeats (VNTR), the technique commonly used by North American forensic laboratories,and dimorphic restriction endonuclease recognition site (RES) developed in this laboratory.Data bases showing allele frequencies were collected for four VNTR hypervariable regions,D2S44, D16S85, INS, and D14S13 from an undefined population group obtained from the GreaterVancouver Area. Allele frequency distributions from this group were similar to the larger andmore statistically defined allele distributions obtained by other laboratories, the RCMP MolecularBiology Unit in Ottawa and Promega associated laboratories in all of which analysed Caucasianpopulations.An alternative technology to VNTR was based on the selection of five regions within thehuman genome that contain dimorphic restriction endonuclease recognition sites (RES). Anucleotide sequence of approximately one kilobase that contained the RES was selected from thegenes of the following proteins: adenine phosphoribosyltransferase (APRT), prealbumin (PALB),adenine deaminase (ADA), carbonic anhydrase II (CAH) and lipoprotein lipase (LPL). Afteramplification of the nucleotide sequence by the polymerase chain reaction (PCR), the presence orabsence of the internal RES was analysed by restriction endonuclease digestion followed by agarosegel electrophoresis. From the pattern (size and number) of the restriction fragments generated, itis possible to make genotype assignments with respect to the RES. Restriction endonucleaseactivity was monitored by the presence of a control DNA fragment which is also cleaved by therestriction endonuclease used for each specific analysis. Statistical calculations demonstrated thatby using 16 unlinked dimorphic RES, rapid individual identification to a certainty of one in 6.5million is possible.111Analysis of casework and/or laboratory simulated forensic specimens indicate that both theVNTR and RES system are discriminating and can be applied to non-ideal specimens. However,the RES system offers a much simpler and more reliable methodology.ivACKNOWLEDGEMENTSI would like to express my gratitude to my supervisor, Dr. J.A.J. Ferris, for providing mewith the opportunity to carry out this project in his laboratory, and for his support and guidancethroughout the course of this study. I would also like to thank the members of my supervisorycommittee, Drs. A.P. Autor, G.R. Gray, B. McGillivray, and G. Tener, for their guidance andhelpful suggestions. My sincere thanks also go to Mr. K.McNeil for his valuable discussions andtechnical assistance during this project, to Dr. M. Altamirano-Dimas for the generous use of hislaboratory equipment and materials, and to Drs. M. Schultzer and E. Wijsman both of whom havehelped tremendously with the statistical aspects of this analysis.My love and thanks goes to my parents who have always had great confidence in myabilities and have urged me forward to meet new challenges. I would also like to express my loveand gratitiude to my husband Jeffrey who not only assisted with drawing of figures and tables inthis thesis, but has shown great patience and has provided unwavering support during my trying butrewarding journey through this thesis project.Conclusions .214REFERENCES .215viviiLIST OF FIGURESFigure 1. Assessment of Accuracy and Precision ofElectrophoretic Separation 352. Conditions For DNA Digestion 423. VNTR Analysis With Probe YNH24 464. VNTR Analysis With Probe 3’HVR a-globin 475. VNTR Analysis With Probe phins 310 486. VNTR Analysis With Probe CMM1O1 497. Two Allele Frequency Distributions For the D2S44 Locus 598. Allele Frequency Distribution For the D2S44 Locus 619. Two Allele Frequency Distributions For the D16S85 Locus 6710. Allele Frequency Distribution For the D16S85 Locus 6911. Two Allele Frequency Distributions For the INS Locus 7412. Allele Frequency Distribution For the INS Locus 7613. Two Allele Frequency Distributions For the D14S13 Locus 7914. Alelle Frequency Distribution For the D14S13 Locus 8115. Comparison of Allele Frequency Distributionsof D2S44 and D14S13 8516. Simultaneous Probing of Two Loci 8717. Comparison of Allele Frequency Distributionsof D2S44 and D16S85 9018. Comparison of Data From Computerized VersusNon-computerized Measurements 9619. D2S44 Allelic Frequency Distribution Comparison 10020. D16S85 Allelic Frequency Distribution Comparison 10121. D14S13 Allelic Frequency Distribution Comparison 10222. Case I - Locus D16S85 109viii23. Case I - Locus D2S44 .11124. Case II - Locus D16S85.11625. Case II - Locus D2S44 11826. Case III - Locus D16S85 12227. Polymerase Chain Reaction 13128. Amplification of a Dimorphic Locus 13229. Dimorphic RES Loci 13630. APRT Locus 14331. LPL Locus 14432. PALB Locus 14533. ADA Locus 14634. CAH Locus 14735. Amplification Products of Five Loci 14936. Probing of Amplification Products 15137. Amplification of Non-Human DNA 16438. Amplification of Human DNA in Non-Human Samples 16739. Amplification of DNA From Aged Stains 19540. VNTR Analysis of DNA from Aged Stains 19641. Amplification of DNA From Bloodstain Exposed toExternal Environment (APRTJPALB/ADA) 19842. Amplification of DNA From Bloodstain Exposed toExternal Environment (PALB) 19943. Inconsistent Amplification 20144. VNTR Analysis of DNA Extracted FromExternal/UV/Soil Bloodstains 20345. Amplification of DNA Extracted From UVIrradiated Bloodstains 20846. Amplification of DNA Extracted From UVIrradiated and Soil Bloodstains 210ixLIST OF TABLESTable 1. General Properties of VNTR Probes Used 442. Accuracy of DNA Typing System 503. Standard Deviation Comparisons of MolecularWeight Standards and Genomic DNA Fragments 524. Relative Standard Deviation 545. Allele Frequencies of Four VNTR Loci 566. Comparison of the Most Common Alleles 647. D2S44 Cumulative Frequency For a SpecificSize Range 658. D16S85 Cumulative Frequency For a SpecificSize Range 719. D14S13 Cumulative Frequency For a SpecificSize Range 8410. Measurement Imprecision 9311. Comparison of Heterozygote Frequencies 10512. Recovery of DNA From Hair Bulbs 12413. Calculated Sizes of Dimorphic RES LociAmplification products 14814. Allelic Frequencies at the APRT locus 15315. Allelic Frequencies at the LPL locus 15416. Allelic Frequencies at the PALB.locus 15617. Subgroup Comparisons at the PALB locus 15718. Allelic Frequencies at the ADA locus 15919. Subgroup Comparisons at the ADA Locus 16020. Allelic Frequencies at the CAH locus 16221. Subgroup Comparisons at the CAH Locus 16322. Dimorphic RES Allelic Profiles Ri 17223. Dimorphic RES Allelic ProfilesP2.17324. Dimorphic RES Allelic Profiles Fl 17425. Dimorphic RES Allelic Profiles F2 175xxiABBREVIATIONSBRL: Bethesda Research LaboratoriesDTT: dithiothreitolEDTA: ethylenediamine tetraacetic acidEtBr: ethidium bromideHetObS: observed heterozygotesHomObS : observed homozygotesHetexp : expected heterozygotesHomexp : expected homozygotesH.W.E. : Hardy-Weinberg EquilibriumLB : Luria-BertaniMol. Wt. Std. : molecular length standardPCR: polymerase chain reactionRES : restriction endonuclease recognition siteRFLPs: restriction fragment length polymorphismsSDS: sodium dodecyl sulfateTris : tris(hydoxymethyl) aminomethaneUV: ultraviolet lightVNTRs : variable number of tandem repeats1GENERAL INTRODUCTIONAt the University of Lyon in France in 1910 Edmund Locard formulated his exchangeprinciple in which he postulated that the criminal always leaves something behind at the scene ofhis crime which was not there before and similarly carries away with him something which was noton him when he arrived [1]. One of the most important applications of the Locard exchangeprinciple for the forensic scientist is to either associate or disassociate biological specimens such asblood and semen stains with persons under suspicion of committing the crime under investigation.Forensic science may be defined as the application of scientific techniques to the investigation ofcrime. Forensic scientists are often required to examine materials associated with an incident,either suspected or known to contravene the law.Until fairly recently, forensic serology routinely involved the application of immunological,serological and biochemical methods to blood and body fluids and stains of blood and body fluids[2-4]. These tests are possible due to genetic markers in human populations which are detected asinherited protein variants. The five major classes of genetic markers are blood group antigens,isoenzymes, serum group antigens, haemoglobin variants, and the human leucocyte antigens (HLA).In most instances, the analysis is based upon the presence of structural differences ofspecific proteins in vivo. Hence, the success of such testing is critically dependent on thepreservation of these proteins in samples. Factors such as age, size, and storage condition(environment) of the biological specimen influences the preservation of the original structure [5].Tissue-specific gene expression will also determine the presence of certain proteins thereby limitingthe types of specimen suitable for analysis. Furthermore, protein and other cellular componentsfrom bacterial contamination of samples [6], mixed body fluid samples [7], or samples originatingfrom more than one individual may make interpretation difficult or even give inaccurate results.Even in the absence of these problems, while exclusion is absolute, a conclusive match isalmost impossible. The limited number of measurable variations in the majority of these proteinsystems means that a large percentage of the population will share any one specific protein variant.2In most instances, the “inclusion” of an individual (positive result) may only be an indication thatthe specimen is consistent with a possible origin from that individual [8] and not positive proof ofassociation. In the last five to six years, extensive efforts have been made to remedy the problemsof uncertain identification and to increase the discriminative potential of forensic specimen testingby utilizing the rapid advances in the analysis of the cellular genetic material otherwise identifiedstructurally as deoxyribonucleic acid (DNA) [9-18].Genetic Variation In The Human Genome:The study of individual genetic variation at the genotypic level as opposed to thephenotypic level has been made possible by the introduction of recombinant DNA technology. Thedirect analysis of DNA not only allows for the determination of specific base sequences responsiblefor altered proteins, but also permits the detection of sequence variation in non-expressed DNA[19]. The human genome, which is 6 X io nucleotides per diploid genome, consists of bothcoding and non-coding regions. Coding regions function in determining the composition andsequence of amino acids making up proteins [20]. The non-coding regions are DNA sequenceswhich either have regulatory roles such as controlling the degree of gene expression in specifictissues, or have no known function.It is believed that only a small percentage of the genome serves a coding function so thatthe majority of the non-coding sequences may be in fact non-functional [21]. However, thesignificance of most of the sequences will probably remain unclear until the Human GenomeProject has been completed. It is in these non-coding sequences that the potential for sequencevariation is the greatest since such sequences would not be under the influence of evolutionaryconstraints placed on coding sequences [20].Due to selection pressures and genetic drift, a 400 amino acid protein changes onlyapproximately once every 200 thousand years [22]. In electrophoretic polymorphism studies of 71different types of proteins in European populations, the frequency of polymorphic loci found was0.28 [23]. Based on this value, Botstein et al. have estimated that the frequency of DNA sequence3polymorphism for protein-coding sequences to be about 0.001 per base pair [24]. However,analysis of polymorphism at the level of the protein does not provide an adequate measurement ofvariation of the total genome. Compared to coding regions, the extent of sequence polymorphismis far greater in the human genome as a whole (both coding and non-coding regions) [25]. Cooperet al. calculated that unique DNA sequence heterozygosity in human autosomes is 0.0039 whichsuggests that the majority of the sequence variation occurs in non-coding regions at a frequency ofabout one every two to three hundred base pairs [25].These findings support the conclusions of Alec Jeffreys in his analysis of the B-relatedglobin gene complex in man [26]. Three DNA sequence variants were detected within this complexof genes and surprisingly, all three variants were found in the intervening (non-protein coding)sequences which made up only 3.4 kilobases out of the total of 41 kilobases. As it seemed unlikelythat this observation occurred purely by chance, Jeffreys suggested that there was a preferentialaccumulation of DNA sequence variation within the intervening sequences. This would indicatethat any possible role played by the intervening sequences would not require fixed sequences.Jeifreys further suggested that if the amount of polymorphism associated with the B-related globingene complex is indicative of the degree of variation found in the rest of the genome, then thefrequency of variation for the human genome would be about one in a hundred base pairs.However, the results of a more recent study suggests that sequence variation in codingsequences contribute significantly to sequence polymorphism of the entire genome [27].Comparison of protein-coding sequences between human and rodent homologous genes indicatethat the rate of nucleotide substitution is highly variable with regards to 1) type of substitution(synonymous and nonsynonymous), 2) different gene regions and 3) different genes [27]. In thisstudy of 17 protein coding sequences the mean rate of synonymous substitution (no change inamino acid) was found to be five fold higher than that of nonsynonymous substitutions (change inamino acid). This is a consequence of selection against any nonsynonymous mutations which causedeleterious effects on the function of a protein whereas synonymous mutations are tolerated.4Different regions within a protein (coded by different gene region) have various functions such thatsome regions are more constraint than others leading to differing substitution rates in the codingsequence (i.e. substitution rate is seven times higher in the C-peptide of the proinsulin genecompared to the A- and B- chain which make up the active portion of the hormone) [27].Synonymous and nonsynonymous substitution rate variation also occurs among genes. Thisdifference may be due to differing functional constraints or differences in the rate of mutation atdifferent loci and the differences in the chromosomal position of the gene [27]. In a comparisonbetween cows and goats, fourfold degenerate sites within the B- and a-globin gene had nucleotidesubstitution rates comparable to that observed with regions not subject to functional constraintssuch as introns, 3’flanking sequences and B-globin pseudogenes [27]. In light of these results, thefrequency of sequence heterozygosity in coding sequences may be much higher than previouslyassumed from proteins studies and limited coding sequence studies.Restriction Fragment Length Polymorphisms (RFLPs):The method used by some of the above mentioned researchers such as Cooper et al. [25]and Jeifreys [26], for their studies of DNA sequence polymorphism is based on the utilization ofbacterial restriction endonucleases. There are different types of restriction endonucleases and themost commonly used type is type II restriction endonuclease. These enzymes recognize specificbase sequences and will cut DNA only at a specific site either within or adjacent to that site [28].Type I restriction endonucleases bind DNA at specific sequences but cleave DNA at random siteswhen the DNA loops back to the bound enzyme [29]. These type I restriction endonucleases arenot applicable to VNTR analysis and will not be considered here.Restriction endonuclease digestion of genomic DNA results in multiple DNA fragments ofvarious sizes. The sizes generated are dependent upon the distances separating any two basesequences recognized by the specific endonuclease. The resulting fragments are then organizedaccording to size by agarose gel electrophoresis and detection of specific sequences (fragments) ispossible using Southern blotting [30] or variations of this blotting procedure. This involves the5transfer of DNA fragments from agarose to a solid support membrane such as nitrocellulose ornylon which is much more easily manipulated than agarose, followed by hybridization with specificnucleic acid probes. The discovery that a DNA probe for one region of the genome detectedrestriction fragments of different sizes in different individuals resulted in the exploitation ofrestriction fragment length polymorphisms or “RFLPs” in molecular biology.Restriction fragment length polymorphisms are due to changes in DNA which alter thedistance separating two restriction endonuclease recognition sites. This can be accomplished bypoint mutations leading to changes in base sequences which either create or eliminate a restrictionendonuclease recognition site (RES) [19]. Point mutations in DNA are responsible for a numberof inherited disorders. While these genetic changes within some RES are deleterious, changes inother RES are apparently harmless to the affected individual. These non-deleterious mutations areinherited in a Mendelian fashion and occur at high frequencies.RFLPs have become valuable tools as markers for genetic mapping, and for genetic diseaseresearch [31]. They have been used to determine the approximate location of genes responsible fordiseases such as Huntington disease [32], Alzheimer’s disease [33], Duchenne muscular dystrophyf34] and cystic fibrosis [35]. By following the inheritance of linked markers, predictive tests areavailable to determine individuals at risk for certain genetic diseases. Analysis of this kind hasmade possible the prenatal diagnosis of several inherited disorders such as sickle cell anemia[36,37], phenylketonuria [38], Duchenne muscular dystrophy [39] and Hemophilia A [40].Inherited alterations in restriction fragment size can also be a result of insertion of DNAinto an area or deletion of DNA from an area bracketed by two sequences recognized by the sameendonuclease. This type of polymorphism was first detected on chromosome 14 by Wyman andWhite who used a segment of single-copy human DNA as a probe [41]. This locus exhibited 8alleles and family studies indicate that these alleles are inherited in a Mendelian fashion. Similarpolymorphisms detectable with single copy (or single locus) probes have since been found neargenes such as the insulin [42,43], and alpha globin [44,45] loci. These loci are hypervariable due to6the presence of core sequences tandemly repeated a variable number of times. Suchpolymorphisms have been termed “variable number of tandem repeats” or “VNTRs”. VNTRs arevery informative for genetic analysis because each locus has the potential for a large number ofalleles resulting from differences in the number of tandem repeats [41]. These differences arepresumably due to recombination by DNA slippage during replication or by unequal exchangesduring mitosis or meiosis [9].Before proceeding further, it is necessary to discuss the use of “allele” in the context ofrestriction fragment length polymorphisms (RFLPs). Historically, “allele” refers to an alternateform of a gene which serves some coding or regulatory function [46]. Others have argued that inmodern times, the term “gene” is used to describe any heritable unit of the genome regardless ofcoding potential and since restriction DNA fragments are inherited as codominant traits accordingto Mendel’s laws, the term “allele” should be defined as alternate forms of any identifiable DNAvariation in the genome [47,48]. In light of the fact that published material from the forensic DNAtyping community utilizes the term “allele” for RFLPs, this author will also adopt this term for thepurposes of consistency.In contrast to the VNTRs described above which exist as unique sequences, there are alsoVNTRs found in multiple copies throughout the human genome. These VNTRs or “minisatellites”were first described by Jeffreys et al. who utilized a 33 base pair sequence from an intron of thehuman myoglobin gene [9]. A tandemly repetitive probe (probe 33.15) made with this 33 base paircore sequence was able to simultaneously detect many variable regions within the human genome.The resultant autoradiograph exhibited numerous bands and the banding pattern was highlyvariable from individual to individual. Band comparisons between pairs of randomly selectedindividuals showed minimal band sharing [10].These minisatellite banding patterns were highly specific to an individual such that theprobability of finding the same pattern, using probe 33.15, in a second unrelated individual was 3 X1O or 1 in 30 trillion [10]. Further analysis with other such minisatellite probes demonstrated7that a combination of two of these probes would be almost totally individual specific. The onlyexception to this specificity is between monozygotic twins in which case the DNA patterns areidentical. The high level of discrimination of the minisatellite probes led to the term “DNAfingerprinting” which is a trade name registered by the British company Imperial ChemicalIndustries (ICI).All fragments from an individual can be traced to one or the other parent withapproximately half of the fragments being of paternal origin and the remaining fragments ofmaternal origin. This is due to the fact that these hypervariable fragments are stably inherited andsegregate in a Mendelian fashion [9]. This observation led to the eventual application of suchmulti-locus (minisatellite) probes to create individual-specific “fingerprints” for humanidentification, including paternity determinations.The first official use of DNA genetic typing for human identification was in a 1985immigration case [49]. The objective was to determine the maternity of a boy who was attemptingto re-enter the United Kingdom on the premise that his mother was already a resident of theUnited Kingdom. Immigration authorities suspected that a switch had occurred and that the boywas in fact the nephew of the woman involved. Conventional blood testing could not differentiatebetween the two possible relationships. DNA testing with two of the Jeifreys’ minisatellite probesshowed beyond a reasonable doubt that the relationship of the boy and woman was one of motherand son.Human identification for immigration purposes is only one application for DNA genetictyping. Forensic identification of unidentified bodies or of crime-associated specimens can also begreatly facilitated and improved with this advanced technology. Initial studies to assess thefeasibility of the application of DNA genetic typing to forensic specimens were performed at theHome Office Forensic Science Service Central Research Establishment in the U.K. [11]. Theresults of these studies confirmed the potential of the procedure. DNA prints were obtained fromspecimens such as semen stains, four year old blood stains, and fresh hair roots. This was further8substantiated by a subsequent evaluation of a blind trial involving 43 blood samples, 11 bloodstainsand 11 semen stains 1161. These promising results eventually lead to the introduction of DNAprofiling to its first criminal case in 1987 [50,51].This historic case not only identified the individual responsible for two rape/murder casesbut it also conclusively excluded another individual of the same crimes. A young man was chargedwith the 1986 sexual assault and murder of a school girl in Leicestershire, England. A similarunsolved murder had occurred in 1983 and conventional testing of semen stains from both victimsdid not rule out the possibility that this youth was also responsible for the earlier murder.However, the results of these tests showed that the semen could have been deposited by at least10% of the male population. Hence, DNA genetic typing of the semen stains and blood from theaccused male was performed, initially by Dr Alec Jeffreys and subsequently repeated by HomeOffice forensic scientists. These results excluded the accused but did confirm that both crimes hadbeen committed by one individual with a chance association of the different seminal stain DNApatterns of only 5.8 X io8 or 1 in 580 billion. The accused male was released and a intensiveblood testing program was undertaken by the Leicestershire Constabulary.More than 5000 males were analyzed with an initial screen by conventional forensicserological tests which eliminated approximately 86% of the population. DNA typing was thenperformed on the remaining 14%. It was discovered that one individual, Cohn Pitchfork, had paida workmate to provide the police with a blood sample on his behalf. When confronted by thePolice, Pitchfork confessed to both of the rape/murders in question. Subsequent DNA testing ofPitchfork’s blood showed a pattern identical to that of semen stains. The successful conviction ofPitchfork was only possible with the help of DNA fingerprinting which excluded a wrongly accusedindividual and in turn identified the real culprit.This case clearly illustrated the potential power of DNA typing as an investigative tool.The Home Office Forensic Service and the British criminal courts have since approved the use ofDNA fingerprinting in forensic cases [52]. The British civil courts have also officially approved9DNA fingerprinting for paternity disputes and as of 1989, the British immigration authorities wereconsidering the routine use of DNA typing for immigration applicants [53].Single Locus VNTR Probes Verse Multi-locus Minisatellite Probes:Although DNA genetic typing originally started with the Jeffreys minisatellite probes, themajority of the North American laboratories are using only the single locus probes. In fact, theRFLPs (minisatellite regions) detected by the minisatellites probes, have since been isolatedindividually to create probes which detect only one minisatellite region at a time and thus serve assingle locus probes [54,55]. The reason for the switch in strategy was the fact that a number ofcomplications involving minisatellite probes had been encountered.Unfortunately, the multi-banding pattern which affords a minisatellite probe its tremendousdiscriminatory power is also one of its downfalls because of the complexity of the numerous DNAbands. These probes simultaneously detect many loci and the recombination rate at these regionsis high [9,56] such that the number and/or sizes of loci detected are not constant from oneindividual to the next. Since there are no expected patterns, there may be some difficulties indetermining which bands are allelic [14]. Extra bands may appear due to partial or incompletedigestion of DNA or bands may be missing due to inadequate amounts of DNA [16]. These errorscould go undetected and lead to false positives or false negatives.Difficulties with interpretation of DNA prints will also occur if the DNA source is a mixedsample i.e. contributed by more than one individual [57]. In fact, contamination of a sample withnon human blood from sources such as birds, dogs, cats and mice, will also produce bandingpatterns specific to the contaminant [58-61].Another drawback is the absence of chromosomal localization of the minisatellites [14].The highly polymorphic nature and complexity of the minisatellites bands hinders attempts atlocalization of such bands [62] to specific areas within specific chromosomes. Difficulties arisewhen attempts are made to determine which band in each of two unrelated individuals actuallyrepresent the same locus. Finally, the size of the alleles detected by minisatellite probes range10from 4 kb to 20 kb [57]. Since the likelihood of obtaining intact full length DNA from old anddegraded samples is low, the large fragments may be missing leading to incomplete DNA prints.In contrast, single locus VNTR probes detect RFLPs in only one defined region and thenumber of alleles or bands is known before the analysis. As the VNTR regions chosen are oneswhich follow Mendelian segregation, each locus will have two alleles with one allele of paternalorigin and the other of maternal origin. If both alleles are identical or have similar sizes, such thattheir difference is beyond the resolution of the electrophoretic system, then only one band will bevisible on the autoradiograph. The presence of more than two alleles for any one of these singlelocus probes is indicative of some methodological error such as incomplete digestion of the DNAor contamination of the DNA sample. Recently however, there has been a report of individualswith three band patterns when probed with a single locus VNTR probe [63]. The three loci so farshown to demonstrate this phenomenon resulting from restriction site polymorphisms are D3S46,D4S139, and D7S22. The nomenclature for DNA segments are as follows:- “D” refers to DNA- number after “D” refers to the chromosomal assignment- “S” refers to unique DNA segment- numbers after US” refers to different sites on the same chromosome In the case of locusD4S139, the abnormal pattern is due to the presence of an internal Hae III site within the VNTRflanked by the two expected Hae III sites. This results in the production of two bands for oneallele such that a normal (2 Hae III sites) allele and an abnormal (3 Hae III sites) allele produce athree band pattern. Nevertheless, the concept of having a predetermined number of bands servesas a primary screen for errors which could lead to inaccurate interpretations. Since one singlelocus probe only detects differences in one region of the genome it does not have thediscriminating power of one minisatellite probe.11In order to approach the level of discrimination afforded by the multi-locus probes, anumber of non-linked single locus probes must be used, essentially recreating the multiple bandingpattern of the minisatellites. However in this case, every band has a known origin.The frequencies of alleles from the different loci can be used together to calculate thechance occurrence of such a DNA banding pattern. The simplicity of the banding patternsobtained with the single locus probes also allows linkage analysis studies to ensure that allelefrequencies of one locus are not influenced by the frequencies of other hypervariable loci used inthe identification procedure. Another advantage of using single locus probes is that they can detecta smaller amount of DNA than that of the multi-locus minisatellite probes which require onemicrogram of genomic DNA in order to produce a DNA fingerprint [57]. This fact is especiallyimportant since the quantity of forensic samples is often very limited.Application of DNA Genetic Typing To Forensic Samples:Single locus probes have been used in forensic case work in North America by both lawenforcement agencies and private commercial companies. Such services for both forensic andpaternity testing have been made available since 1987 by the first two private DNA typinglaboratories in the United States, Lifecodes Corporation (Valhalla, NY) and Cellmark Diagnostics(Germantown, MD). These two companies have reported tremendous success with this newtechnology. The first conviction using DNA evidence in the United States was prosecuted inOrlando, Florida in 1987 [64]. Since then, there have been numerous criminal trials in which DNAprofiling has been admitted as evidence [65,66]. The United States Federal Bureau of Investigationhas aggressively developed forensic DNA analysis techniques and have imparted their knowledge tostate and local crime laboratories in the U.S. [67]. As well as on going research, the FBI arepresently providing a forensic DNA typing service for criminal case materials.Forensic laboratories in Canada including the Royal Canadian Mounted Police CentralForensic L.aboratory in Ottawa, and the Centre of Forensic Sciences in Toronto have also putextensive efforts into this revolutionary technique [8,68]. Canada’s first trial experience with DNA12evidence in December 1988 resulted in the exoneration of an accused, James Alexander Parent,who was charged with a number of related sexual offenses [69]. DNA evidence indicated thatParent was not the assailant in several of the offenses that he had been charged with. In April1989, DNA evidence was used in court to positively identify an individual as the assailant in thesexual assault of a 68 year-old woman [8,69]. Comparison of DNA extracted from semen stainsfound on the victim’s nightgown and bedspread with that from the accused’s blood sample, showedidentical DNA banding patterns. The judge presiding over the “voir dire” in the District Court ofOttawa, ruled that the DNA evidence was admissible. RCMP experts testified that the chanceoccurrence of the DNA print from these forensic samples matching those of the accused was foundto be less than 1 in 70 billion people. Faced with such overwhelming odds, the accused PaulMcNally, plead guilty to the charges.Controversies Surrounding DNA Genetic Typing:In many of the earlier cases, DNA evidence was not challenged in the courts becausedefense attorneys had difficulty finding scientists to question the accuracy of DNA typing [70].However, in 1989 the scientific community started to question the accuracy and validity of DNAanalyses performed by the various DNA typing laboratories in North America. In addition toseveral other cases involving questionable DNA evidence, the highly publicized New York murdercase of People v. Castro in August 1989, proved that these concerns were justified [71,72]. TheDNA analysis of the Castro case was performed by Lifecodes which is the largest private forensicDNA laboratory in the United States. Defence expert witness Eric Lander, from the MassachusettsInstitute of Technology in Cambridge, along with the two experts for the prosecution, signed astatement declaring that the DNA evidence presented by Lifecodes was not “scientifically reliableenough” to support its conclusions which incriminated the accused [71,73]. Based on thisdeclaration, the presiding Judge Gerald Sheindlin ruled that the DNA evidence was not admissiblein this case. The issues raised in the Castro case by Eric Landers forced both the forensic andscientific community to examine more critically the entire concept and technology of DNA analysis.13Although DNA genetic typing with single locus VNTR probes has made a major impact inforensic science as well as in other disciplines concerned with individual identification, severalissues still need to be dealt with by the “experts” in this field [72,74,75]. One of the major issues isthe unpredictability of the sizes of the DNA fragments after restriction endonuclease digestion. Ahypervariable (VNTR) locus may have 50 to 100 alleles such that it is impossible to predict thesizes of the DNA fragments [71]. DNA extracted from forensic samples are frequently of poorquality (j degraded) thus eliminating the detection of the large alleles which require the presenceof intact DNA molecules. Individuals heterozygous for a probed VNTR locus would beerroneously typed as being homozygous for a small allele. The poor quality of the forensic samplesmay also result in band shifting to such an extent that false negatives may occur. Furthermore,VNTR alleles may differ in size by as little as one base [76]. Assignment of a particular band to adiscrete allele would be difficult if the sizes of bands differ by only a few base pairs (j 1000 bp+1- 50 bp may have two possible allelic assignments). For DNA fragments >1 kb, slight changes inDNA length may not be resolved by agarose gel electrophoresis. Both of these problems interferewith the acquisition of reliable quality data and accurate interpretation of results.Another important issue revolves around the need for implementation of standard qualitycontrol and quality assurance procedures by various DNA laboratories [75,77]. The problem in thisarea is that different laboratories, especially the commercial ones, tend to keep their work a secretand only reluctantly reveal their procedures when forced to by the courts [74]. In addition, thereare as yet no government or scientific agency-enforced proficiency testing programs to accreditforensic laboratories for DNA analysis [70]. However, within the last two years, the FBI havepublished several articles concerned with issues of quality assurance, proficiency testing andstatistical standards for DNA analysis [78-80]. Furthermore, members from several agencies acrossNorth America have recently formed a group called “TWGDAM” or Technical Working Group onDNA Analysis Methods. The purpose of TWGDAM is to set up guidelines for DNA typinglaboratories so as to promote consistency wherever possible at the different facilities. The14Canadian Society for Forensic Science has set up a similar group (subcommittee) for the samepurpose in Canada.At the present time, the United States Congress is considering the DNA ProficiencyTesting Act of 1991 [81]. This bill would promote a step towards the requirement of standardguidlines and regular proficiency testing in laboratories performing DNA forensic testing.The last major issue is one concerning the calculation of the probabilities of chance“matches” between DNA restriction patterns from non-related individuals. The highly publicizedclaims of astronomical odds against “matches” arising at random such as 67,500,000,000:1 odds inGeorgia v. Caidwell [71], and 5,000,000,000:1 odds in Florida v. Andrews [82], are undoubtedlyimpressive to trial juries. However, the statistical basis for such calculations may not be applicablefor some VNTR loci if the reference population used to build the data base consists ofheterogenous subpopulations or if the combination of loci used are in linkage disequilibrium [70-72]. The method currently used for probability calculations is based on The Hardy-WeinbergEquilibrium which is actually an extension of Mendel’s laws of inheritance [83].The Hardy-Weinberg Equilibrium (which gave rise to the Hardy-Weinberg Law) wasproposed in 1908 independently by G.H Hardy in England and by Wilhelm Weinberg in Germany[84]. This equilibrium describes “the expected relationship between the frequencies of alleles inlocal populations and the frequencies of individuals of various genotypes in these same populations”[83]. In the context of RFLPs, the Hardy-Weinberg Equilibrium relates the frequency of thevarious bands (alleles) in a population to the frequency of individuals possessing pairs of bands(genotypes) in a population [73].The Hardy-Weinberg Law states that “under certain conditions of stability both allelicfrequencies and genotypic ratios remain constant from generation to generation in sexuallyreproducing populations” [85]. Hence, using the Hardy-Weinberg formula, the probability of aperson having a specific banding pattern can be computed once the frequencies of the individualalleles have been established. In order to increase the discriminating power of single locus probes,15the genotype frequencies of each loci probed can be multiplied together to give the probability ofconcurrent incidence. This “product rule” [86] is only valid if the combination of loci analyzed areunlinked (at linkage equilibrium) the presence of certain alleles at one locus has no influenceon the alleles at another locus [71].The controversy surrounding the application of the Hardy-Weinberg formula stems fromthe fact that with respect to certain loci, mixed populations may not be in the Hardy-WeinbergEquilibrium. This means that allele frequencies may differ from one ethnic group to another suchthat population data bases developed with one population may be inaccurate for anotherpopulation. To explain this discrepancy, one must realise that in a continuously evolving world, thefour conditions necessary to achieve the Hardy-Weinberg Equilibrium ie. a large population size,no mutations or mutational equilibrium, no immigration or emigration and random reproduction,are never met [85]. Although appreciable changes in allele frequencies due to violations of thesefour conditions may occur quickly with respect to the evolutionary time scale, it is slow with respectto real time. Therefore, it is still possible to observe populations in Hardy-Weinberg Equilibriumdespite breaches of the “specific conditions”. However, due to the existence of small and especiallyisolated communities in which there is inbreeding and/or reduced gene flow (no addition of newalleles by inter-racial matings), there would be an increase in the number of individualshomozygous at various loci because the alleles they carry are identical by descent or parentage [83].As a result, individual communities may have different gene frequencies for various alleles.Evidence for allelic heterogeneity between American blacks, Caucasians, and Hispanics isavailable for at least four loci [87,88]. Even within these racial subpopulations there may be amixture of ethnic subgroups leading to deviations from the Hardy-Weinberg Equilibrium. Thisdeviation was observed by Eric Lander in his analysis of the Hispanic data base used as thereference population in the problematic Castro case [71,73]. “Subgroups” may by themselves be inHardy-Weinberg Equilibrium but the pooling of subgroup data before the calculation of the HardyWeinberg expectations could result in a divergence from the equilibrium (Wahiund’s Principle)16[83]. If the total population data base does not take in to account heterogenous subpopulations,there could be an under-estimation of the frequencies of some alleles. The error is furthercomplicated by the possibility of disequilibrium between loci since loci in heterogenous populationsmay not be in equilibrium even if the loci are on separate chromosomes [71]. This could result inestimated probabilities of matched DNA profiles being lower than the true probabilities and henceunjustly more incriminating to the suspect [89]. It is clear therefore that allele frequency databases for each subpopulation are required to avoid erroneous results.When dealing with RFLPs (VNTRs), another avenue for error leading to non Hardy-Weinberg expectations is the limit of resolution of bands in agarose gel electrophoresis. Asmentioned previously, VNTR regions have a continuum of allele sizes and slight differences maynot be resolved so that two similar sized alleles may appear as one allele. This would lead to asimilar over representation of individuals homozygous for certain loci and cause a populationwhich is in Hardy-Weinberg Equilibrium for the loci studied, to appear to not be in equilibrium.All the issues related to the use of RFLPs in forensic identification discussed here are themajor problems associated with this technology and are not by any means trivial. Due to thetremendous potential of this technology, numerous laboratories both government and commercial,are actively in search of new strategies and methods to overcome some of the existing problems asevidenced by the vast amount of literature concerning DNA genetic typing. The recent discovery ofthe polymerase chain reaction [90-92] has provided more avenues to deal with the identification offorensic specimens.17THESIS HYPOTHESISForensic identification is a critical component of any criminal investigation. Informationthat associates an object or a person to a specific crime may be used to corroborate otherevidentiary material or may provide the law enforcement agencies with a “starting point” fromwhich the investigation proceeds. Unfortunately, with the exception of fingerprints, a forensic“match”, using conventional techniques, of an individual to a biological specimen(s) found at acrime scene, will in most cases only indicate that the individual is a possible contributor of thespecimen. The biological components, proteins, tested are not polymorphic or variable enoughwithin the population to provide sufficient discrimination to distinguish one individual from anotherat a statistically significant level. In order to observe greater variation, the analysis must proceed toa more fundamental level, i.e. the basic heritable material of the cells (deoxyribonucleic acid orDNA) some of which codes for the proteins and some of which is phenotypically silent.In 1985, with the use of DNA probes, Dr. Alec Jeifreys of the University of Leicester(England) reported the detection of tandem-repetitive “minisatellite” regions dispersed throughoutthe human genome [9]. These minisatellites (multi-locus variable number of tandem repeats) werehighly polymorphic and were capable of creating individual-specific DNA “fingerprints”. The DNAbanding pattern obtained by analysis with two such minisatellite probes was specific such that theprobability of two individuals with the same pattern was <<5 x io19 [10]. These impressiveresults caught the immediate attention of the forensic community and the media. It appeared thatthe opportunity had arrived for the positive identification of criminals by analysis of specimens suchas hair bulbs, blood, semen and stains of blood or semen.Although the DNA typing (fingerprinting) procedure described byJeifreys utilizes a combination of techniques that have been already established in the scientific andmedical community, it cannot be assumed that the reliability of these techniques is good enough tomake life and death judgements on a suspect. There must be extensive research into many areas18such as DNA polymorphism, effects of environmental conditions on DNA integrity particularlyduring the subsequent typing procedure, and the reliability of the techniques (reproducibility andaccuracy).The aim of this thesis is to determine the extent to which hypervariable regions, specificallythe areas containing the single locus variable number of tandem repeats (VNTRs), within thehuman genome can be used as a means of distinguishing between individuals in a forensic capacity.This investigation also focuses on the validity of data collected by non-conventional populationsampling procedures (i.e. from undefined population). In addition, another DNA typing systeminvolving dimorphic restriction endonuclease recognition sites is offered as an alternative for theexisting hypervariable VNTR (single locus and multi-loci) systems that may solve some the obviousproblems of the present VNTR technology such as questionable allelic assignments and inability todetermine allele frequency equilibrium by the Hardy-Weinberg test.Objectives1. To build a population data base for four hypervariable VNTR regionsutilizing DNA samples from undefined individuals of the GreaterVancouver Area.2. To establish technical conditions for VNTR analysis of genomic DNA.3. To apply VNTR technology to forensic samples and to determine itsapplicability to non-laboratory samples.4. To compare the frequency distribution of VNTR alleles collected fromdefined and undefined populations.5. To set up another system of RFLP analysis utilizing dimorphicrestriction endonuclease recognition sites (RES) and the polymerasechain reaction (PCR).6. To build a population data base for several dimorphic RES utilizingDNA samples from undefined individuals and from Orientals in thepopulation of the Greater Vancouver Area.7. To test the discriminatory potential of this dimorphic RES system bythe analysis of two families and a number of undefined individuals.1920MATERIALSSolutions:1. Alkali transfer buffer - 1.5 M NaC10.25 M NaOH2. Denaturation buffer- 1.5 M NaC10.5 M NaOH3. Denhart’s (100X) - 2% ficoll2% polyvinylpyrrolidone2% bovine serum albumin4. HTE buffer - 0.100 M tris-base0.040 M EDTA(pH 8.0)5. LTE buffer - 0.010 M tris-HC10.001 M EDTA(pH 7.5)6. Luria-Bertani medium (LB)- 0.5% Bacto-yeast extract1.0% Bacto-tryptone1.0% NaC1(pH 7.5)7. Post hybridization washSolution I- 2X SSPE0.1% SDSSolution II- 1X SSPE0.1% SDSSolution III - 0.1X SSPE0.1% SDS8. SSC (1X) - 0.150 M NaCI0.015 M a36HO7(2H)(pH 7.0)222. Insulin 5’HVR (detects INS)- kindly provided by Dr. Graeme I. Bell,Department of Biochemistry and Biophysics, University of California3. YNH24 (detects D2S44) - kindly provided by Dr. Y. Nakamura, Howard HughesMedical Institute and Department of Human Genetics, University of Utah4. CMM1O1 (detects D14S13)- kindly provided by Dr. Y. Nakamura, Howard HughesMedical Institute and Department of Human Genetics, University of Utah5. HauplS - Adenine phosphoribosyltransferase (APRT) probe kindly provided byDr. Peter J. Stambrook, Department of Anatomy and Cell Biology, University ofCincinnati College of Medicine6. ADAN16 - Adenine deaminase (ADA) probe purchased from the American TypeCulture Collection (ATCC), catalogue # 57227.7. H25-3.8- Carbonic anhydrase II (CAH) probe purchased from the American TypeCulture Collection (ATCC), catalogue # 57263.8. HLPLG18 - Lipoprotein lipase (LPL) probe kindly provided byDr. S. S. Deeb, Department of Medicine, Division of Medical Genetics, University ofWashington23CHAPTER 1VNTR ANALYSISSection I - Data Base GenerationSection II - Forensic Case Samples24INTRODUCTIONPrior to 1985, DNA genetic typing for individual identification was virtually unknown to theforensic community. The first successful use of DNA fingerprinting in a criminal case, in theUnited Kingdom in 1987 [50], catapulted this revolutionary typing system into the public andforensic arena. The obvious potential of DNA genetic typing and the excitement fuelled by themedia, resulted in a great demand by law enforcement agencies for such a forensic service.However, as the system was still in its infancy at that time, a great deal of research anddevelopment was required to establish the reliability and validity of the typing results.At the onset of this thesis project in late 1987, most North American forensic laboratorieswere just beginning to focus more on single locus VNTR probes as opposed to minisatellitesprobes as developed by Jeifreys in the United Kingdom [9,10]. One of the major tasks was to buildpopulation data bases for the VNTR loci utilized in the typing procedure. Each laboratory begangenerating its own data base because of problems of consistency in allelic frequencies of data basescollected by sampling different geographic locations. It was not known whether allelic frequenciesgenerated by sampling one population would be applicable to another population due to variationsin subpopulation (L ethnic subgroups) frequencies. Another reason for the separate data baseswas that there was no unanimity in the choice of probes among all the forensic laboratories. Sincethe forensic application of DNA analysis was new, there was no consensus as to which particularprobes would be most appropriate.The complete absence of research into this novel field in British Columbia was the mainimpetus for this thesis project. With the dubious applicability of data bases generated elsewhere, itwas important to build data bases for the local population (Greater Vancouver Area). Anotherimportant consideration was to determine and identify problems which may make the promisingDNA genetic typing system (by VNTR) invalid and/or unreliable. Potential problems could beenvisioned with respect to:1) data bases2) variability of the VNTR loci3) technical procedures4) applicability to forensic specimens which is influenced by thecollection and storage of such specimens2526METHODS1. PREPARATION OF DNAIsolation of DNA From Liquid Blood (Protocol A):Deoxyribonucleic acid was extracted from approximately 6 ml of EDTA blood according tothe preferential cell lysis method (David Hoar - Childrens’ Hospital, Calgary Alberta: personalcommunication). Briefly, red blood cells were osmotically lysed with 5 volumes of a solutioncontaining 0.017 M of Tris-HC1 (pH 7.65) and 0.139 M of NH4CI. The remaining intact nucleatedcells were washed twice with saline and resuspended in 2 ml of HTE buffer (100 mM Tris-HC1, 40mM EDTA, pT-I 8.0). Two milliliters of WBC lysis buffer (100 mM Tris-HC1, 40 mM EDTA, pH8.0, 0.2% SDS) were immediately injected into the nucleated cell suspension to lyse the cellsinstantaneously. The DNA released was purified by two phenol/chioroform/isoamyl alcohol(25:24:1) extractions and one chloroform/isoamyl alcohol (24:1) extraction. The aqueous layercontaining the DNA was then precipitated with two volumes of 95% ethanol. The precipitatedDNA was pelleted and finally resuspended in LTE buffer (10 mM Tris-HC1, 1 mM EDTA, pH 7.5).Isolation of DNA From Liquid Blood (Protocol B):Deoxyribonucleic acid was extracted according to the protocol described by the FBI [93].Briefly, liquid blood samples were frozen at -70°C in 700 ul aliquots. A frozen sample was thawedat room temperature, mixed with 800 ul of lx SSC (150 mM NaC1, 15 mM sodium citrate, pH7.0), and centrifuged for one minute in a Eppendorf microcentrifuge. One milliliter of thesupernatant was discarded and replaced with one milliliter of lx SSC. This mixture wascentrifuged for one minute in the microcentrifuge after which all the supernatant was discarded.The pellet was resuspended with a 405 ul solution of 0.18 M sodium acetate, 0.62% SDS and 0.25mg/mI Proteinase K. This suspension was then incubated at 56°C for one hour, extracted oncewith 120 ul of phenol/chloroform/isoamyl alcohol (25:24:1), and once with 400 ulchloroform/isoamyl alcohol (24:1). The DNA in the aqueous layer was precipitated with one27milliliter of cold absolute ethanol, and centrifuged for 30 seconds before the supernatant wasdiscarded. The DNA pellet was resuspended in 180 ul of TE4, incubated at 56°C for 10 minutes,and reprecipitated with 0.2 M NaAcetate and 500 ul of cold absolute ethanol. The precipitate waspelleted with a 10 second centrifugation in the microcentrifuge, washed with one milliliter of roomtemperature 70% ethanol, and vacuum or air dried before it was resuspended in 100-200 ul TE4and incubated overnight in a 56°C water bath.Isolation of DNA From Forensic Samples (Protocol C):Deoxyribonucleic acid was extracted from bloodstains, hair shaft cells and tissue fragmentsaccording to the protocol described by Gill et al. [15]. Briefly, hair bulbs or bloodstains cut intosmall pieces, or chopped up tissue fragments were incubated on a shaker overnight at 37°C inextraction buffer (10 mM Tris-HC1, 10 mM EDTA, 100 mM NaCl (pH 8), 2% SDS, 100 ug/mlproteinase K, and 39 mM DTT). The sample was then extracted with one volume of liquifiedphenol and the aqueous layer was then extracted one volume chloroform/isoamyl alcohol (24:1).The aqueous layer was removed and treated with ethanol overnight at -20°C to precipitate DNA.After centrifugation, the DNA pellet was resuspended in LTE at room temperature.Purification of DNA Extracted From Stains:The resolubilized DNA was placed on a Millipore disc filter (0.05 uM) and dialysed for 2-4hours in LTE.Quantitation of DNA:Various concentrations of calf thymus DNA (Hoefer Scientific Instruments) were used asstandards for fluorometric quantification of linear DNA. A small volume (4 ul) of the DNAsample prepared by the methods described above, was diluted with (1000 ul) Hoechst 33258 (bisbenzimidazole) stain (0.1 uglml in TNE). The fluorometric density of this diluted sample wasobtained with a Hoeffer TKO 100 mini fluorometer.282. AMPLIFICATION AND PREPARATION OF DNA PROBESTransformation of Recombinant Plasmids into Competent E.Coli(DHSa) Cells:Recombinant plasmids were amplified by transformation into DH5a bacterial cells andsubsequent culturing of the transformed cells. The transformation procedure was performed asdescribed in the protocol provided with the E.coli cells purchased from BRL. Briefly, a 20 ulaliquot of competent cells was removed from the -70°C freezer and thawed on ice. Fivenanograms of recombinant plasmid was added to the tube of thawed cells and mixed gently with thepipette tip. The tube mixture was left on ice for 30 minutes and then heat shocked at 37°C for 20seconds. The tube was placed on ice and 0.95 ml of Luria-Bertani (LB) medium was added beforeit was incubated for one hour at 37°C with aeration (on top of a horizontal platform shaker movingat 200 rpm). The cells from this culture were then spread on to LBIX-gal agar plates containing aspecific antibiotic. The choice of the antibiotic used was dependent upon which antibiotic resistantgene was present within the recombinant plasmid used in the transformation. X-gal (5-bromo-4-chloro-3-indolyl-B-D-galactoside) is a chromogenic substrate of B-galactosidase and cleavage of thissubstrate results in the production of a blue color. The plates containing the transformed bacterialcells were incubated at 37°C until colonies appeared. Cells containing either the recombinant ornon-recombinant plasmids formed colonies on the antibiotic plates due to the presence of theantibiotic resistant gene contained in the plasmid. Recombinant plasmids which contained adefective B-galactosidase gene due to the insertion of the foreign DNA of interest, produced whitecolonies when transformed into these E.coli (DH5a) cells because X-gal is not cleaved to producethe blue color. These white colonies were selected for subsequent manipulations (jç plasmidisolation.29Small Scale Amplification and Isolation of Plasmid DNA:The small scale amplification of plasmid was performed as described in the Cold SpringHarbor manual [94] except for minor alterations. Bacterial cells containing recombinant plasmidswere inoculated into 5 ml of LB medium containing antibiotic corresponding to the antibioticresistance gene within the recombinant plasmid. The inoculated medium was incubated overnightat 37°C with aeration. A 1.5 ml volume of this culture was centrifuged at 13,000 xg for one minuteand the resulting pellet was resuspended in 100 ul of an ice-cold solution of 50 mM glucose, 10mM EDTA, 25 mM tris-HC1 (pH 8.0). This suspension was left at room temperature for 5minutes before adding 200 ul of an ice-cold solution of 0.2 N NaoH and 1.0% SDS. After mixing,the suspension was stored on ice for 5 minutes and then 150 ul of an ice-cold solution of 5 Mpotassium acetate (pH 4.8 was added. The resulting suspension was mixed by vortexing the tube inan inverted position for 10 seconds. After storing on ice for 5 minutes, the suspension wascentrifuged for 5 minutes at 13,000 xg to pellet out the precipitated cellular DNA and debris. Thesupernatant was then extracted twice with one volume of phenol/chloroform/isoamyl alcohol(25:24:1) and then ethanol precipitated at room temperature for 5 minutes. The pellet from thisprecipitation was washed once with 70% ethanol, dried in a vacuum desiccator, dissolved in LTEand treated with RNase to remove contaminating RNA.Excision of Foreign DNA Insert From Recombinant Plasmids:Recombinant plasmids dissolved in LTE were digested with the appropriate restrictionendonuclease to cleave at the sites immediately flanking the insert of interest. The digestconditions were those recommended by the suppliers of respective restriction endonucleases. DNAfragments in the digest were separated by electrophoresis through 0.8% low melting point agarosegel (containing 1 ug/mI ethidium bromide) in lx TBE or lx TAE buffer. The insert band asdetermined by size, was excised from the gel and placed in a 1.5 ml tube containing LTE. The tubewas placed in a 55°C heating block until the agarose melted (approximately 5 minutes). Theresulting suspension was extracted once with one volume of liquifled phenol, once with one volume30of phenol/chioroform/isoamyl alcohol (25:24:1), once with chioroform/isoamyl alcohol (24:1) andthen ethanol precipitated to purify and collect the insert DNA. The purified DNA was finallyresuspended in LTE.Labelling and Purffication of DNA Probes:The DNA probes were labelled using the random primers DNA labelling system kitpurchased from BRL. The kit protocol was followed except that the labelling reaction was allowedto proceed for 4-6 hours. Purification of the labelled probe was achieved with a Sephadex G-50column. The column was prepared by packing Sephadex G25-80 into a commercially availableQuik-Sep disposable chromatography column (Isolab Inc.). The packed column was washed twicewith LTE. The probe was then passed through the Sephadex column by centrifugation for threeminutes at 5000 xg to remove the unincorporated dNTPs. A portion of the eluted probe (2 u11300ul of eluent) was mixed with 10 ml of Aquasol (Dupont’s universal LSC cocktail) and analyzedusing a liquid scintillation counter (LKB 1217 RackBeta) to determine the specific activity of theprobes.3. VNTR ANALYSISRestriction Endonuclease Digestion of DNA Samples and Agarose Gel Electrophoresis:Five micrograms of DNA were digested with 12.5 - 30 units of either Pvu II or Hae III.Digestion conditions for these restriction endonucleases were those recommended by the supplier,BRL, using a digest incubation period of 4 hours. DNA digests were separated by electrophoresisthrough 1.2% agarose gel in lx TBE buffer (0.089 M of Tris-borate, 0.089 M boric acid, 0.025 MEDTA, p1-1 8.3). The progression of the DNA through the gel was detected with 0.04%bromophenol blue and 0.4% xylene cyanol. Electrophoresis was stopped when the xylene cyanoldye had migrated to 1 1/2 - 2 centimeters from the end of the gel (approximately 4 hours at 8031volts and 35 mAmps). The gel was then stained for 10 minutes in EtBr (1 ug/ml), soaked indistilled water for 10 minutes and viewed over an ultraviolet light box.Molecular length Standards and Controls:Molecular length standards or markers (BRL 1 kb ladder / Pharmacia 4X174 RF-Hae III)and a control DNA sample were included in every gel. (Initial experiments utilized lambda-HindIII I X174 RF-Hae III fragments as molecular length standards.) The molecular length standardlanes bracketed all the digested genomic DNA lanes eg. in a 14 lane gel, standards in lanes 1, 7 and14. The standard in lane 7 was included to correct for variations in the middle of the gel. One ortwo DNA sample with known VNTR patterns i.e. sample analyzed several times previously, wasalso included to serve as a control sample for the quality of the typing procedure. A minimum oftwo control DNA samples, one male and one female, were included in each gel during analysis ofcase material.Transfer of DNA From Gel to a Solid Support:A modified Ainersham alkali transfer protocol was used to transfer the electrophoreticallyseparated DNA fragments from the gel to a Hybond nylon membrane. Briefly, the agarose gel wassoaked in denaturation buffer (1.5 M NaCl and 0.5 M NaOH) for 30 minutes at room temperature.The gel was then placed in a capillary blotting apparatus with an alkali transfer buffer (1.5 M NaCIand 0.25 M NaOH) for 4-6 hours. The hybond membrane containing the transfer DNA waswashed in 5X SSC for 2-5 minutes and baked at 80°C for 2 hours to fix the DNA.Hybridization of32P-DNA probes to membrane-fixed DNA:The baked Hybond membrane was incubated with prehybridization solution (5X SSPE, 5XDenhardt’s solution, 1% SDS and 1% powdered non-fat milk) in a 65°C shaking water bath for 30-60 minutes. The DNA probe (prepared as previously described) was first denatured by heating in a100C water bath for 10 minutes, cooled on ice for 5 minutes and then added to theprehybridization solution containing the Hybond membrane. The concentration of probe in the32hybridization solution was 1.5 - 2.5 nglml. Hybridization was allowed to proceed in a 65°C waterbath (Blue M) for at least 12 hours.Post hybridization washes:Following hybridization, the membrane was washed twice with solution I (2X SSPE and0.1% SDS) at room temperature for 10 minutes (each time), once with solution II (1X SSPE and0.1% SDS) at 65°C for 15 minutes, and a final wash with solution III (0.1X SSPE and 0.1% SDS)at 65°C for 5 - 15 minutes. The incubation period of the final wash varied depending upon theprobe used in the hybridization mix. The decreasing salt concentration of each successive solutionand the increasing temperature of incubation, increased the stringency of hybridization. Thewashed membrane was wrapped in Saran-Wrap before it was exposed to Kodak X-OMAT AP orAR X-ray film at -70°C for various time periods.Probe removal from membrane for reprobing:In order to increase the level of discrimination, more than one VNTR region was analyzed.Hence a membrane was sequentially probed up to four times with different probes. Between eachreprobing, the filter was striped of the previous probe. Removal of the bound probe was achievedby an incubation at 45°C for 30 minutes in 0.4 M NaOH followed by another incubation at 45°Cfor 30 minutes in a solution of 0.1 X SSC, 0.1% (w/v) SDS, and 0.2 M Tris-HC1 pH 7.5.Determination of the Molecular length (Sizes) of DNA Fragments:The sizes of the genomic DNA restriction fragments were determined by comparing thedistance migrated by these DNA fragments from a fixed point (arbitrary origin) at the top of thegel, with the distances migrated from the same fixed point by various fragments of the molecularlength standards. Distance measurements were made to the middle of the band. The molecularlength standards have been sequenced by the commercial suppliers and hence are of known sizes.A line was generated by plotting the distance migrated by the molecular length standards againsttheir respective sizes. This line was then used to extrapolate the sizes of the genomic DNA33fragments according to their migration distances from the arbitrary origin. The measurements weremanually performed by a single person. The extrapolation was performed by a computer using aprogram model [95] based on Southern’s proposal that the reciprocal of DNA mobility plottedagainst fragment length is linear [96]. However, a correction factor was added to each measureddistance value for the genomic DNA fragments before they were entered into the computer. Thiswas to correct for the skewness of the molecular length standard lanes within the same gel (jidentical sized fragments at progressively higher or lower positions in the gel going from one laneto the next). This skewness is due to internal problems in the electrophoretic system such as,unequal temperatures in centre of the gel as opposed to the outer edges because of heating effectsof the electric current, and minute aberrations in the gel due to impurities in the agarose. Thecorrection factor was calculated by dividing the distance differences between two known identicalfragments in two flanking molecular length standards lanes by the number of lanes between thestandards lane plus one of the standard lanes.Example: Molecular length standard lanes- 1, 7 and 14Number of lanes between standard lanes 1 and 7 5Measured distance difference between 6.04 kb marker inlanes 1 and 7 = 1.0 mmTherefore, correction factor = 1.0 mm I (5 + 1) = 0.17 mmHence, lane 4 distance for allele approximately6 kb in size = measured distance + (number of lanesfrom first marker lane x 0.17 mm)= measured distance + (3 x 0.17 mm)Assessment of the Precision of the Typing System:A mixture of DNA samples from three different individuals was digested and separated byelectrophoresis as previously described for VNTR analysis. The DNA samples were chosen so thatwhen probed sequentially with two probes (3’HVRIYNH24), a number of bands (alleles) were34detected at regularly spaced intervals throughout the length of the gel. By having the DNA bandsat regular intervals spread from the top to the bottom of the gel, it was possible to get a range ofDNA fragment sizes (alleles) representative of the alleles of the four VNTR regions probed. Thisdigested DNA sample mixture was run in six lanes on the same gel and repeated in 15 gels.Molecular length standards (1 kb ladder / X174 RF-Hae III) lanes were also included in each ofthe 15 gels and were situated such that the molecular length standards and the genomic DNAmixtures were in alternating lanes (i.e. lanes 1,3,5,7,9,11,13- standards, lanes 2,4,6,8,10,12 -genomic DNA) (Figure 1). Lanes 1, 7 and 13 were used as standard lanes for the genomic DNAmixtures to simulate the conditions used in the regular typing procedure. Lanes 1, 7 and 13 wereused as standard lanes for the determination of the fragment sizes of the other standard lanes(lanes 3, 5, 9, and 11) which were in this situation treated as fragments of unknown sizes.Determination of DNA fragment sizes were performed as described in the previous section (page32-33).Within each group, i.e. the lanes containing genomic DNA (Figure 1 - even number lanes)or the molecular length standard lanes (Figure 1 - odd number lanes), the experimental sizevariation for identical fragments in the respective lanes within a gel and between gels were used tocalculate the confidence intervals for the specific size range. The confidence intervals for each sizerange were in turn used to create bins for allelic assignments.The rationale for analyzing the degree of size variation of identical fragments not only ingenomic DNA lanes (fragments not sequenced such that their true DNA size was unknown) butalso in the lanes containing the molecular length standards (commercially sequenced DNA ofknown size), was to establish whether there was a difference in variability between commerciallyobtained DNA and the more complex genomic DNA extracted in this laboratory.Assessment of the Accuracy of Experimental Size Determinations:The molecular length standard lanes (Figure 1 - odd numbered lanes) in the gels describedin the preceding section, were also used to assess the accuracy of the DNA typing procedure. As35Figure 1 : Assessment of Accuracy and Precision ofElectrophoretic SeparationProbe 3’HVR a-globin was used to detect DNA fragments digested with Hae III. Oddnumber lanes contain molecular length standard DNA (BRL 1 kb/4X174 RF-Hae III).Even number lanes contain Hae III digested genomic DNA from three different individualsin order to produce simultaneously a number of fragment of various sizes. Lanes 1, 7and 13 were used as molecular standard lanesi.e. lanes 1 and 7 were used as standards for lanes 2-6lane 7 and 13 were used as standards for lanes 8-12The molecular length standard DNA in lanes 3, 5, 9, and 11 were treated as unknowns sothat accuracy and precision measurement could be determined with both commercially boughtnon-genomic DNA and genomic DNA extracted from this laboratory.36Figure 1 : Assessment of Accuracy and Precision ofElectrophoretic Separation—— — — •1——1 234567891011121314—— — — — ——II —. — — — —37stated previously lanes 1, 7 and 13 were used as standard lanes and the other standard lanes (3, 5, 9and 11) were treated as fragments of unknown sizes. Comparison of the experimentally calculatedsize with the true sizes provided an indication of the accuracy of the typing system in thislaboratory.38SECTION IRESULTS AND DISCUSSIONPopulation Samples:The VNTR data in this thesis was collected from an undefined population. This populationsample consist of individuals admitted to The Vancouver General Hospital (live patients or morgueautopsy patients). Patients admitted to the hospital had blood drawn for routine tests and theleftover blood was used for DNA typing analysis. Blood drawn from morgue patients at autopsywere also used in this analysis. No other selection criteria was placed on the samples. Details arenot known about these samples as hospital policy does not allow questions such as ethnic or racialorigins to be included in the information sheets unless the blood samples were initially designatedfor research purposes.In 1987 when this project was being planned, the issue of collection and identification ofpopulation control samples was a matter of great concern to my supervisory committee. At thattime in Vancouver, there was a great deal of concern within the medical and biotechnology fields asto the issue of confidentiality of samples and the public press concerns had been raised about theaccessibility of results of any form of DNA genetic typing to police, governments and otheragencies. The supervisory committee decided that I would be allowed to use blood samples fromthe Haematology department which had been collected for other purposes, tested and were to bediscarded. I would be allowed access only to the names of the individuals for the determination ofracial origin provided that all subsequent information regarding the patients was destroyed and thatI would not be allowed to persue the issue of racial identification beyond this process. Access toautopsy material was limited by the fact that the B.C. Coroners Act specifically forbids the use ofCoroners’ autopsy material for research or teaching purposes. This has resulted in limited andconservative statistical interpretation of these population studies. However, within the context of a39forensic investigation, this type of limitation is more often the rule and as a result, these populationstudies are very relevant to practical forensic case investigations.Duplication of samples is not a significant problem in this sampling because samples fromlive patients who may provide blood more than once, were checked by patient names before therecords of the names were destroyed for patient confidentiality. Errors in checking could haveoccurred but should not be large enough to significantly alter the frequency values. Morguepatients have blood drawn only once and hence no sample could have been repeated.Technical Conditions For VNTR Analysis of Blood DNA:Two different methods for DNA isolation from liquid blood were utilized in this study.Protocol A was used for the majority of the DNA isolations for VNTR analysis because it:1) provided large quantities of high molecular length DNAsuitable for DNA typing2) was technically undemanding3) it had been verified by long use in this laboratoryHowever, protocol B was used to extract DNA from blood samples processed during the latter partof this study. The first method begins with a primary osmotic lysis of the red blood cells(anucleated), followed by lysis of the white blood cells (nucleated) with a detergent (SDS). Thesecond method is similar except that there is an initial freezing/thawing step before the osmoticlysis of the red blood cells. Other than the differences in methods of cell lysis between the twoprotocols, there is also a difference in the quantity of blood processed. Protocol B is designed toprocess a much smaller volume of whole blood: ie. 700 ul compared with the usual volume of 5-7ml by protocol A. The amount of DNA extracted from 700 ul of whole blood is usually sufficientfor 5 VNTR analyses. Excess DNA from the large scale extraction by protocol A is usually notrequired for the average VNTR analysis. Protocol B is a more efficient method to use since lessmaterials are required for this small scale extraction. In the event that more DNA is needed,additional blood stored frozen in 700 ul aliquots can be extracted.40Furthermore, since protocol B is used by the FBI laboratory and it seemed appropriate toadopt procedures developed by the FBI laboratory as they are the predominant group in theUnited States responsible for the development of forensic DNA analysis techniques.Standardization of forensic DNA analysis techniques would greatly facilitate efforts towards qualitycontrol and quality assurance. Data gathered using standardized techniques where possible will bemuch more valuable for comparative purposes.Technical conditions for preparation of DNA from forensic specimens:The methodology described for protocol C was the basic procedure followed. Sinceforensic specimens can come in a variety of forms, additional procedures were added or slightchanges were made to existing procedures where appropriate. These alterations will be discussedindividually for each case in section II of this chapter.Restriction Endonuclease Digestion of Genomic DNA:Originally restriction endonuclease Pvu II was used to cleave genomic DNA for VNTRanalysis. Subsequently, Hae III was utilized in place of Pvu II because it was decided that Hae IIIwould be more appropriate for VNTR analysis of forensic specimens [97]. This decision was basedon that fact that1) Hae III recognizes a four base sequence G-G-C-C and will cut DNAmore frequently than five or six base cutters such as Hinf I andPvu II. This leads to smaller fragments (alleles) which enablesgreater resolution of small size differences between fragments.2) Hae III is not sensitive to methylation which could interfere withcleavage of DNA.3) Hae III maintains its activity under adverse conditions such as thosefound in contaminated forensic samples.4) Hae III is inexpensive compared to other endonucleases.41The success of this entire DNA typing procedure is dependent on the degree of DNAdigestion by the restriction endonuclease Hae III. Incomplete digestion of DNA will result in morethan the expected two DNA bands detected by each VNTR probe. This complicates theinterpretation of the banding pattern seen on autoradiogram and may lead to inaccurateconclusions. A number of experiments were performed to establish the quantity of restrictionendonuclease and the length of the incubation time required for complete digestion of DNAobtained from blood. Restriction endonuclease concentration was varied from 2.5 - 6.0 units/ugDNA and the incubation time was varied from 2 hours to 20 hours (overnight). Completedigestion of DNA was obtained in every case even with the lowest concentration of restrictionendonuclease (2.5 units/ug DNA) incubated for the shortest time interval (2 hours) (Figure 2).Incomplete digestion would have been detected by the presence of multiple bands (more than 2)within each lane. Hence for subsequent digests, the incubation time was 4 hours with a restrictionendonuclease concentration of 2.5 units/ug DNA.Detection of Hypervariable Regions of DNA Using Multiple VNTR DNA32P-Probes:Hybond membranes bound to electrophoretically separated digested genomic DNA weresequentially probed with four single locus hypervariable VNTR DNA32P-probes. Membraneswere stripped between each successive probing as described in the methods (p44). The fourhypervariable VNTR probes were:1) YNH24 [98]2) 3’I-IVR/a-globin [44,45]3) phins 310 [42,4314) CMM1O1 [98]These probes were selected because they were freely available during the on-set of this study.General properties of the four probes are listed in Table 1. Each of these single locus probes42Figure 2: Conditions For DNA DigestionRestriction endonuclease concentrations of 2.5 - 6.0 units/ug DNA were incubated with5 ug genomic DNA for periods up to 16 hours (0/N). The electrophoretically separateddigested samples were hybridized with VNTR probes YNH24 (panel A) and phins 310 (panel B)to determine the degree of digestion of DNA samples.Lane 1 = 2.5 units/ug DNA2 = 4.5 units/ug DNA3 = 6.0 units/ug DNAFigure2:ConditionsforDNADigestion21w41w8hr0/N1231231231234.07)——— aaáaflfl9fla2.O3WW’wO.87>2hr4hr8hr0/N123123123123BATable 1 : General Pronerties of VNTR ProbesProbe Locus Size (bp) Core Repeat (bp) Identified byHBV-2YNH24 D2S44 2000 31 oligonucleotideYHVR Human a-globinD16S85 40’J() 17(c-g1obm) pseudogene (iiç1)phins 310 INS Human insulin(5HVR - insulin) (chromo 11) 879 14- 15 eDNAHuman Myoglobin-2CMM1O1 D14S 13 2200 15 oligonucleotide4445detected a maximum of two bands per lane (individual) representing the alleles for the particularlocus (Figures 3-6).Accuracy of the DNA Typing Procedure:The validity of the DNA typing procedure is critically dependent upon the precision andaccuracy of experimentally calculated allelic sizes. Precision (deviation from the mean sizedetermination) plays a major role in the derivation of confidence intervals and will be discussed inthe following section. The accuracy of this system was determined by comparison of theexperimentally obtained value with the true fragment size.Fragments of known sizes (molecular length standards) as determined by sequencing by thesupplier, were run in four lanes and repeated in 15 gels resulting in a total of 60 intra- and inter-gel comparisons for each standard size. The mean size determination for each standard wascompared to the true fragment size (Table 2). The greatest deviation was the 4.072 kb fragmentwhose mean size determination of 4.278 kb was 5.0% away from its actual size. However, a highdegree of accuracy was obtained with some of the other fragment sizes 1.018 kb and 0.872 kbfragments. There does not appear to be any correlation between fragment sizes and degree ofexperimental accuracy. Furthermore, the direction of errors is not consistent in that someestimates are above and some are below the true value. Other than measurement errors,inaccuracy of this typing system may be a possible result of the use of an inappropriate method forthe extrapolation of DNA size to migration distance from the origin of the gel [99].The method used in this analysis is one proposed by Schaffer and Sederoff [95] whichincorporates a full set of standards for the extrapolation of migration distance to DNA size. Elderand Southern proposed that only the four neighbouring standards rather than a whole range ofstandards, be used for extrapolation [99]. Factors such as heterogeneity in the gel and temperatureand voltage gradients across the gel will affect the mobility of DNA fragments leading to deviationsfrom the simple reciprocal relationship of migration distance and molecular size. Therefore,46Figure 3 : VNTR Analysis With Probe YNH24Genomic DNA from 11 individuals was probed with YNH24 which detects VNTRs at theD2S44 locus (lanes 1-11). Three molecular standard DNA lanes are present (lane M),bracketing the genomic DNA lanes.[6 days exposure to Kodak X-OMAT AP film at -70°C with screens]060)’ •1234 67891011I3.O6 I1.35a.YNH2447Figure 4 : VNTR Analysis With Probe 3’HVR a—GlobinGenomic DNA from 11 individuals was probed with 3’HVR a-globin which detects VNTRs atthe D16S85 locus (lanes 1-11). Three molecular standard DNA lanes are present (lane M),bracketing the genomic DNA lanes.[4 days exposure to Kodak X-MAT AR film at -70°C with screens]Ml 234 5M 6789IO6O)3.O61.35)j10 11 MS.Alpha -Globin48Figure 5 : VNTR Analysis With Probe phins 310Genomic DNA from 11 individuals was probed with phins 310 which detects VNTRs at theINS (insulin) locus (lanes 1-11). Three molecular standard DNA lanes are present(lane M), bracketing the genomic DNA lanes.[5 days exposure to Kodak X-MAT AP film at -70°C with screens]M12 3 4 5MG 7 8 91011 MI3.06)-•1wS0.60)-p31049Figure 6: VNTR Analysis With Probe CMM1O1Genomic DNA from 11 individuals was probed with CMM1O1 which detects VNTRs at theD14S13 locus (lanes 141). Three molecular standard DNA lanes are present (lane M),bracketing the genomic DNA lanes.[6 days exposure to Kodak X-MAT AP film at -70°C with screens]91011 M3.O6 •.CMM1O150Table 2: Accuracy of DNA Typing System*Actal Size Mean Deteimined Variation from(kb) Size (kb) actual size (%)6.106 5.992 1.95.096 5.23 1 2.64.072 4.278 5.03.064 3.187 4.02.030 2.052 1.11.636 1.606 1.81.353 1.318 2.61.078 1.052 2.41.018 1.011 0.70.872 0.864 0.90.603 0.621 3.0n=60*- Molecular weight standard DNA51according to Elder and Southern, incorporation of standards distant from the unknown to bemeasured (method of Schaffer and Sederoff) results in less accurate size determinations.The published computer program for the method of Schaffer and Sederoff was the onlyprogram readily available to this author during the period of data analysis. Hence the method ofSchaffer and Sederoff was utilized in the absence of a more suitable and accessible alternative.Precision of DNA Typing System:Assessment of the precision of the DNA typing system was accomplished by repetitiveexperiments with genomic DNA and molecular length standards. The experimental size variationof identical DNA fragments in different lanes within and between gels was determined by singlefactor analysis of variance, often abbreviated “ANOVA” [100]. The precision (deviation from themean size determination of identical samples) of a system which determines bin sizes, is distinctfrom measurement imprecision (maximum size difference between identical samples) which definesthe “match” criterion. Measurement imprecision will be discussed on pages 92-94.For comparable sized fragments, the variance and thus the corresponding standarddeviation was greater in the genomic DNA lanes than in molecular standards lanes (Table 3). Thismay suggest that the genomic DNA extracted in this laboratory is not as pure as commerciallyobtained molecular length standards. Impurities could affect the migration of the DNA fragmentsthrough the gel resulting in greater size variations. An alternate explanation may be the differencein amount of DNA electrophoresed through the gel. The molecular length standards lanes containnanogram amounts of DNA whereas the other lanes contain 2.5 micrograms of digested genomicDNA. DNA mobility is affected by the total mass in the sample loaded in the well [101].Increased mobility, band width and band distortion occurs with increased quantities (>0.2 g per0.15 cm2 of surface area in the loading well) of DNA.The magnitudes of variance and standard deviation are dependent upon the magnitude ofthe DNA fragment size [102]. Therefore, to compare the degree of precision for different sizedDNA fragments, the “coefficient of variation” or “relative standard deviation” was calculated52Table 3: Standard Deviation Comparisons of Molecular Weight Standardsand Genomic DNA Fragments (Inter/Intra Assay Variation)A B*V.J5 Variance Standard Deviation *Genon,jc Variance Standard Deviation2(kb) 2 a (kb) DNA (kb) a a (kb)5.992 0.00668 ±0.081 5.376 0.01519 ±0.1235.231 0.00400 ±0.064 4.333 0.00610 ±0.0784.278 0.00266 ± 0.052 3.96 1 0.00449 ± 0.0673.187 0.00121 ±0.035 3.159 0.00212 ±0.0462.052 0.00058 ± 0.024 2.610 0.00120 ± 0.0341.606 0.00030 ±0.017 1.560 0.00036 ±0.0191.318 0.00024 ± 0.016 1.490 0.00041 ± 0.0201.052 0.00010 ±0.010 1.270 0.00031 ±0.0171.011 0.00010 ±0.010 0.830 0.00007 ±0.0080.864 0.00006 ± 0.007 0.660 0.00006 ± 0.0070.62 1 0.00005 ± 0.007 n = 6 lanes x 15 gels = 90n = 4 lanes x 15 gels = 60MWS = molecular weight standard DNA*= mean determinations2a = sample variancea = sample standard deviation53(Table 4). Relative standard deviations were 1.4% in the molecular length standards group and2.3% for the genomic DNA fragment group. The values obtained for the latter group is ofgreater importance since genomic DNA samples are the unknowns in regular (actual) DNA typinganalysis. There appears to be a trend, in the genomic DNA group, towards a greater deviation withincreased DNA fragment size (Table 4 - panel B). This observation is not surprising sincemeasurement error is usually a function of restriction fragment length [103]. A measurement errorof a given distance results in a greater deviation in larger fragments than smaller fragments.Precision may be influenced by a single factor or a combination of factors such as gelimpurities, gel imperfections, and band measurement errors. Gel impurities should be minimal asBRL’s Ultra Pure agarose and distilled filtered water was utilized. However, slight variations in themigration of DNA in outer and middle lanes did occur in some gels as a possible result ofinconsistent heat dissipation. This curvature in the gel might have been eliminated with the use ofa continuously flowing gel buffer system.Distances of each band are determined from a fixed point to the centre of each band.Errors in band measurement may have affected the degree of precision since these measurementswere made manually by one individual. Measurement errors may occur particularly with thickbands which make the determination of the centre of the band difficult especially without the aid ofa densitometer, that is, by relying on visual inspection during band measurements. Despite thepossibility of error introduced by manual assessments, the frequency distribution of alleles analyzedwith the P131’s image analysis system was not significantly different (95% level) from thedistribution of the same alleles measured manually. The greatest discrepancy between the twoassessments was in bin 15 (alleles ranging in the size range of 1666-1795 bp) which had a frequencyvalue difference of 3.1%. The validity of the typing results will be further discussed in one of thefollowing sections (p94-9’7). This author would like to thank the FBI forensic laboratory for theuse of their image analysis system which has made this comparison possible.54Table 4: Relative Standard Deviation (Inter/Intra Assay Variation)A*W47.,i (kb) V (%)5.992 1.45.231 1.24.278 1.23.187 Li2.052 1.21.606 1.11.318 1.21.052 1.01.011 1.00.864 0.90.621 1.1n=60B*Genomjc V (%)DNA (kb)5.376 2.34.333 1.83.961 1.73.159 1.42.610 1.31.560 1.21.490 1.31.270 1.30.830 1.00.660 1.1n=90MWS = molecular weight standard DNA*= mean determinantsV = relative standard deviation = coefficient of variation= (a! mean) x 10055The RCMP laboratory has assessed the precision of their typing results by repeated analysisof a 2.73 1 kb fragment [47]. The relative standard deviation of the mean size determination (2.724kb) is 0.7%. Although typing results of the 2.610 kb genomic fragment from this study (Table 4 -panel B) exhibited a high degree of precision (1.3%), the RCMP laboratory’s results with the 2.724kb fragment is almost twice as precise. This discrepancy does not suggest that the procedures usedin this laboratory are inadequate but merely that larger confidence intervals are required for theallelic bin system. Further comparisons with other fragment sizes were not possible due to theabsence of published data.Determination of Bin Sizes Using Confidence IntervalsTo account for intra-gel and inter-gel variabilities and measurement imprecision, alleles ofthe HVR for each locus were assigned to a bin corresponding to the appropriate size range (Table5). Hence, several resolvable DNA fragments may contribute to the numbers generated for eachbin. Since each bin was then considered as one allele, its frequency would always be an overestimation of the true frequency for any one specific DNA fragment length.In order to set the arbitrary boundaries for each bin, it was necessary to determine theconfidence intervals for a range of DNA fragment sizes representative of the alleles of the fourVNTR regions analyzed. The standard deviations obtained from the repetitive experiments(precision studies) were used to establish these intervals.The mean determined sizes of the molecular length standards spanned a greater range ofsizes than the genomic DNA fragments 5.992-0.621 kb versus 5.376-0.660 kb. Since the aim ofthese experiments was to create a binning system to encompass a wide range of allelic sizes, themean molecular length standards sizes were used for confidence interval calculations. However,the variance of most similar sized genomic DNA fragment was used to derive standard deviationvalues of respective mean molecular length standards. It was necessary to use larger variances ofthe genomic DNA fragments because confidence intervals derived from these values were used toset up bins for genomic DNA alleles.=)L-JL’3t’J——\C00çj4J—00—acU-J‘C\D00-C(M-C)C———c,...—U(M44t-3‘Jt)).—-C00——C‘.0tJCQt3C00‘.0—a0U‘C00C—at’J—aC\0cM-‘L)4t.)C00Q-ta‘.0Ct-JM—a—a(Mc)cM—aocC0%—90-a-aççj44.3C00QtJ-‘.0t-JCCC\t’JC00(ML)‘.0—acMLtJ—00%4L’J4C—aC‘.0cM4L’JC000%3L’J‘.00%-J—cM‘.0C4cM0%——a-0%0%cMcMcMMcJcMc0%4—D——C00Cb°°°4——————————————————————-CCCt.)—•.jt-JL’JL)L)4c-c-b)L’3———c4oo‘-.‘—a-J00— cM————————————CCCCCCCCCC——00CCCCCCCCCCC——1-3——0—1-34.1-31-34400—a00•-44CCCCCCCCCCbbPbbbbbc(MC57Example: (Values are listed in Table 3)Mean mol. wt. std. = 5.992 kb Mean genomic DNA = 5.376 kbvariance = Cr2 = 0.0066 variance = u2 = 0.0152std. dev. = u = ± 0.081 std. dcv. = a = ± 0.123But use: Mean mol. wt. std. = 5.992 kbvariance = = 0.0152As it was not practical to obtain the variances for an unlimited number of DNA sizes, anaverage variance was used for each region bracketed by the mean molecular length standard values.Confidence intervals (mean ± 3u) for bins were established such that a genomic DNA fragmenttyped repeatedly in different lanes in the same gel or in different gels will be assigned to the samebin in 99.7% of the cases.Example: (Values listed in Table 3)1) Mean mol. wt. std. = 5.992 kb - —‘ —‘ Variance cr2 0.01522) Mean mol. wt. std. = 5.231 kb —‘ — —‘ Variance = 2 = 0.01523) Mean mol. wt. std. = 4.278 kb —‘ —‘ —‘ Variance = or2 = 0.0061Variance of region between 5.231 and 4.278 kb= average of of 5.231 and 4.278 kb= (0.0152 + 0.0061) ± 2 = 0.0106Standard deviation = a = ± 0.103 kbConfidence interval = = * 2 x 3or= 2 x 3(0.103)= 0.6189 kb* The 3u value is multiplied by 2 because delta has to account for ± 3a.58Hence, bin size for the region bound by the 5.23 14 and 4.2778 kb standard fragments,should be 619 bp j bins spaced 619 bp apart. The endpoints of each bin do not overlap toovercome the problem of bin assignment for alleles that are “on the line” of two bins.Population Data Base For the Allelic Frequencies of the 3”HVR of the D2S44LocusGenomic DNA from an undefined group of 578 individuals from the Greater VancouverArea was analyzed for allelic types (sizes) at the hypervariable region of the D2S44 locus. Thiscorresponded to the analysis of the alleles of 1156 chromosomes. Comparison of bin (allelic)frequencies of the first 578 chromosomes with the total frequency values obtained with 1156chromosomes revealed no significant difference (at 5% significance level) between the twofrequency distributions (Figure 7). This suggests that a sample size of only 289 individuals (578chromosomes) is large enough to provide an adequate estimate of the frequencies of the DNAfragment sizes (alleles). However, because methodological errors are replicated in a single study,this is not a reflection on the accuracy of the DNA fragment sizes in relation to their true sizeswhich is unknown without DNA sequencing.The frequency distribution for the alleles of D2S44 is shown in Table 5 and also graphicallyin Figure 8. For statistical analyses, the frequencies for bins 1-7 (DNA fragments 986 bp) will becombined as with bins 26-29 (fragments 4501 bp) in order to reduce the significance placed onthese rare alleles. This is in accordance with the procedures followed by the FBI’s DNA specialistswho suggested that bins which have less than 5 observed events are not statistically significant [104].In order to avoid situations where a DNA fragment, lying on or near the boundary of two bins, isassigned a frequency value lower than the true frequency of the sample population, the highestvalue of either the identified bin or its two flanking bins will be used in the calculations of theindividuality of the DNA profile.The frequencies of these D2S44 alleles show a slightly overlapping bimodal distributionwith the first mode at bin 14 (1536-1665) and the second (smaller) mode at bin 22 (3028-3247 bp)Figure 7 : Two Allele Frequency Distributions For the D2S44 LocusGenomic DNA from an undefined group of 578 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D2S44 locus.Allele frequency distributions of sample sizes (578 chromosomes and 1156 chromosomes)are compared.5920Figure7D2S44(1156/578 Chromosomes)15zm[L1234567891011121314151617181920212223242526272829*Bin•Total (1156Chromsomes)ElFirst 578Chromsomes*BinnumberasinTable5CFigure 8 : Allele Frequency Distribution For the D2S44 LocusGenomic DNA from an undefined group of 578 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D2S44 locus.Allele frequency distribution of this locus is shown for 1156 chromosomes.61Figure8D2S44(1156Chromosomes)14 12-00100000008 6 40t,C”(‘400N0001234567891011121314151617181920212223242526272829*Bin*BinnumberasinTable5C’63(Figure 8). The distribution obtain in this study is very similar to those obtained by both thePromega Corporation (collection of data from other laboratories) [105] and the RCMP laboratory[47]. Promega is a Wisconsin based company committed heavily to the promotion andadvancement of DNA typing. Its frequency distribution for D2S44 has one mode at approximately1660 bp (Table 6) and another smaller mode at approximately 3030 bp. Similarly, the frequencydistribution obtained by the RCMP showed the first mode at the size range of 1508-1637 bp (Table6) and the second smaller mode at the size range of 2863-3033 bp.The slight differences in the mode (most common allelic sizes) of the three distributionsare due to the fact that Promega data collection does not use a binning system and the RCMP havedifferent bins from the ones used in this study. Depending on the position of the arbitraryboundaries of the bins, the position of the mode for one distribution may be plus or minus one binfor another distribution. To solve this discrepancy, it may be more appropriate to compare thecumulative frequencies for a specific range of DNA fragment sizes for which the boundaries bisectat the low points of the distributions (Table 7). The frequency values obtained in this studyindicate that approximately 50.8% of the D2S44 alleles are within the size range of 1427-2365 bp(bins 13-18) and 21.1% are within the range of 2807-3621 bp (bins 21-23). The Promega frequencyvalues for two similar sizes ranges are approximately 45% and 21.5% and the RCMP values areapproximately 48% and 25.2%. In each of the three analyses, these two size ranges account forabout 2/3 to 3/4 of all D2544 alleles. The comparable values of each of the two cumulativefrequencies in all three distributions reaffirms the reliability of the sample data to provide anaccurate estimate of true (population) allelic frequencies and true allelic sizes.Population Data Base For the Allelic Frequencies of the 3’HVR of the D16S85LocusGenomic DNA from an undefined group of 608 individuals from the Greater VancouverArea were analyzed for their allelic types (sizes) at the 3’hypervariable region of the D16S85 gene.This corresponded to the analysis of the alleles of 1216 chromosomes. The various D16S85 alleles64Table 6 : Comparison of the Most Common AllelesMost CommonLocus Lab Sample SizeSize Range (bin) FrequencyUBC 578 1536-1665bp 11.5%(14)* *D2S44 Promega 487 —1660bp 9%RCMP 565 1508 - 1637bp 13.8%(9)UBC 608 775-822bp 13.9%(5)Dl 6S 85 Promega -- -- --0000 - 1077bp 49.1%RCMP 596 (1->5)____678 -726bpUBC 586 29.4%(3)5’HVRInsulin Promega -- -- --RCMP -- -- --1427- 1535bpUBC 530 15.0%(13)* *D14S13 Promega 891 -‘l5lObp 16%RCMP -- --Bracketed numbers are bin numbers- UBC Sample Population : undefinedPromega Collection Sample Population : Whites (unknown origin)RCMP Sample Population Whites (from Ottawa, Canada)*- Promega values are approximations only.65Table 7: D2S44 Cumulative Frequency for a Specific Size RangeLab Size Range (bp) Cumulative Frequency1427-2365 50.0%1st UBC (13) (18)Mode* *Promega —1501 - 2138 45 %1508 - 2351 48.0%RCMP (9) (13)UBC 2807-3631 21.8%(21) (23)2’Mode Promega —2815 - 3637*22 %*2863 - 3674RCMP (17) (19) 25.2%Bracketed numbers are bin numbers*- Promega values are approximations only.66were assigned to bins as described for the D2S44 alleles (Table 5). Comparison of the bin (allelic)frequencies of the first 608 chromosomes with the total frequency values obtained with 1216chromosomes revealed no significant difference (at 5% significance level) between the twofrequency distributions (Figure 9). This suggests that a sample size of even only 304 individuals(608 chromosomes) is large enough to provide an adequate estimate of allelic frequencies, thusconfirming the previous results with D2S44.The frequency distribution of alleles (bins) is skewed to the right with the mode or themost common fragment in the size range of bin 5 (775-822 bp) (Figure 10). Bin 5 in combinationwith its two surrounding bins, 4 and 6, account for approximately one third of all the alleles of the3’HVR of the D16S85. The remainder of the alleles are spread between the other bins and exceptfor bin 1, the frequencies are all less than 5.0% for each bin.Frequency values for bins with rare observations will be combined so as to reduce theirsignificance during statistical analyses. The distribution obtained in this study for alleles of thehypervariable region of D16S85 is similar to those obtained by the RCMP [47]. (Promega does nothave published data for this locus). The frequency distribution published by the RCMP laboratorygroup is also skewed to the right with the most common DNA fragment sizes in the range of 0000-1077 bp (Table 6) and the remainder of the alleles with sizes of up to approximately 7000 bp andoccurring at frequencies of < 6.0%. Since the binning systems are different for the twodistributions, the cumulative frequencies for a specific size range were compared (Table 8). TheRCMP have combined all the alleles 1077 bp into one bin with a frequency of 49.1%. Similarlythe cumulative frequency for the all the bins in this study which account for the DNA fragmentssizes of 1069 bp (bins 1-8) is 50.5%. Bins 1-8 have not been combined in this study because:1) the number of observed events in each bin is> 16 and is thussignificant2) the size of each bin was determined according to a 99% confidenceinterval (± 3o-)Figure 9: Two Allele Frequency Distributions For the D16S85 LocusGenomic DNA from an undefined group of 608 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D2S44 locus.Allele frequency distributions of sample sizes (608 chromosomes and 1216 chromosomes)are compared.6720.Figure9D16S85(1216/608Chromosomes)15 10__iiIll___________1234567891011121314151617181920212223242526272829*. Bin•Total(1216)Chromosomes]First608Chromosomes*BinnumberasinTable5Figure 10 : Allele Frequency Distribution For the D16S85 LocusGenomic DNA from an undefined group of 608 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D16S85 locus.Allele frequency distribution of this locus is shown for 1216 chromosomes.6920Figure10D16S85(1216Chromosomes)15-41000005-400-400C—-4-4-4-4C”01234567891011121314151617181920212223242526272829*Bin*BinnumberasinTable571Lab Size Range (bp) Cumulative FrequencyUBC 0000 - 1069 50.5%t(1) t(8)0000 - 1077RCMP *(1•••> 5) 49.1%Bracketed numbers are bin numberst - bins not combined* - bins combinedTable 8: D16S85 Cumulative Frequency for a Specific Size Range723) the allele frequency for any one bin should be already an overestimation of any one fragment size (allele).Therefore, the grouping of all the alleles < 1069 bp would be an overly conservative estimate.However, it would not be unreasonable to reduce the number of bins in this size region (as seenwith the FBI binning system) by merging two or more bins. Alternatively, for statistical calculationsof the individuality of a DNA profile, the bin frequencies of alleles three standard deviations (for a99% confidence interval) from the measured allele size of interest may be combined to ensure anover estimation of the true frequency.For example: (refer to Table 5 for bin frequencies)Frequency of D16S85 allele with size 2120 bp = Xstandard deviation for this size range = a = 36.7 bpX + 3a 2120 + 3(36.7) = 2230 bpX - 3u = 2120 - 3(36.7) = 2010 bpFrequency of 2120 bp = Frequency of alleles 2010-2230 bp= Frequency of bin 17 + bin 18 = 8.3%This approach combines the “fixed-bin” system used for the data base generation of this thesis andthe “floating-bin” system used by Lifecodes Incorporated. The “fixed-bin” system providesconservative allele (bin) frequencies by assigning alleles of a specific size range to a particular binwhose width is determined by the standard deviation of the alleles of that size category [104,106].Lifecodes’ “floating-bin” system involves measurement of alleles at 100 bp increments and the allelefrequency used in calculations were obtained by adding the values contained within two standarddeviations of the mean size determination [87]. (A disavantage of the “floating-bin” system is thatan assigned allele size of an evidence sample is assumed to be the mean value and such anassumption may lead to a lower frequency estimate for the allele [104]. No further comment canbe made regarding this system due to the absence of published details.) The combination of thesetwo system makes the frequency estimates much more conservative than that obtained by each73method alone. This conservative method of calculation would also safeguard against concerns [107]that an allele will be assessed a lower frequency than its true value if the arbitrarily assignedboundaries of a bin bisects the peak of a distribution.Population Data Base For the Allelic Frequencies of the 5’HVR of the InsulinGene:Genomic DNA from 586 undefined individuals from the Greater Vancouver Area wereanalyzed for their allelic types (sizes) at the S’hypervariable region of the insulin gene. Thiscorresponded to the analysis of the alleles of 1172 chromosomes. Various insulin alleles wereassigned to bins as described for the D2S44 alleles (Table 5). Comparison of the bin (allelic)frequencies of the first 586 chromosomes with the total frequency values obtained with 1172chromosomes revealed no significant difference (at 5% significance level) between the twofrequency distributions (Figure 11). This suggests that a sample size of only 293 individuals (586chromosomes) is large enough to provide an adequate estimate of the allelic frequencies, againconfirming previous results.The frequencies of the alleles show a bimodal distribution with the first mode at bin 35(678-726 bp) and the second (smaller) mode at bin 18 (2146-2365 bp) (Figure 12). As before,frequency values for bins with rare observations will be combined so as to reduce their significanceduring statistical analysis. Numerous bins with extremely low frequencies separating the two modesis clearly indicative of two very distinct size ranges of DNA. The small alleles (bin 1-5) account forapproximately 75% of the total alleles and the larger alleles (bins 17-19) account for nearly 24%.Alleles with intermediate sizes (bins 6-16) are rarely observed. There is an absence of publisheddata for the allelic distribution of the 5’ hypervariable region of the insulin gene cleaved withrestriction endonuclease Hae III. However, Bell et al. have collected data at this locus by cleavinggenomic DNA with restriction endonuclease Bgl I [42]. Their allelic frequencies exhibited acomparable bimodal distribution with the smaller DNA fragments (approximately 2.7 - 3.3 kb)accounting for 75% of alleles and the remaining 25% made up of the larger alleles (approximatelyFigure 11 : Two Allele Frequency Distributions For the INS LocusGenomic DNA from an undefined group of 586 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the iNS locus.Allele frequency distributions of sample sizes (586 chromosomes and 1172 chromosomes)are compared.74Figure115’HVRInsulin(1172/586 Chromosomes)30 253*Bin•Total (1172Chromosomes)First586Chromosomes*BinnumberasinTable5Figure 12: Allele Frequency Distribution For the INS LocusGenomic DNA from an undefined group of 586 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the INS locus.Allele frequency distribution of this locus is shown for 1172 chromosomes.7630 25 20.w I-.Figure12S’HVRInsulin(1172Chromosomes)35N —15 10 5 0Il•)123‘fl00000000000—0ooooo0I••.IIIIIIIIIIII567891011121314151617181920212223*. Bin000000—000—--EL*BinnumberasinTable5784.3 - 4.9 kb) [43]. The DNA fragment sizes (alleles) from this laboratory and those of Bell et al.,cannot be directly compared since different restriction endonucleases were used. Restrictionfragment sizes are not only determined by the VNTR size but also the DNA length contributed byflanking regions which are dependent upon the particular restriction endonuclease used for thegenomic DNA digest.Contrasting results have been shown for a black American population sample usingrestriction endonuclease Sac I [108]. The allelic frequency distribution obtained with the blackAmerican data is unimodal compared with the bimodal distributions demonstrated for the twostudies discussed above. This suggests that there are subpopulation differences at the 5’HVR ofthe insulin gene and that the undefined populations sampled by this author and that of Bell et al.are either similar or consist of a mixture of subpopulations that do not have significantly differentallele frequencies.Population Data Base For the Allelic Frequencies of the HVR of the D14S13Locus:Genomic DNA from 530 undefined individuals from the Greater Vancouver Area wereanalyzed for their allelic types (sizes) at the hypervariable region of the D14513 locus. Thiscorresponded to the analysis of the alleles of 1060 chromosomes. The various D14513 alleles wereassigned to bins as described for the D2S44 alleles (Table 5). Comparison of bin (allelic)frequencies of the first 530 chromosomes with the total frequency values obtained with 1060chromosomes revealed no significant difference (at 5% significance level) between the twodistributions (Figure 13).The frequency distribution of the alleles (bins) is skewed to the right with the mode ormost common DNA fragments in the size range of bin 13 (1427-1535 bp) (Figure 14). Frequencyvalues for bins with rare observations will be combined so as to reduce their significance duringstatistical analysis. The distribution obtained in this study for the alleles of D14S13 is similar tothat obtained by the Promega Corporation collection [1091. The RCMP laboratory group has notFigure 13 : Two Allele Frequency Distributions For the D14S13 LocusGenomic DNA from an undefined group of 530 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D14S13 locus.Allele frequency distributions of sample sizes (530 chromosomes and 1060 chromosomes)are compared.7920Figure13D14S13(1060/530 Chromosomes)15C101234567891011121314151617181920212223242526272829*Bin•Total (1060Chromosomes)fjFirst 530Chromosomes*BinnumberasinTable5Figure 14 : Alelle Frequency Distribution For the D14S13 LocusGenomic DNA from an undefined group of 530 individuals from the Greater Vancouver Areawas analysed for alleleic types (sizes) at the hypervariable region of the D14S13 locus.Allele frequency distribution of this locus is shown for 1060 chromosomes.8120Figure14D14S13(1060Chromosomes)I,,15I..,1000 0000C’‘C5Cr00 -‘Itn—CCCCC01234567891011121314151617181920212223242526272829*Bin*Binnumber asinTable5GO83published data for this locus. The frequency distribution published by the Promega Corporation isalso skewed to the right with the mode at approximately 1510 bp (Table 6). For comparisonpurposes, cumulative frequencies for a specific size range were calculated for two distributions(Table 9). The frequency values obtained in this study indicate that approximately 50% of theD14S13 alleles are within the size range of 1427-2145 bp (bin 13-17) and the remainder of thealleles occurring at frequencies of < 5%. For a size range of approximately 1374-2055 bp, a similarcumulative frequency value of 55% is seen in the Promega distribution with the remainder of thealleles occurring at frequency values of < 5%. The agreement between these two sets of datareinforces the reliability of the sample data to provide an accurate estimate of the true frequenciesand true allelic sizes.Comparison of the D2S44 and D14S13 distributions indicate that the first mode of D2S44almost coincides with the mode of D14S13 (Figure 15). This would present an analytical problemif the two probes for these loci were used to detect simultaneously the alleles of the two loci.There would be a high probability (since the common alleles overlap) that the alleles of one locuswould be of a very similar or identical size to the alleles of the other locus. Since overlappingbands cannot be detected, an individual heterozygous at both loci may be falsely categorized asbeing homozygous for one or both loci. Hence simultaneous probings of different loci should notinvolve loci with closely overlapping common alleles.Simultaneous Detection of Two Single Locus VNTR Regions:A combination of several single locus VNTR probes (to detect multiple regions) are usuallyrequired to create a discriminating DNA profile. Sequential probing of membrane-fixed DNA maybe time consuming depending on the amount of fixed DNA and the efficiency of the labelledprobe. It may also result in lower quality autoradiographs in the final probings. An alternative isto probe simultaneously for more than one locus at a time as seen in Figure 16. Two alleles eachwere detected for five individuals tested when one probe was used per hybridization reaction(Figure 16 - Panels A and B). When two probes were used per hybridization reaction, four alleles84Table 9: D14S13 Cumulative Frequency for a Specific Size RangeLab Size Range (bp) Cumulative FrequencyUBC 1427-2145 50.0%(13) (17)Promega —1374- 2055* 55%*Bracketed numbers are bin numbers*- Promega values are approximations only.Figure 15 : Comparison of Allele Frequency Distributions of D2S44 and D14S13Allele frequency distribution of D2S44 (1156 chromosomes) and D14S13 (1060 chromosomes)are compared to determined the areas of overlapping alleles.8520-Figure15DistributionComparisonofD2S44andD14S1315U I-.10___kiiii1234567891011121314151617181920212223242526272829*Bin•D2S44DD14S13*BinnumberasinTable5Figure 16: Simultaneous Probing of Two LociGenomic DNA from 5 individuals (lanes 1-5) were probed with VNTR probes 3’HVRa-globin(Panel A), YNH24 (Panel B) and combination of 3’HVR a-globin/YNH24 (Panel C).[3 days exposure to Kodak X-MAT AR film at -70°C with screens]87Figure16:SimultaneousProbingofTwoLoci2342345234I..,•••SiBa’-SCA00 0089were detected for each individual, corresponding to two alleles per locus (Figure 16 - Panel C).The presence of four alleles for each individual (Panel C: lanes 1-5) eliminates the possibility of anallele(s) of one locus covering the allele(s) of the other locus; a result that would lead to anincorrect conclusions.The combination of probes used were the 3’HVR/c-globin and YNH24 which detected theD16S85 and D2S44 hypervariable regions respectively. This combination was chosen because thesize range of the most common alleles of D16S85 was distinct from that of D2S44 (Figure 17).Although these two loci do have some overlapping alleles, these alleles are less common. Thusthere is a lower probability of an individual having rare alleles of the same size for both loci. Insituations where < 4 alleles are detected for simultaneous hybridizations with two single-locusprobes, the analysis would have to be repeated with sequential probings with one probe perhybridization reaction. Thus, more time and labour would be required than if the analysis wereperformed initially with sequential probings. The possibility of overlapping alleles increases inproportion to the number of regions detected simultaneously. Hence, the possible benefit in timesaving of simultaneous probings (two or more) may not be great enough for multiple probings toserve as a viable alternative for sequential probings.DNA Profiles:Genetic typing of 470 individuals did not show any identical bin profiles at all four of theVNTR loci, D2S44, D16S85, INS and D14S13. Two individuals did exhibit the same bin pattern at3/4 of the loci, but the bin numbers at the fourth locus were far apart. Identical bin profiles refersto alleles that fall into the same bin which in turn represents a range of allele sizes. Hence, even ifbin numbers (alleles) at all four loci were identical, the two profiles may not constitute a real“match” of alleles. In order for a match to be declared, the two identical bin samples must bereanalysed on the same gel to provided a more accurate measurement. This concept will bediscussed further in the following paragraphs regarding measurement imprecision.Figure 17: Comparison of Allele Frequency Distributions of D2S44 and D16S85Allele frequency distribution of D2S44 (1156 chromosomes) and D16S85 (1216 chromosomes)are compared to determined the areas of overlappingalleles.9020Figure17DistributionComparisonofD2S44andD16S8515G.) 10nj1234567891011121314151617181920212223242526272829*BinD2S440D16S85*BinnumberasinTable592Measurement Imprecision:The ultimate goal of the RFLPIVNTR DNA typing procedure is to determine whether twoDNA banding profiles are identical and hence may be attributed to one individual. Two alleles,from two different lanes in the same gel, which fall into the same bin do not necessarily indicate apositive “match” for those DNA fragments. Conversely, two alleles with visually matchingmobilities may fall into different bins depending on the arbitrary boundaries of the bins. The binsystem only serves to provide conservative frequency estimates for alleles of the VNTR loci. Thesefrequency values are used to calculate the uniqueness of a DNA profile once a “match” has beendeclared.An objective basis for declaring a “match” between two samples is possible by incorporatingan experimentally determined measurement imprecision value [47]. The degree of measurementimprecision is defined as the maximum size difference between identical DNA samples. The sizesof two DNA bands, from two different samples, must be within the window of measurementimprecision of each other before they are assessed as a “match”. The degree of measurementimprecision for various sized genomic DNA fragments within or between gels was determined(Table 10). The degree of measurement imprecision within a gel appears to be relatively constant,ranging between 3.7- 5.5%. However, there does not appear to be any correlation betweenmeasurement imprecision and fragment length especially in the measurements within a gel. Sinceactual casework usually involves comparison of samples within the same gel, it would beappropriate to use a value of 5.5% as the “window” for measurement imprecision for “match”determinations.One can argue that if the “window” for measurement imprecision was too large, DNAfragments which are of different sizes will be falsely declared as “matched”. As a consequence ofthe resolution of agarose gel electrophoresis and minute differences possible between VNTRalleles, false “matches” can occur even in the absence of measurement errors. The presence of“windows” is to protect against false negatives (non matches) resulting from measurement errors.93*G Measurement MeasurementImprecision ImprecisionDNA (kb) (within gel) (between gels)5.376 4.9 % 9.0 %4.333 4.1 % 9.8 %3.961 5.5 % 7.1 %3.159 4.8% 6.5%2.610 4.4 % 6.0 %1.560 4.2% 6.6%1.490 5.5% 73 %1.270 4.9 % 6.3 %0.830 3.7 % 4.8 %0.660 4.2 % 6.0 %* = Mean determinationsMax - MmMeasurement imprecision = x 100MeanTable 10: Measurement Imprecisionn=9094The occurrence of these false positive “matches” is equalized by the binning system of VNTRalleles. Bin sizes are all greater than the maximum measurement imprecision within gels (5.5%).Hence, as previously discussed, during data base generation there is a tremendous over estimationof the frequency of any one allele. Therefore, in the event of a false “match” with one VNTR, thefrequency of the allele (bin) will be higher than its actual frequency which reduces the uniquenessof that allele and the corresponding value of the “match”. Since a “matched” DNA profile isgenerated by analyzing several VNTR loci, the unlikely occurrence of multiple false “matches” forany two samples will reduce the likelihood of an inaccurate interpretation of the typing results[110]. Furthermore, a quantative match criterion is used only after an initial visual match has beenmade.Imprecision within a gel for the 2.610 kb fragment (4.4%) in this study is higher than thevalue from the RCMP laboratory for the 2.724 kb fragment (approximately 2.9%). This differencemay be because the RCMP laboratory value is based on the “maximum size difference encounteredwhen two identical samples are compared relative to the same set of flanking size markers” [47]and more than one set of markers are present in one gel. In contrast, the values calculated fromthe data of this study included all measurements across the entire gel thus, accounting for moreerrors. The comparability of the imprecision values calculated from data from both laboratories isconfirmed by the values obtained from measurements between different gels which are 6.0% (2.610kb) for this laboratory and 5.5% (2.724 kb) for the RCMP laboratory [47].Validity of Subjective (Manual) DNA Band Assessments:One of the major concerns of any laboratory is the validity of generated data and thereliability of analyses. The allelic frequencies for the four VNTR loci analyzed in this laboratorywere vulnerable to errors due to subjective DNA band measurements by a single individual.However, analyses and comparison of data with other laboratories which utilize automated(objective) assessments, confirm the validity and reliability of this system. This confirmation isbased on the fact that:951) -the frequency distribution (Figure 18) of 364 allelesobjectively analyzed with the FBI’s image analysing system was notsignificantly different (95% level) from the distribution of the samealleles measured manually (Figure 18)-the discrepancies ranged from 0.0% in bin 5 to 3.1% in bin 152) -frequency distributions of the alleles of the four VNTR loci arecomparable to those of the Bell et al., RCMP laboratory and/orPromega Corporation3) -a high degree of precision (or small relative standard deviation)was obtained4) -the level of measurement imprecision for the 2.6 kb size range wascomparable to that obtained by the RCMP laboratoryThe validity of subjective DNA band measurements is corroborated by a European study inwhich a comparison was made between fragment sizes calculated manually with a plotted curve orcalculated using a computer program and digitizing tablet [111]. Their analysis of 72 DNA(fragments) bands indicated that there was no statistically significant differences between the twoalternative methods of DNA sizing.Figure 18: Comparison of Data From Computerized VersusNon-computerized MeasurementsData for the D16S85 locus (364 chromosomes) was collected by using manual bandassements or the FBI laboratory’s image analysis system in Quantico, Virginia. Allelefrequency distributions of the two sets of data are compared to determined whetherthere is any significant difference between data sets.9615.Figure18DataFromNon-ComputerizedvsComputerizedMeasurements(364Chromosomes)10U[ill__________1234567891011121314151617181920212223242526272829*BinINon-ComputerizedMeasurements121ComputerizedMeasurements*BinnumberasinTableS98CONCLUSIONDuring the last four years, there has been a tremendous amount of work devoted to theresearch and development of DNA genetic typing for human identification. New technicalrefinements have appeared at a very rapid rate from government, corporate and universitylaboratories. Accordingly, various protocols in this laboratory have been refined or replaced withprocedures developed either here or in other facilities.The accuracy (amount of deviation from the actual size) of the typing results obtainedcould benefit from some improvement. However, the maximum level of inaccuracy (approximately5.0 % of molecular size) was less than the window (greater than 13 % of molecular size) allottedfor the bin of that size range. This would allow for a large enough “wobble” such that theassignment of DNA fragments to bins would not be drastically altered to cause appreciable changesin allelic frequencies.The level of precision of this DNA typing system as measured by the relative standarddeviation, is much higher than the level of accuracy. Although both aspects of the system areimportant, precision is more critical since the reliability of the results is dependent upon obtainingthe same results in repetitive trials. The degree of accuracy is relative from one sample to the nextand thus with a binning system, the allelic frequency distribution will not be altered. Bin sizes forallele assignments were determined according to the degree of precision of the typing system sothat DNA fragments of the same size can be assigned to the same bin in 99.7% of the cases. Thelevel of precision of this laboratory (maximum relative deviation of ± 2.3% - Table 4) iscomparable with those of other forensic laboratories who quoted a value of ± 2.5% (or 5%) at the1991 Promega International Symposium on Human Identification [A. Autor- personnelcommunication].Measurement imprecision, defined by the maximum size difference between identicalsamples, serves as the basis for declaring a match of alleles from different samples. The maximumdegree of measurement imprecision (within a gel) of this typing system for DNA fragments99between 5.3 - 0.6 kb is approximately 5.5% of the fragment size. Thus, fragment comparisonsvarying up to 5.5% can still be considered as “matched” (inclusion). In most instances, the valuesfor measurement imprecision within the same gel will be used because casework material will moreoften than not be on one gel.Allelic frequency distributions generated for four VNTR loci are comparable to publisheddata from major laboratories such as that of the RCMP, and Promega Corporation associatedlaboratories. The population sampled by the RCMP and Promega associated laboratories areCaucasians, whereas this study analyzed an undefined population. Detailed comparisons ofcumulative frequencies of specific size ranges of alleles at various loci obtained by the three groupshave shown striking similarities with a estimated maximum difference of 5% at the D14S13 locus(Table 10). A more general representation of the similarities can be seen in Figures 19-21. Thecomparability of these different data bases may suggest that the1) undefined population samples consists of mainly Caucasians, and/or2) undefined population samples consists of certain racial or ethnicsubpopulations which have the same or very similar frequency valuesas those of Caucasians3) frequency values are homogenous for all subpopulationsThe first two alternatives are indistinguishable here because the ethnic or racial origin ofindividuals in the undefined populations unknown. However, the first suggestion is a reasonablepossibility since the Vancouver population is composed of 75.6 % whites (non-Hispanic), 22.1%Asian, 0.3% Black, 0.5% Hispanic, and 1.5% Native North American [112]. The latter suggestionis probably invalid since analysis of three racial subpopulations (American Blacks, Caucasians, andHispanics) have shown subpopulation heterogeneity for at least five loci (HRAS-1, D14S1, D2S44,D14S13 and 5’HVR insulin) [87,88,108]. Analysis of the D2S44 locus by the RCMP has revealedheterogeneity in the Caucasian and Native Indian Population [113]. Furthermore, there has been areport of slight differences in allelic frequencies of the same racial group (Blacks) in differentFig.19:D2S44AllelicDistributionComparisonsSize(kb)—usc—I—RoMp—*-PromegaSizevaluesrepresentmidpointsofconsecutivesizerangesFig.lOb:RCMPD2S44AllelicFig.19c:PromegaD2S44AllelicFig.19a:IJBCD2S44AllelicDistributionDistributionDistributionI0S46681m1kb)Sue1kb)SIze1kb)Si..Value.r.pr...ntmidpointsSi..6.10..repr...ntmidpoint.SeaVeiuflr.p6.nntmidpointsOnctsIe.oatibsSitsenS..Of0,nS.OutiV..IrOnO.,Of000050uti6.bee15n4..©Fig.20:D16S85AllelicDistributionSizevaluerepresentmidpointsofconsecutivesizerangesComparison—UBC+-RCMPFig.20a:UBCD16S85AllelicDistribution$60F e q U e n C y $10F$50r e$40q U e$30n C Y$20Size(kb)—USC0245010$10 $002468Size(kb)10SloeVeluerepresentmidpointsofconsecutivesloersngesFig.20b:RCMPD16S85AllelicDistributionF e q U 0 n C y $Size(kb)10SloeveluerepresentmidpointsOfconsecutivesloerur,oeeF e q U C n C ySloevalue.aremidpointsatConaCcuti,.Cbsrang..Fig.21a:UBCD14S13AllelicDistributionSinsvaluesar.midpoint.atConsecutivesinsrangesFig21:D14S13AllelicDistributionComparisons26 20 16 10F r e q U e n C y$61802468¶01214Size(kb)—UBOFig.21b:PromegaD14S13AllelicDistribution02468Size(kb)uBc-+—PromegaSizevaluesaremidpointsofconsecutivesizeranges10121416e q U e n C y $8101214Size(kb)Promeget’-)103geographical locations across the U.S.A. [114]. This phenomenon of frequency differences in thesame racial group was not observed in a comparison of Caucasian populations from Dusseldorf andAmsterdam [1111.The results obtained in this study suggest that this data collected by non-conventionalsampling procedures such as sampling from an undefined population from the Greater VancouverArea, would be applicable to other subpopulations where there is a majority of Caucasians and nosignificant restriction on gene flow due to barriers such as geographical isolation or religiousbeliefs. This would be true for loci which do not have large variations in allele frequency betweenracial subpopulations. Slight variations in allelic frequencies of the other racial subgroups present,would not significantly alter the overall frequencies because the contribution made by any onesubgroup would be small. Furthermore, allelic frequencies generated in this thesis for theundefined population is based on a fixed bin system which minimizes potential subpopulation(racial or geographical) differences [110]. The application of this system to data from theTechnical Working Group on DNA Analysis Methods participants, suggests that there are nodramatic differences in allelic (bin) frequencies for Caucasian (or black) geographicalsubpopulations across the U.S.A. and Canada [110].Allelic frequency data bases generated for one geographical population would not always beapplicable to another population due to differences in types and proportions of racialsubpopulations. However, each geographically distinguishable population could use its own database, or data collected from the racial subpopulation which account for a large majority of thespecific test population in question, or data collected from a geographical population to which itsracial subgroup makeup bears close resemblance. The possibility of utilizing databases collectedfrom geographically distinct areas would eliminate the neccessity of building separate databases foreach and every region. Another alternative would be to incorporate the highest value obtained forthe specific allele from any one data base (geographical or racial subpopulation) in the calculations104for uniqueness of DNA profile. This conservative approach (used by the RCMP) will safeguardagainst biasing against individuals of any one subpopulation leading to false interpretations.There has been considerable debate as to the validity of the data bases generated for theVNTR loci. Much of this controversy relates to the existence or non existence of Hardy-Weinbergequilibrium for the loci used in DNA genetic typing. It is highly probable that VNTR loci are inequilibrium but are perceived to deviate due to the lack of an adequate test for equilibrium at lociwith non-discrete alleles [E. Wijsman - personal communication]. This opinion is in accordancewith others who believe the Hardy-Weinberg test does not apply to VNTRs because of thecontinuum of allele sizes [B. Budowle - personal communication] and that the test is full ofartifactual pitfalls [115]. With present analytical capabilities, such systems are predisposed to excesshomozygosity [E. Wijsman-personal communication]. Excess of homozygotes have been shown forVNTR systems regardless of whether they are or are not in equilibrium [104]. The frequency ofobserved heterozygotes (Table 11) in this study is comparable to that obtained by other laboratories(as reported by the Budowle et.al. [106]) and by the RCMP laboratory [47]. Further analysis withrespect to excess of homozygosity / heterozygosity, was not performed. Since present methods usedto determine excess are inappropriate, the results from these methods would not be meaningful.This author does not feel that one should use a method that has been shown to be inappropriate(E. Wijsmen- unpublished results), to confirm the inconclusive results of other researchers.More recently Devlin et.al. [103,116] have purposed another method for determiningequilibrium by using only data consisting of heterozygotes and omitting the homozygotes(“contaminated data”). There have been some criticisms of this method [117]. However, thismethod has not yet been applied to a large amount of VNTR data or analysed by otherresearchers. Although data in this thesis may benefit from some analysis with this new method,this analysis was not performed since this computer program is not publically avaliable.This thesis does not attempt to deal with the Hardy-Weinberg equilibrium because of thecomplexity of this problem. There is no universal agreement as to what valid statistical method105Table 11 : Comparison of Heterozygote FrequenciesFrequency of HeterozygotesLaborator D16S85 D2S44 INS D14S13UBC 87% 92% 75% 90%RCMP 69% 91% N/A N/AFBI 90% 91% N/A 90%UBC data from this studyFBI = data from various laboratoriesas reported by Budowle et. al. [107]N/A no available dataRCMP = data published in reference 47106could be utilized to test for equilibrium. Forensic DNA experts at the Promega DNA Symposiumin April of 1991, have reported that judges have recently stopped requesting information for theHardy-Weinberg test for VNTR loci in cases where DNA identification is presented in evidence [A.Autor - personal communication].107SECTION IIFORENSIC CASESCASE I - Body Identification (1988)Background Information:On April 5 1988, two CF18 aircrafts scrambled from CFB Comox in the early hours tosearch for a fishing vessel in distress off the west coast of Vancouver Island. The master of thevessel had radioed a “May-Day” distress call giving his position as south of Brooks Peninsula.While one aircraft maintained an altitude of 23,000 feet, the other aircraft descended toapproximately 2,000 feet to try to locate the fishing vessel visually. After repeated failed passes,the pilot of the low attitude aircraft told the master of the fishing vessel that he was going toengage his “after-burners” such that the vessel might see and/or hear the aircraft. No furthercontact was made with the pilot of that particular aircraft. The fishing vessel was rescued hourslater by the Canadian Coast Guard.Another search was initiated for the missing CF18 aircraft. Five days later, the crash site ofthe aircraft was found. The aircraft had hit the side of a mountain with its “after-burners” full on,impacting at a speed approaching mach 1 (600-700 mph). Very little remained of the aircraft andonly approximately 2.5 kg of tissue from the pilot was recovered from the crash site.Since the number of individuals capable of flying a CF18 aircraft is quite limited, theidentity of the pilot was not particularly in doubt. However, to reduce chances of mistaken identityand hence the level of anxiety of the pilot’s family, an attempt was made towards some means ofidentification. DNA genetic typing was utilized as no other means of identification was possible(i.e. no dental remains, fingerprints, or material suitable for radiological or anthropologicaltechniques).108DNA Typing Results:The parents of the missing pilot provided blood samples for comparison with the tissuesrecovered from the crash site. This analysis was performed to determine whether the profile of thetissue DNA is consistent with that of a offspring of the two individuals who provided the bloodsamples (parents of missing pilot). An exclusion (non-consistent DNA profile) is inferred if anyparticular sized DNA fragments found from the tissue is not equalled by a fragment from at leastone tested parent.The DNA samples were probed with three VNTR probes : YNH24, 3’HVR cx-globin, andphins 310. The autoradiographs from two of the probings can be seen in Figures 22 and 23. Ineach case, DNA bands (fragments) found in the tissue DNA were also present in the profile ofDNA from the parents of the missing pilot. These results do not exclude the tested parents aspossible parents of the donor of the tissue. To evaluate such an inclusion, two hypotheses must beconsidered as discussed by D.A. Stoney [118]. These hypotheses are: a) that these tested parentspassed the DNA alleles found in the tissues, and b) that a random couple have passed these alleles.In order to determine which hypothesis to accept, the probability of the observed alleles occuringunder each of these hypotheses is calculated. The ratio of these two probabilities are then used toproduce the relative likelihood of hypotheses “a” and “b”.Calculations: (according to formulations by D.A. Stoney [112])R Probability parents would pass tissue DNA types (genotypes)Probability a random couple would pass tissue DNA typesa) The numerator is dependent upon the zygosity of the tested parents andcan be a factor of 1.0, 0.5 or 0.25.eg. Couple Possible tissue types and frequencies1. XX, XX XX = 1.002. XY, XY XX 0.25 ; XY = 0.50 ; YY = 0.253. XX, XY XX = 0.50 ; XY = 0.504. XZ, YW XY = 0.25 ; YZ = 0.25 ; ZW = 0.25 ; XW = 0.25109Figure 22: Case I - Locus D16S85Genomic DNA samples were probed with to detect 3’HVR a-globin alleles at the D16S85locus.Lane m = molecular length standards1 = control genomic DNA sample with known alleles (female)2 = II It (male)3II It It It It It It (female)54 = genomic DNA sample (blood) from missing pilot’s mother53 = genomic DNA sample (blood) from missing pilot’s father52 = genomic DNA sample (tissue) found at crash siteLarge arrow: allele possibly contributed by tested motherLarge arrow and two small arrows : allele possibly contributed by tested father[1 day exposure to Kodak X-MAT AR film at -70°C with screens]23.13—+9.42—_+4.36—41.35-41.08110Figure 22: Case I - Locus D16S85028 ALPHA - GLOBIN (3’)CONTROL M F Pm 1 2 ‘ 54 51 52 m-1 day111Figure 23 : Case I - Locus D2S44Genomic DNA samples were probed with YNI-124 to detect alleles at the D2S44 locus.Lane m = molecular length standards1 = control genomic DNA sample with known alleles (female)2 = H (male)3 = It (female)54 = genomic DNA sample (blood) from missing pilot’s mother53 = genomic DNA sample (blood) from missing pilot’s father52 = genomic DNA sample (tissue) found at crash siteLarge arrow: allele possibly contributed by tested motherLarge arrow and two small arrows : allele possibly contributed by tested father[3 days exposure to Kodak X-MAT AP film at -70°C with screens]Note: Molecular length markers are not visible in this autoradiograph.Since the same membrane was probed each time, markers visible on theD16S85 autoradiograph (Figure 19) was used for extrapolation of DNA sizes.:3*112Figure 23 : Case I - Locus D2S44028 YNH24m 1 2 3 54 53 52 mI6.56—+4.36—41.353 days113Since the tested parents were heterozygous for different alleles (analogous to couple 3 in theexample above) at each of the three loci, the numerator in this case would be (0.25)(0.25)(0.25)or (0.25).b) The denominator is the product of the tissue DNA allele frequency in the population.Locus Allele Size FrequencyD2S44 5.868 kb * 0.1544.658 kb * 0.154D16S85 7.086 kb * 0.0124.053 kb * 0.050INS 0.864 kb * 0.3730.783 kb 0.373*- not the frequency value for that bin (allele)- the highest frequency value of either the identified bin or its twoflanking bins is used to err on the conservative sideFreq. of occurrence = 2pq x 2pq x 2pq= 2(0.154 x 0.154) x 2(0.012 x 0.050) x2(0.373 x 0.373)= 0.0000158c) Likelihood ratio “R” = (0.25) / 0.0000158 = 989The tested parents are 989 times more likely as a randomly selected couple to pass this setof DNA types found in the tissue. Considering this statistic and the extremely low portion ofcouples with offsprings who are capable of operating a CF18 aircraft, the identity of the pilot wasconfirmed. This analysis was performed in 1988 during the infancy of DNA genetic typing and ofthis laboratory itself. Standard procedures were not outlined and this laboratory started out using114the restriction endonuclease Pvu II for the digestion of genomic DNA. Restriction endonucleaseRae III have since become the standard for the North American laboratories and was adopted bythis laboratory in 1989. The genomic DNAs in this analysis were digested with Pvu II and theallele frequency values listed in the above calculations are not from Table 5 (Chapter 1) whichrepresent values obtained with Rae III. The frequency values obtained with Pvu II are not listed inthis thesis because analysis of the Pvu II data was minimal and these results cannot be corroboratedwith data from other laboratories. To establish some form of consistency between forensiclaboratories, the Pvu II values have not been used since 1989. Although the results of this caseanalysis may not be acceptable by today’s standards, it was adequate in 1988 in terms of the “state-of-the-art” for that time. This case serves to illustrate the applicability of DNA genetic typing tobody identification.CASE II - Associating Suspect With Victim (1988)Background Information:An assault victim was unwilling to identi1’ his assailant, who was a gang member, becauseof fear of retaliation by the gang. The police had suspected a certain individual as the assailant andconfiscated a pair of sweat pants from his room in his parents’ house. The sweat pants contained19 little dots (largest spot was <5 mm in diameter) of what appeared to be bloodstains. DNAextracted from these 19 stains were pooled and compared DNA extracted from blood DNA of thevictim. With conventional serological tests, one would not pooi samples because of the possibilityof more than one contributor. However, DNA typing with single locus VNTRs will only have amaximum of two bands on the autoradiograph for each probe (except in rare occasions as discussedin the introduction) so that multiple bands would be indicative of mixed samples. Since thequantity of recovered DNA was expected to be small, all the samples were used for one analysis.115DNA Typing Results:Successful DNA banding patterns were obtained with three VNTR probes (YNH24, 3’HVRa-globin, and phins 310) and two of the probings are shown in Figures 24 and 25. In each instance,the bands of the DNA extracted from the sweat pants (Figures 24, 25 - lane SP 115) matched thoseof the blood DNA of the victim (Figures 24, 25 - lanes VB 116). Calculations incorporating thefrequency of alleles from the three VNTR loci probed, were performed to determine theuniqueness of the DNA profile:Probe Allele Size FrequencyYNH24 3.986 kb 0.1953.819 kb 0.1953’HVR a-globin 6.305 kb 0.0245.512 kb * 0.045phins 310 0.617 kb 0.0420.602 kb 0.042*- not the frequency value for that bin (allele)- the highest frequency value of either the identified bin or its twoflanking bins is used to err on the conservative sideFreq. of occurrence = 2pq x 2pq x 2pq2(0.195 x 0.195) x 2(0.024 x 0.045) x2(0.042 x 0.042)= 0.00000058 = 1/1,725,513The probability of an individual having these six alleles (using these three probes) is approximately1 in 1.7 million. Since the population of British Columbia is approximately 3.5 million, one canquite reasonably conclude that the suspect was present during the assault in order to be splatteredwith blood from the victim. This evidence was never brought to trial because the ownership of thesweat pants could not be proven. Initially, the brother of the suspect had claimed that the sweatpants belonged to the suspect but he changed his mind at a later date.116Figure 24: Case II - Locus D16S85Genomic DNA samples were probed with 3’HVR a-globin to detect alleles at the D16S85locus.Lane m = molecular length standards1 = control DNA sample with known alleles - 5.0 ug (female)2 = “ “ (male)115 = genomic DNA sample from suspect’s pants - all1st 116 = genomic DNA from victim’s blood sample - 5.0 ug2nd 116 = genomic DNA from victim’s blood sample - 2.5 ug[3 days exposure to Kodak X-MAT AR film at -70°C withscreens]Figure 24: Case II - D16S8588-50CONm 1 2ALPHA - GLOB INSP VB115 116 116117m*1./3 days118Figure 25 : Case II - Locus D2S44Genomic DNA samples were probed with YNH24 to detect alleles at the D2S44 locus.Lane m = molecular length standards1 = control DNA sample with known alleles - 5.0 ug (female)2 (male)115 = genomic DNA sample from suspect’s pants - all1st 116 = genomic DNA from victim’s blood sample - 5.0 ug2nd 116 = genomic DNA from victim’s blood sample - 2.5 ug[3 days exposure to Kodak X-MAT AP film at -70°C withscreens]Note: Molecular length markers are not visible in this autoradiograph.Since the same membrane was probed each time, markers visible on theD16S85 autoradiograph (Figure 21) was used for extrapolation of DNAsizes.Figure 25 Case II - D2S4488-50 YNH24Sa9119mVBeSSISSSSa*__0aS120This analysis was also performed in 1988 with the restriction endonuclease Pvu II. Hence,comments from case I also apply in this case.CASE III - Hair Analysis (1989)Background Information:A doctor was accused of having sexual relations with a female patient. The female patienthad charged the doctor with this impropriety at the conclusion of the affair. She had apparentlycollected a number of pubic hairs which were claimed to be from the doctor and proof of theassault. Conventional hair fibre analysis failed to conclusively implicate the accused doctor. DNAtyping was sought as a last resort at identification of the 16 pubic hairs.Method (additions to that already mentioned in methods section)The pubic hairs had been mounted onto microscope slides for conventional testing and hadto be place in an 80°C xylene bath for 10 minutes to remove the cover slip and mounting medium.The nine extracted hairs were rinsed with 95% ethanol (to remove traces of xylene which mayinterfere with the typing procedure) and the two ends of each hair were cut-off and incubated inextraction buffer. Both ends of each hair were used to ensure that every hair shaft (which mayhave attached cells) was included. Microscopic examination showed that the unknown hair shafts(lower end) and follicles had degenerated, possibly due to the age of the hair and/or to the effectsof the mounting on slides. Because of the dehydrated state of the hair bulbs, they were incubatedin extraction buffer for 7 days instead of the usual overnight incubation. The remaining proceduresfor DNA extraction are as described previously (p27: protocol C).Five day old plucked hair shaft ends from laboratory personnel were included in theanalysis as controls for the extraction procedure of hair shafts. Half of these hairs (16 hairs) weretreated with a 80°C xylene bath to serve as a control for the effects of xylene on the typngprocedure.121DNA Typing Results:The most sensitive probe, 3’HVR a-globin, was used to determine whether a DNA patternwas obtainable for the hair DNA. Unfortunately, no DNA bands were observed for the unknownhair sample (Figure 26- lane 2) after a four day autoradiographic exposure and did not appeareven after a 14 day exposure (autoradiograph not shown). The control blood and hair DNAsamples each produced two DNA bands with the lower band appearing quite faint in the print ofthe autoradiograph. The absence of band(s) in the unknown sample lane may be due to thequantity and/or quality of DNA. The entire sample (DNA extracted from cells attached to 16 hairshafts) is present in lane 2 (Figure 26) and although fluorometric readings indicate that theequivalent of approximately 0.16 ug of DNA is present, this may actually be an artifact due tofluctuations of the fluorometer when dealing with such a dilute concentration of DNA (0.009ug/ul). Xylene treatment of DNA did not affect the DNA typing procedure as indicated by theequivalent amount of band intensities of lanes 3 and 4 (Figure 26). However, the amount of DNArecovered per hair bulb was noticeably less for the xylene treated control hair samples (Table 12).The recovery of DNA from the unknown hair samples was extremely low and could be due to anumber of things such as:1) degradation of DNA as a consequence of age2) treatment with mounting media3) treatment with xylene4) small quantities of cells attached to lower end of hair shaft a result of shedding asopposed to plucking.Although this analysis was not successful, it does demonstrate the applicability of DNAtyping to hair fibre analysis as a valuable means of identification. Furthermore, the complexity ofthe affects of various components and substances on the entire typing procedure including DNAextraction, suggests that this system is not immune to problems and requires much more researchand development.122Figure 26: Case III - Locus D16S85Genomic DNA samples were probed with 3’HVR a-globin to detect alleles at the D16S85locus.Lane 1 = control A (female) DNA sample from blood - 0.5 ug2 = DNA from unknown hair sample - all3 = control B (female) DNA sample from hair treated with xylene0.5 ug (1/6 of extract from 16 hairs)4 = control B (female) DNA sample from hair with no xylene treatment -0.5 ug (1/9 of extract from 16 hairs)5 = DNA from doctor’s blood sample (male) - 0.5 ug6 = control B (female) DNA sample from blood - 0.5 ug7 = control C (male) DNA sample from blood - 0.5 ug[4 days exposure to Kodak X-MAT AR film at -70°C with screens]t’3t)124Table 12: Recovery of DNA From Cells Attached to Hair ShaftsSample Age Method of Treatment Extracted DNARemoval (ng/hair)Control A 5 days pluckedno treatment 275xylene 180Unknown Unknown 11.2(>> 5 days) shed-mounting media-xylene125DISCUSSIONThe three forensic cases presented in this section serve to illustrate a few of the possibleapplications of DNA genetic typing. In cases 1 and 2, no mention was made of “matching” criteria(5.5% of fragment size) as determined by measurement imprecision within gels. The declared“matches” were only made visually (subjectively), without the objective boundaries describedpreviously (section I - measurement imprecision). This was because these cases were analyzed withthe restriction endonuclease Pvu II system (which also uses different size gels and molecularlength standards) and the measurement imprecision experiments have not been and will not beperformed by this laboratory. Once the switch was made to the Hae III system, it was morereasonable to fully concentrate on that system as the Pvu II system was not compatible with otherNorth American Forensic laboratories. However, the calculated values of identity should not besignificantly changed since the sizes of the bins for Pvu II system were greater than 5% of thefragment size.The number of details to consider in DNA analysis of forensic specimens is enormous andthe unknowns encountered in case III give an idea of the types of questions that have to beanswered. For example, the effects of aging on DNA stability of biological samples is of suchsignificance that, many laboratories have devoted much work to analysis of aged specimens withrespect to DNA analysis for both sex typing and individual profiling.Human sex determination of one week old bloodstains (equivalent to 10 ul of whole blood)by dot hybridization with a radioactive labelled Y-specific probe (pY3.4) was reported by Tyler etal. [119]. Furthermore, sex determination with a 4 year old blood stain (equivalent to 100 ul wholeblood), a single hair, 5 ul whole blood and 2 ul semen was demonstrated with radioactive labelledprobe pHY2.l, using either dot or blot hybridization [15]. More recently, Yokoi and Sagisaka haveshown that another Y-speciflc radioactive labelled probe (pHYlO) provides a more rapid (two day)dot hybridization sex determination test for blood and bloodstains [1201. The sensitivity of this126radioactive labelled pHY1O probe is at a range of 1 ul of blood and is 30-50 times greater than thatof its photobiotin labelled counterpart.With Jeffreys’ minisatellite probe 33.15, Gill et al. have obtained DNA fingerprints frombloodstains up to 4 years old, 11 week old semen stain and hair roots [11]. However, blood andsemen stains stored at room temperature for more than four years were found to be not typeableby this same laboratory [16]. Sperm DNA from a single vaginal swab taken at up to 36 hours postintercourse, have been typed by probe 33.15 [16]. Contamination of sperm DNA with vaginal cellDNA was eliminated by using a differential cell lysis method which preferentially lyses female cells[11]. Successful DNA typing with hypervariable single locus probes have been accomplished for 28day bloodstains stored at room temperature and 2 to 3 year old stains stored in a cold room [12].Studies have indicated that hypervariable single locus probes can also be used to successfully typesemen stain DNA and sperm DNA isolated from vaginal swabs [13]. Additionally, DNA extractedfrom human bone has found to be suitable for DNA typing and DNA yields are approximately 10to 20 fold higher in spongy tissue than in compact tissue (based on weight of tissue) [121].Other than the quantity of DNA, the major determinant of the success of a DNA typinganalysis of a forensic specimen is the quality of DNA. The presence of extractable DNA from agedspecimens was demonstrated by the cloning of a 3.4 kb DNA fragment from a 2,400 year oldEgyptian mummy [122]. However, agarose gel electrophoresis indicated that the most of the DNAwas < 0.5 kb with only a minor portion of DNA> 5.0 kb. The rate of postmortemautodegradation of DNA in human rib bone was concluded to increase with increasing humidityand temperature of storage environment [123]. An extensive study of postmortem stability of DNAin human organs by Bar et a!. demonstrated that the amount of degraded DNA correlated with thelength of the postmortem period [124]. However, case histories indicated that this DNAdegradation was also influenced by the environmental temperature at the site of death and/or thepresence of infectious disease prior to death. DNA from different organs were stable for variouslengths of time. Liver DNA was completely digested 24-36 hours postmortem, and DNA was stable127for up to five days in the kidney, spleen, and thyroid glands. Greatest stability was found in organssuch as the brain, psoas muscle and lymph nodes. Blood DNA gave conflicting results with goodstability in some cases and poor in others. This variability is thought to be a result of nonhomogenous sampling of blood clots. Blood clots, even from decayed bodies, produced good yieldsof high molecular length DNA. Although environmental conditions such as humidity andtemperature play a critical role with regards to the quantity and quality of DNA recovered fromforensic specimens, there are undoubtedly other contributing environmental factors. The effects ofultra-violet radiation, rain, air pollution, soil, synthetic dyes (in clothes), oil, and gasoline, must allbe taken into consideration and some of these issues will be discussed further in the Addendum.128CONCLUSIONThe application of DNA genetic typing in forensic science has enormous potential.However, the nature of forensic specimens makes this system of analysis complex and it isimportant to determine the possible affects of a wide range of variables on the success, validity andreliability of the analysis. Results of such testing have tremendous implications on the lives of thevictim and the accused in cases of assault. Although research in the last few years has been veryinformative, there is still an obvious need for continuing research and development of this system.129CHAPTER 2DIMORPHIC RES ANALYSISSection I - Data Base Generation- Reliability StudiesSection II - Discriminating Power Studies130INTRODUCTIONOne of the greatest limitations of DNA technology as it applies to forensic work is thequantity and quality of DNA available for analysis. With the application of thermostable Taq DNApolymerase for use in DNA synthesis in the polymerase chain reaction (PCR) [125], the meansbecame available to eliminate or substantially reduce these problems. The polymerase chainreaction allows enzymatic amplification of minute amounts of specific nucleotide sequences bymore than a million fold in copy number (Figure 27). This powerful technique has simplified manyworking protocols in molecular biology and enabled genetic manipulations to be carried out withtrace amounts of starting material. Amplification of nucleic acids has been applied in molecularmedicine and has resulted in more efficient diagnostic tests for various genetic diseases [92,126-128]and more sensitive tests for the detection of microbial pathogens [129,130].Currently, the polymerase chain reaction is being used to amplify polymorphic VNTRregions [131433] which may be used for individual identification purposes. With this type ofanalysis, it is possible to complete the identification procedure within 24 hours [1311. In contrast,the current procedure used for non-PCR identification may require 6 to 8 weeks. Although PCRidentification can be performed more quickly using minute amounts of DNA and is less labourintensive, several other problems previously discussed involving VNTRs still exist. Furthermore,the VNTR.PCR procedure has an obstacle of its own that may limit its application in forensicscience. Since a hypervariable locus may have 50 to 100 alleles, it will be impossible to predict theallelic sizes of the amplified DNA fragments. Hence, nonspecific and/or unfaithful amplificationproducts may be erroneously identified as true alleles leading to invalid interpretations.In light of problems with methodology, interpretation, and statistics of RFLP analysis usingthe VNTR (or VNTR-PCR) system, it seemed appropriate to search for alternate strategies forRFLP analysis. This eventually lead to the idea of DNA profiling with PCR amplification ofRFLPs created via a simpler mechanism, by utilizing point mutations (Figure 28). Such pointmutations which could result in either the elimination or the creation of a restriction endonucleaseFigure 27: Polymerase Chain Reaction131a.L. TargetSequenceCDi Denature and anneal primersI—NA—a)-- Primer extensioni Denature and anneal primers0’-•EV,-.— p—---Ahiiiiiiiiimiiiii01111111111i1111111 111Ipa)Primer extension-oCD• Denature and anneal primers—pa)Primer extension__IIII’GfLo Short productI I I I I I I I‘‘‘‘ I I o Short product132TaqI TaqI__________I__________800bp‘p.!!IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIITaq I1“i’tH’fl’H”H+ Taq I enzyme[.4—. 300 bp—“HiiiiiiiiiiiiiiiIISOObp liiiFigure 28: Amplification of a Dimorphic Locus[..m— 500bp i.J.4 300bp -Taq I133recognition site (RES), will have only two possible alleles (dimorphic). Thus, for a particular RESthat is located in a specific chromosomal pair, an individual may possess three genotypes:homozygous negative, homozygous positive, and heterozygous.The presence or absence of a specific RES may be considered as two alleles of one specificlocus. If the alleles of two particular loci (“X” and “Y”) were unlinked, different combinations ofthe alleles may occur leading to nine possible genotypes (eg. XXYY, XxYY, xxYY, XxYy, Xxyyetc.) in the population. If a number of such unlinked genomic loci are available, analyses of theirrespective allelic frequencies may provide a potential system for DNA profiling.Amplification of such loci with PCR solves DNA quantity problems and simplifies the entire typingprocedure. In contrast to the VNTR system which identifies individuals according to the sizes ofthe DNA fragments (alleles), this dimorphic RES system identifies by comparing specific DNAsequences.Polymerase chain reaction amplification of five dimorphic RES loci is analyzed to collectallelic frequency data bases for these loci and to also determine their potential for generation ofdiscriminating DNA profiles. The five loci are: adenine phosphoribosyltransferase gene (APRT),lipoprotein lipase gene (LPL), prealbumin gene (PALB), adenine deaminase gene (ADA), andcarbonic anhydrase II gene (CAH).134METHODSPopulation Samples:Undefined population sample are as described in chapter 1 for VNTR analysis (p38-39).The “Oriental” population was also taken from the same group as that for the undefined populationbut with an additional stipulation of an “Oriental-like” surname (i.e. Wong, Lee, Nyen). To reducethe likelihood of “Non-Orientals” with “Oriental-like” surnames as a result of inter-racial marriages,the females of this group also possessed an “Oriental-like” first name (i.e. Mae-Ling, Su Lan).Extraction of DNA from liquid blood for PCR analysis:DNA was extracted from liquid blood samples as described in protocol A and B for VNTRanalysis (chapter 1 - methods).Quantitation of DNA:Refer to chapter 1 - methods sectionLoci With Dimorphic RES:The selection criteria for loci were that each should contain an unique dimorphic RES1) with allelic frequencies of as close to 50-50 as possible and not moredivergent than 70-302) with available sequence information to reduce work load3) which is known to be not associated with a disease.Oligonucleotide Primers:The specific primers (two for each locus) for the APRT, PALB, ADA, CAH and LPL lociwere designed by selecting oligomeric sequences (20 bases) from published sequence data [134-138]. The selection criteria for the sequences were that each should:1) flank a polymorphic RES and the distance separating each primer from the RES be <0.70 kb1352) contain only a single cleavage site for the specific restriction endonuclease in theselected region3) exhibit high specificity in terms of sequence homology with the locus of interest (lowaffinity with other loci)4) not contain long stretches of polypurines or polypyrimidines.The respective sequences selected for the construction of the primers were:a) APRT 1 : 5’-TFGTGAGATrGAGCCCCCGA- 3’APRT 2: 5’-AGCTGAAAGGCCAGTGACAT- 3’b) PALB 1 : 5’-TCTCAGCCTGATGGGAACCA- 3’PALB 2: 5’-GATCTGAGTGAGCCATACCA- 3’c) ADA 1 : 5’-AGCAGCCAGCCAGTAAAATG- 3’ADA 2: 5’-AGCGTATCCTCCTCTTCCAA- 3’d) CAN 1 : 5’-CACAGTGTCTCAGGTCCACA- 3’CAH 2: 5’-AAGGGCTCAGTCTCTGGTGT- 3’e) LPL 1 : 5’-AAAGGAATGGTCGGAAAATG- 3’LPL 2: 5’-GTTTGAGACTAGCTTGGCCA- 3’The APRT, primers were synthesized by the Biochemistry Department at the University of BritishColumbia, the PALB primers were synthesized by Pharmacia and the ADA, CAl-I and LPL primerswere synthesized by this author using the PCRMate in the Pathology Department (MedicalMicrobiology) at the University of British Columbia.When these primer sets were used individually in the PCR to amplify the APRT, PALB,ADA, CAN, and LPL loci, a 0.73 kb, 1.01 kb, 0.96 kb, 0.72 kb, and 1.03 kb product fragmentrespectively would be expected. The polymorphic restriction endonuclease recognition sites in thefive loci were analyzed (Figure 29).136Figure 29: Dimorphic RES LociAPRT Locus PALB LocusAPRT 2 PALB 213’——— 5’ 3’——— 5’5’——— 3’ 5’___ 3’I IAPRT 1 Taq I Pi 1 Fnu 4H10.48kb 0.25kb-0.36 kb 0.65kb0.73kb 1.01 kbADA Locus CAH LocusADA2 CAH21 13’———5’ 3’ ———5’5’——— 3’ 5’ ——— 3’ADAI CAH1iaq-0.37 kb 0.59 kb 4 0.50kb 4 0.22 kb--4 0.96kb 4 0.72kbLPL LocusLPL 23’———5’LPL1 Pll‘oligonucleotideprimers (20 bases)4—0.46 kb I’ 4 0.57 kb-‘4 1.03kb137PCR for the APRT Locus From Genomic DNA Extracted From Blood:DNA samples from undefined individuals from the population of the Greater VancouverArea, were selected. Two hundred nanograms of each genomic DNA sample were added to 0.3uM of each of the two APRT primers, 1.7 mM of MgCl2,200 uM of each of the fourdeoxynucleotide triphosphates (dNTPs) (Perkin Elmer Cetus/Pharmacia), and 1.25 units of TaqDNA polymerase (Perkin Elmer Cetus/BRL), in a total volume of 50 ul. An equal volume ofmineral oil was layered onto the aqueous solution. The PCR was carried out in a thermal cycler(Perkin Elmer Cetus) for 31 cycles with an initial (before the first cycle) denaturation of 3 minutesat 94°C. The reaction cycle conditions were: 94°C (30 seconds) for denaturation, 61°C (30seconds) for primer annealing, 72°C (60 seconds) for extension, and with an auto-extension periodof 3 seconds (at 72°C) at the end of each cycle. In the later experiments, the denaturationtemperature was increased to 95°C and the cyclic denaturation time was increased to either 40 or60 seconds.PCR For the PALB Locus From Genomic DNA Extracted From Blood:Two hundred nanograms of genomic DNA from the same samples described above, orgenomic DNA from Orientals, were added to 0.2 uM of each of the two PALB primers, 2.0 mM ofMgC12,200 uM of each of the four dNTPs, and 1.25 units of Taq DNA polymerase in a totalvolume of 50 ul. The PCR cycle conditions were carried out as described for the APRT locus.PCR For the ADA Locus From Genomic DNA Extracted From Blood:Two hundred nanograms of genomic DNA from the same samples described for the APRTlocus, or genomic DNA from Orientals, were added to 0.3 uM of each of the two ADA primers,3.0 mM MgC12,200 uM of each of the four dNTPs, and 1.25 units of Taq DNA polymerase in atotal volume of 50 ul. The PCR cycle conditions were carried out as described for the APRTlocus.138PCR For the CAN Locus From Genomic DNA Extracted From Blood:Two hundred nanograms of genomic DNA from the same samples described for the APRTlocus, or genomic DNA from Orientals, were added to 0.3 uM of each of the two CAl-I primers,1.5 mM MgC12,200 uM of each of the four dNTPs, and 1.23 units of Taq DNA polymerase in atotal volume of 50 ul. Equal volume of mineral oil was layered onto the aqueous solution. ThePCR was carried out in a thermal cycler (Perkin Elmer Cetus) for 31 cycles with an initial (beforethe first cycle) denaturation of 3 minutes at 95°C. The reaction cycle conditions were: 95°C (40seconds) for denaturation, 61°C (30 seconds) for primer annealing, 72°C (60 seconds) forextension, and with an auto-extension period of one second (at 72°C) at the end of each cycle. Inlater experiments, the cyclic denaturation time (at 95°C) was increased to one minute.PCR For the LPL Locus From Genomic DNA Extracted From Blood:Two hundred nanograms of genomic DNA aliquoted from the same samples described forthe APRT locus, were added to 0.17 uM of each of the two LPL primers, 4.0 mM MgC12,200 uMof each of the four dNTPs, and 1.23 units of Taq DNA polymerase in a total volume of 50 ul. Anequal volume of mineral oil was layered onto the aqueous solution. The PCR was carried out in athermal cycler (Perkin Elmer etus) for 31 cycles with an initial (before the first cycle)denaturation of 3 minutes at 95°C. The reaction cycle conditions were: 95°C (40 seconds) fordenaturation, 59°C (30 seconds) for primer annealing, 72°C (60 seconds) for extension, and withan auto-extension period of one second (at 72°C) at the end of each cycle. In the laterexperiments, the cyclic denaturation time (at 95°C) was increased to one minute the primerannealing temperature was increased to 61°C and the MgCl2 concentration was decreased to3.0 mM.Restriction Endonuclease Digestion of the PCR Products and Gel Electrophoresis:The PCR products were precipitated with 2 volumes of 95% ethanol and 1/10 volume of0.3 M sodium acetate (pH 5.2). The resulting pellet was resuspended in 12 ul of sterile water and139separately digested with the appropriate restriction endonuclease. PCR products obtained from theAPRT and CAH loci were digested with Taq I, PCR products obtained from the PALB locus weredigested with Fnu 4111, PCR products obtained from ADA locus were digested with Pst I, and PCRproducts obtained from the LPL locus were digested with Pvu II. The digest conditions were thoserecommended by the suppliers of the respective restriction endonucleases. DNA fragments in thedigests were separated by electrophoresis through 1% agarose gel (containing 1 ug/ml ethidiumbromide) in 1X TBE buffer (0.089 M Tris-borate, 0.089 M boric acid, 0.025 M EDTA, pH 8.3) orin 1X TAE buffer (0.04 M Tris-acetate, 0.002 EDTA). Molecular length standards included inPCR electrophoresis gels were a combination of lambda-Hind III/4X174 RF-Hae III (BRL).Control For PCR Amplification:The conditions for PCR amplification of the five loci were established using one stockDNA sample. This stock sample was used in each subsequent PCR run as a control to confirmPCR conditions. Since contaminated DNA would be present in extremely small quantities relativeto the large amount (200 ng) of test sample DNA, negative controls were not included in all of thetests.Control For the Activity of the Restriction Endonuclease:Plasmid DNA fragments containing an unique restriction endonuclease recognition site(RES) were included as controls to confirm the activity of the restriction endonuclease used ineach digest reaction of the PCR products. To avoid complicating the zygosity patterns, theseplasmid fragments were chosen so that the digested plasmid fragments were either larger or smallerthan the undigested or digested PCR products. Two different Taq I control fragments were usedfor the two loci (APRT/CAH) with the polymorphic Taq I RES. The control fragment (0.42 kb)for the APRT locus was prepared from pBR322 and digestion of this fragment with Taq I results intwo fragments (differing in length by 10 bp) that appear as one band at the 0.21 kb range. Thecontrol fragment (2.48 kb) for the CAFI locus was also prepared from pBR322 and digestion of this140fragment with Taq I results in two fragments (differing in length by 60 bp) that appear as bands inthe 1.27 and 1.21 kb range. The only reason for having two Taq I control fragments was that thesmaller control fragment was prepared initially and since it appeared to be easier to visualize largerfragments, another control fragment was made for the CAH locus. The control fragment (0.52 kb)for the PALB locus was prepared from pUC 19 and digestion of this fragment with Fnu 4H1results in two fragments (differing in length by 60 bp) that is observed as bands at the 0.27 and 0.21kb range. The control fragment (2.77 kb) for the ADA locus was prepared from pBR322 anddigestion of this fragment with Pst I results in two fragments (differing in length by 30 bp) thatappears as one band at the 1.40/1.37 kb range. The control fragment (3.05 kb) for the LPL locuswas also prepared from pBR322 and digestion of this fragment with Pvu II results in two fragments(differing in length by 30 bp) that appear as one band at the 1.54/1.51 kb range.Reliability of PCR Amplification:The reliability of the amplification by PCR was monitored by one or more of the followingmethods:1) DNA sequencing of the oligo primers2) probing of the amplified products with a DNA probe that is homologous to the targetlocus3) correlating the number and sizes of the restriction fragment(s) that were obtainedexperimentally with that of the calculated size.Transformation of Recombinant Plasmids into Competent E.Coli (DH5a) Cells:Recombinant plasmids containing inserts which probe for the amplification products at theAPRT, ADA, CAH and LPL loci were transformed into E.coli (DH5a) cells for amplification.Refer to chapter 1 - methods sectionSmall Scale Amplification and Isolation of Plasmid DNA:Refer to chapter 1 - methods section141Excision of Foreign DNA Insert From Recombinant Plasmids:Refer to chapter 1 - methods sectionLabelling and Purification of DNA Probes:Refer to chapter 1 - methods sectionHybridization of32P-DNA probes to membrane-fixed DNA:Refer to chapter 1 - methods sectionPost hybridization washesRefer to chapter 1 - methods sectionProbe removal from membrane:Refer to chapter 1 - methods sectionSpecificity of the Oligonucleotide Primers for Human DNA:Attempts were made to amplify non-human DNA in order to establish the specificity of theoligonucleotide primers. The different types of DNA used were: 1) bacterial DNA (E.coli.,C.perfrmnens, P.syringe, and B.subtilis), 2) yeast DNA (C.albicans), 3) plasmid DNA (pACYC184), 4) mouse DNA (cell line 3T3) and 5) DNA extracted from blood of a dog, cat, and chicken.Amplification of these DNA was allowed to proceed for at least 35 cycles (cf. 31 cycles for DNAextracted from human blood).142SECTION IRESULTS AND DISCUSSIONReliability of the PCR (one primer set I reaction):Analysis of the electrophoretogram showed that only one band could be detected in each ofthe amplified DNA products and their respective sizes were as expected, j 0.73 kb for the APRTlocus, 1.01 kb for the PALB locus, 0.96 kb for the ADA locus, 0.72 kb for the CAH locus and 1.03kb for the LPL locus. The results of the restriction endonuclease digestion of the respective PCRproducts agreed with that of theoretical calculations (Figures 30-34). The zygosity of each locus foran individual was determined by referring to the number and sizes of the restriction fragmentsgenerated in each case (Table 13). Individuals that are homozygous for the absence of the RES ata specific locus will show only one DNA band (locus DNA fragments intact) while those who arehomozygous for the presence of the RES, will show two bands (result of complete cleavage of alllocus DNA fragments at the respective RES).Individuals that are heterozygous for the restriction endonuclease site will show threebands. One of these bands represents intact locus DNA fragments indicating that the enzyme siteis absent in one chromosome. The other two bands represent cleaved DNA fragments indicatingthat the particular endonuclease enzyme site is present in the second chromosome.In addition to correlating the number and sizes of the restriction fragment(s) that wereobtained experimentally with that of the calculated size for each of the five loci analyzed, thespecificity of the PALB amplification was ensured by DNA sequencing of the two oligo primers bytheir manufacturers (Pharmacia). Alternately the specificity of the APRT, ADA, CAH, and LPLamplifications were confirmed by hybridization of the amplification products of each loci with theirrespective genomic DNA probes. The Southern blots used in each hybridization contained theamplification products of all five dimorphic loci (Figure 35). Specific hybridizations (a single bandcorresponding to the calculated size) were obtained with the APRT (Haupl5), CAH (H25-3.8), and143Figure 30: APRT LocusHuman genomic DNA was amplified at the APRT locus and the amplified material (0.73 kb)was digested with restriction endonuclease Taq I. A DNA fragment (0.42 kb) with arecognition site for Taq I was included in the digest reaction to serve as an internalcontrol for the activity of the restriction endonuclease.Lane 1 = Homozygous positive (0.48 and 0.25 kb fragments)2 = Heterozygous (0.73, 0.48, and 0.25 kb fragments)3 = Homozygous positive4 = Homozygous positive5 = Undigested control fragment (0.42 kb)6 = Taq I digested control fragment (0.21 kb)0.730.480.420.250.21144Figure 31 : LPL LocusHuman genomic DNA was amplified at the LPL locus and the amplified material (1.03 kb)was digested with restriction endonuclease Pvu II. A DNA fragment (3.05 kb) with arecognition site for Pvu II was included in the digest reaction to serve as an internalcontrol for the activity of the restriction endonuclease.Lane m = molecular length standards1 = Undigested control fragment (3.05 kb)2 = Pvu II digested control fragment (1.54/1.51 kb)3 = Homozygous positive (0.57 and 0.46 kb)4 = Heterozygous (1.03, 0.57, and 0.46 kb)5 = Homozygous negative (1.03 kb)Note: The 0.57 and 0.46 kb fragments are fainter than that seen in the original gelbecause the print is a picture of a picture.(3.O52.32(ii1.54/1.511.350.57o.ooL145Figure 32: PALB LocusHuman genomic DNA was amplified at the PALB locus and the amplified material (1.01 kb)was digested with restriction endonuclease Fnu 4111. A DNA fragment (0.52 kb) with arecognition site for Fnu 4111 was included in the digest reaction to serve as an internalcontrol for the activity of the restriction endonuclease.Lane M = molecular length standards1 = Undigested control fragment (0.52 kb2 = Fnu 4111 digested control fragment (0.27 and 0.21 kb)3 Homozygous positive (0.65 and 0.36 kb)4 = Heterozygous (1.01, 0.65, and 0.36 kb)5 = Homozygous positive1.35>.:.1.O10.87>.O.650.60>. <J0.520.36— 0.270.21146Figure 33 : ADA LocusHuman genomic DNA was amplified at the ADA locus and the amplified material (0.96 kb)was digested with restriction endonuclease Pst I. A DNA fragment (2.77 kb) with arecognition site for Pst I was included in the digest reaction to serve as an internalcontrol for the activity of the restriction endonuclease.Lane m = molecular length standards1 = Undigested control fragment (2.77 kb)2 = Pst I digested control fragment (1.40/1.37 kb)3 = Homozygous positive (0.59 and 0.37 kb)4 = Homozygous negative (0.96 kb)5 = Heterozygous (0.96, 0.59, and 0.37 kb)Note: The 0.37 kb fragment is fainter than that seen in the original gelbecause the print is a picture of a picture.2.321.35o.6oc2.771.40 /1.37O.960.590.37m12345Figure 34: CAH LocusHuman genomic DNA was amplified at the CAH locus and the amplified material (0.72 kb)was digested with restriction endonuclease Taq I. A DNA fragment (2.48 kb) with arecognition site for Taq I was included in the digest reaction to serve as an internalcontrol for the activity of the restriction endonuclease.Lane m = molecular length standards1 = Undigested control fragment (2.48 kb)2 = Pst I digested control fragment (1.27/1.21 kb)3 = Homozygous positive (0.50 and 0.22 kb)4 = Heterozygous (0.72, 0.50, and 0.22 kb)5 = Homozygous negative (0.72 kb)1472.321.35o.6o><2.481.21O.22Table13:CalculatedSizesSENSITIVITYTORESTRICTIONENZYMESIZEOFLOCUSENZYMEAMPLIFIEDHOMOZYGOUSHETEROZYGOUSHOMOZYGOUSREGION(1band) -I-(3bands)-1+(2bands)+1+0.73kb0.48kbAPRTTaqI0.73kb0.73kb0.48kb0.25kb0.25kb1.01kb0.65kbPALBFnu4H11.01kb1.01kb0.65kb0.35kb0.35kb0.96kb0.59kbADAPstI0.96kb0.96kb0.59kb0.37kb0.37kb0.72kbCAHTaqI0.72kb0.72kb0.50kb0.50kb0.22kb0.22kb1.03kbLPLPvuII1.03kb1.03kb0.57kb0.57kb0.46kb0.46kb00149Figure 35 : Amplification Products of Five LociHuman genomic DNA was amplified at the APRT, PALB, ADA, CAl-I and LPL loci.Lanes M and m = different quantities of molecular length standards1 = amplification products of APRT locus2 = PALB3 =I? It ADA4 = H CAN5 = LPL1.35O.87o.6oMmi150LPL (HLPLG18) DNA probes (Figure 36) which indicate that the appropriate regions of thegenome have been amplified. The ADA (ADAN16) hybridized extremely well with the ADAamplification products (Figure 36 - lane 3) but the probe also hybridized weakly to the APRTamplification products (Figure 36 - lane 1). This shows that although the correct locus (ADA) hasbeen amplified, the ADA genomic probe has a region that is partially homologous to the amplifiedAPRT region.It is of prime importance to include a DNA fragment of known size containing one specificrestriction endonuclease site as an internal control to assess the activity of the restrictionendonucleases (Figure 30-34). Addition of this control is necessary because the results obtainedfrom an inactive or partially active restriction endonuclease might affect the subsequent zygosityassignment. Thus, by referring to the restriction fragments generated from the internal controlDNA, anomalous restriction endonuclease activity can be detected. Accuracy of the results andappropriate interpretation of the data are thereby ensured.Another important consideration is the fidelity of the Taq I DNA polymerase. It isestimated that the frequency of base misincorporation by the polymerase is approximately 1/400[125]. A base substitution resulting from any misincorportion event will only be detectable if thisevent occurs during the first few cycles and if the initial copy number is low (less than 10) [139].Furthermore, since these are random misincorporations, the amount of PCR products that containa misincorporated base within dimorphic RES will be negligible (i.e. either the creation orelimination of the RES).A final point to consider is the possibility that heterozygous individuals could be incorrectlycategorized as homozygous negative individuals if the smaller bands were not visible as a result oflow quanitities of PCR products. This would only occur if both of the two smaller bands(representing the digested fragment) were not visible. The presence of even one of the twofragments would infer the presence of the other small fragment. The fluorescence of the larger ofthe two small bands would be always at least 0.5 times as intense as the largest undigestedFigure 36: Probing of Amplification ProductsLane 1 = amplification products using APRT primers2=3=4=5=PALBADACAHLPLTop left panel : probed for the APRT locusTop right panel : probed for the LPL locusBottom left panel : probed for the ADA locusBottom right panel: probed for the CAH locus12345ADAH1511 2 34 512345CAH152fragment. To ensure against such classification mistakes one may wish to consider southernblotting and hybridization techniques for checking low PCR product samples.Allelic Frequencies of the Dimorphic Restriction Endonuclease Recognition Site at the APRTLoci:The zygosity frequency distribution of a specific dimorphic restriction endonucleaserecognition site (RES) in the second intron of the adenine phosphoribosyltransferase (APRT) locuswas studied with genomic DNA samples that were obtained from an undefined population from theGreater Vancouver Area. The frequency values for the APRT locus were 75% Taq I site positiveand 25% Taq I site negative (Table 14), and agreed with the previous report [140]. Chi squareanalysis [141] of the observed genotypes indicate that the dimorphic variation at the Taq I site ofthe APRT locus is in Hardy-Weinberg Equilibrium and that such frequency data should remainrelatively stable in the absence of disturbing forces.Allelic Frequencies of the Dimorphic Restriction Endonuclease Recognition Site at the LPL Loci:The zygosity frequency distribution of a specific dimorphic restriction endonucleaserecognition site (RES) in the sixth intron of the lipoprotein lipase (LPL) locus was studied withgenomic DNA samples that were obtained from an undefined population from the GreaterVancouver Area. The frequency values for the LPL locus is 51% Pvu II site positive and 49% PvuII site negative (Table 15). These values are not significantly different from those of Li et al. [142]who obtained frequency figures of 62% site positive and 38% site negative. The observedgenotypes from this study indicate that the LPL locus for this population sample is in HardyWeinberg Equilibrium.Allelic Frequencies of the Dimorphic Restriction Endonuclease Recognition Site at the PALBLocus:The zygosity frequency distribution of a specific dimorphic RES in the third intron of theprealbumin (PALB) locus was studied with genomic DNA samples that were obtained from anTable14:APRTAllelicFrequenciesALLELICZYGOSITYFitnesstoHWEFREQUENCIESLOCUSNumberofHomozygousHomozygousRESRESx2peoplenegativeHeterozygoussitivenegativepositive(1df)tested(-I-)(-1+)(+1+)APRT1731263980.250.750.22>50%U2E(obs1,-exp)Z=“Goodnessoffit”ofdatatonullhypothesis=1expwhereobs,istheobservednumberofindividualswiththe“U’genotypeandexp,istheexpectednumberofindividualswiththe“U’genotypeP=theprobabilityofobtainingthepresentresultsifthenullhypothesis(H.W.E.)iscorrectP<5%rejecthypothesis(locusnotinequilibrium)H.W.E.=Hardy-WeinbergEquilibriumdf=degreesoffreedomCalculations:(n=173)Phenotypes+1++1-Expectedgenotypefrequency0.5620.3750.062Expected#ofindividuals97.264.910.7Observed#ofindividuals9863122=[(9897.2)2]+[6364.9)21+[12p7)2]=0.2201X97.264.910.7Table15:LPLAllelicFrequenciesALLELICZYGOSITYFitnesstoHWEFREQUENCIESLOCUSNumberofHomozygousHomozygousRESRESx2peoplenegativeHeterozygouspositivenegativeposive(1df)tested(-I-)(-1+)(+1+)LPL1524069430.490.511.29>10%2E(obsj,-exp.)X=“Goodnessoffit”ofdatatonullhypothesis=exp1,whereOb5jistheobservednumber ofindividualswiththe“ij”genotypeandexp1,istheexpectednumberofindividualswiththe“VgenotypeP=theprobabilityofobtainingthepresentresultsifthenullhypothesis(H.W.E.) iscorrect*P<5%rejecthypothesis(locusnotinequilibrium)H.W.E.=Hardy-WeinbergEquilibriumdf=degreesoffreedomCalculations:(n=152)Phenotypes+1++1--IExpectedgenotypefrequency0.2600.5000.240Expected#ofindividuals39.576.036.5Observed#ofindividuals4369402[(43-395)2]+{(69-76)2]+[(40-36.5)2]=1291X—7636.5155undefined population and a Oriental population of the Greater Vancouver Area. Individualselected for the Oriental population were chosen on the basis of their surnames. Although it ispossible that some selected females were not Orientals, the proportion of such individuals shouldbe low so that the allele frequencies should not be significantly altered from their true frequencies.Allelic frequencies obtained for a undefined population at the PALB locus were 66% Fnu 4H1 sitepositive and 34% Fnu 4111 site negative (Table 16), and chi square analysis of observed genotypesdemonstrated that the PALB locus is in Hardy-Weinberg Equilibrium. However, the allelicfrequencies obtained in this study were different from that reported by Yoshioka et al. [143]. Thesamples analyzed by Yoshioka’s report were from a Japanese population consisting of severalpatients with familial amyloidotic polyneuropathy (FAP) and 10 “normal” people. The allelefrequencies of both Japanese subgroups were 50% for the presence or absence of the RES whichsuggested that this variation was not responsible for FAP. The discrepancies between allelicfrequencies of this report and that of Yoshioka’s, could be due to the differences in ethnic originsof the selected sample population. As previously discussed in Chapter 1, it has been reported thatsignificant variations in the allelic frequency distribution of highly polymorphic loci occurred invarious ethnic groups, American Blacks, Caucasians and Hispanics [87,88]. To test thisassumption, the sample population was altered such that Orientals were excluded from theundefined sample and this group was re-categorized as the Non-Oriental group. The excludedOrientals were then combined with another Oriental test sample and this new combined group wasdesignated the “Oriental group”. The differences between the allelic frequencies of the two groupsare not significant at the 5% level. Chi square analysis of the genotypes also show that the PALBlocus for both groups are in Hardy-Weinberg Equilibrium (Table 17). Unless there are uniquedifferences within the Oriental subpopulation (j Chinese, Japanese, Vietnamese) which can notbe determined here, it is unlikely that the observed discrepancies of the allelic frequencies of thisstudy and that of Yoshioka et al. are due to ethnic differences. A more likely explanation is theTable16:PALBAllelicFrequenciesALLELICZYGOSITYFitnesstoHWEFREQUENCIESLOCUSNumberofHomozygousHomozygousRESRESx2peoplenegativeHeterozygouspositivenegativepositive(1dl)tested(-I-)(-1+)(+1+)PALB1931688890.340.662.09>10%22(obs,-exp.)X=“Goodnessoffit”ofdatatonullhypothesis=exp,whereobsistheobservednumberofindividualswiththe“V’genotypeandexpistheexpectednumberofindividualswiththe“i”genotypeP=theprobabilityofobtainingthepresentresultsifthenullhypothesis(H.W.E.) iscorrect*P<5%rejecthypothesis(locusnotinequilibrium)H.W.E. =Hardy-WeinbergEquilibriumdf=degreesoffreedomCalculations:(n=193)Phenotypes+1++1-Expectedgenotypefrequency0.4360.4490.116Expected#ofindividuals84.186.622.3Observed#ofindividuals8988162—[89-84.1)2]+[(88-86.6)2]+[(16-22.3)2J=2088X—84.186.622.3.Table17:SubgroupComparisonsatthePALBLocusZYGOSITYALLELICFITNESSFREQUENCIESTOH.W.B.PALBNumberofHomozygousHomozygousRESRES2PpeoplenegativeHeterozygouspositivenegativepositiveX(1df)tested(-I-)(-1+)(+1+)Undefined1931688890.340.662.09>10%Non-1801683810.350.651.98>10%OrientalsOrientals134959660.290.710.76>10%158difference in the sample size of the two reports j 193 individuals in this analysis versus < 20individuals in the Japanese analysis.Allelic Frequencies of the Dimorphic Restriction Endonuclease Recognition Site at the ADALocus:The zygosity frequency distribution of a specific dimorphic RES in the second intron of theadenosine deaminase (ADA) locus was studied with genomic DNA samples that were obtainedfrom an undefined population and a Oriental population from the Greater Vancouver Area. Foran undefined population, the allelic frequency values were found to be 73% Pst I site positive and27% Pst I site negative (Table 18). These results are comparable to those obtained by Tzall et al.who obtained frequency values of 67% and 33% for Pst I site positive and negative, respectively[144]. Chi square analysis indicate that the allelic frequencies of the two studies are not significantat the 5% level. However, in contrast with Tzall’s study, data obtained for an undefined populationin this study demonstrated that the ADA locus is not in Hardy-Weinberg Equilibrium. Furtheranalysis was performed for comparison with an Oriental subgroup. The Non-Oriental and Orientalgroupings were organized as described for the PALB locus. The allelic frequencies of the NonOrientals and the Orientals are not significantly different, but the chi square values are noticeablydifferent (Table 19). The smaller the chi square value, the smaller the difference between theobserved and expected values. Although the data from the undefined and Non-Oriental groupsdemonstrate non-equilibrium (P< 5%), the chi square value of the Oriental group (with P > 90%),indicate that the ADA locus in Orientals is definitely in equilibrium as there is no differencebetween the observed and expected values. The possibility exists that there are other subgroupswithin the Non-Oriental group which have different allelic frequencies and by themselves, maydemonstrate equilibrium at the ADA locus but deviate from equlibrium when their frequencies arepooled. The observed HWE in Tzall’s study may be attributed to their population sample whichconsisted of a more homogenous group i.e. 29 third generation Utah individual’s and 16 individualsof European extraction.Table18:ADAAllelicFrequenciesALLELICZYGOSITYFitnesstoHWEFREQUENCIESLOCUSNumberofHomozygousHomozygousRESRESx2peoplenegativeHeterozygouspositivenegativepositive(1dtested(-I-)(-1+)(+1+)*ADA175780880.270.734.70<5%2(obs,-exp.)2X=“Goodnessoffit”ofdatatonullhypothesis=1whereobsistheobservednumber ofindividualswiththe“i-”genotypeandexp,istheexpectednumber ofindividualswiththe“L”genotypeP=theprobabilityofobtainingthepresent resultsifthenullhypothesis(H.W.E.) iscorrect*P<5%reject hypothesis(locus notinequilibrium)H.W.E. =Hardy-WeinbergEquilibriumdf=degreesoffreedomCalculations:(n=175)Phenotypes+1++1--IExpectedgenotypefrequency0.5330.3940.073Expected#ofindividuals92.868.812.8Observed#ofindividuals888072=[(88-92.8)2}+[8o-68.8)2]+[7-12.8]=4.699—X92.868.812.8Table19:SubgroupComparisonsattheADALocusZYGOSITYALLELICFITNESSFREQUENCIESTOH.W.E.ADANumber ofHomozygousHomozygousRESRES2peoplenegativeHeterozygouspositivenegativepositiveX(1df)tested(-I-)(-1+)(+1+)Undefined175780880.270.734.65<5%Non-1616800.270.735.30<2.5%Orientals110639650.230.770.01>90%OrientalsI161Allelic Frequencies of the Dimorphic Restriction Endonuclease Recognition Site at the CAHLocus:The zygosity frequency distribution of a specific dimorphic RES located one kilobase fromthe 5’ end of the carbonic anhydrase II (CAH) locus was studied with genomic DNA samples thatwere obtained from an undefined population and a Oriental population of the Greater VancouverArea. For an undefined population, the CAH locus was found to be in Hardy-WeinbergEquilibrium with allelic frequencies of 58% Taq I site positive and 42% Taq I site negative (Table20). These frequency results are comparable to those obtained by Lee et a!. who obtained valuesof 50% for the presence or absence of the Taq I site [145]. Chi square analysis indicate that thetwo data sets are not significant at the 5% level.As with the PALB and ADA loci, comparison of allelic frequencies were made with anOriental sample group. The Non-Oriental and Oriental groupings were organized as described forthe PALB locus. Both groups demonstrate Hardy-Weinberg Equilibrium but their frequency valuesare reversed (Table 21). These frequency differences of the two groups are highly significant witha Chi square value corresponding to a P < 0.005. This suggests that there are subgroup differences(at least between Orientals and Non-Orientals) at the CAl-I locus but these differences do notdisturb equilibrium. The number of Orientals in the undefined group is small (11 out of 93), henceallele frequencies of the undefined group are identical to those of the Non-Oriental group.Specificity of Oligonucleotide Primers for Human DNA:Forensic specimens may be susceptible to contamination by a variety of agents, forexample: 1) yeast 2) bacteria and 3) body fluids of animals. Since such impure specimens mayproduce inaccurate results, the specificity of the oligonucleotide primers for human DNA is ofutmost importance. Each set of primers were used in attempts to amplify DNA from 10 differentsources of possible contaminants. The PCR was allowed to proceed for 35 cycles, which exceedsthat for PCR with human DNA (31 cycles), to ensure that amplified products would not be misseddue to low levels of amplification. Amplification products of a size appropriate for eachTable20:CAHAllelicFrequenciesALLELICZYGOSITYFitnesstoHWEFREQUENCIESLOCUSNumber ofHomozygousHomozygousRESRES2peoplenegativeHeterozygouspositivenegativepositive(1df)tested(-I-)(-1+)(+1+)CAH931843320.420.580.29>50%22E(obs4,-exp.)X=“Goodnessoffit”of datatonullhypothesis=exp,whereobsistheobservednumber ofindividualswiththe“tj”genotypeandexpj,istheexpectednumber ofindividualswiththe“1.”genotypeP=theprobabilityofobtainingthepresent resultsifthenullhypothesis(H.W.E.) iscorrect*P<5%rejecthypothesis(locus notinequilibrium)H.W.E. =Hardy-WeinbergEquilibriumdf=degreesoffreedomCalculations:(n=93)Phenotypes+1++1--IExpectedgenotypefrequency0.3360.4870.176Expected#ofindividuals31.345.316.4Observed#ofindividuals3243182—[(32-313)2}+ [43-453)2J+[(18-16.42}—0288X31.345.316.4—Table21:SubgroupComparisonsattheCAHLocusZYGOSITYALLELICFITNESSFREQUENCIESTOH.W.E.CAHNumber ofHomozygousHomozygousRESRESpeoplenegativeHeterozygouspositivenegativepositivePtested(-I-)(-1+)(+1+)(1df)Undefined1843320.420.580.29>50%NonOrientals8216290.420.580.45>50%**Orientals3350160.590.410.18>50%164Figure 37 : Amplification of Non-Human DNAAmplification at the ADA(Panel A) and APRT (Panel B) loci was attempted with 200 ngDNA from a variety of non-human sources.Lane m = molecular length standards1 = human DNA2 =dogDNA3 =catDNA4 = chicken DNA5 = mouse DNA (cell line 3T3)6 = C.albicans DNA (yeast)7 = pACYC (plasmid)8 = E.coli DNA (bacteria)9 = C.perfringens DNA (bacteria)10 = P.syringe DNA (bacteria)11 = B.subtilis (bacteria)- amplification proceeded for 35 cycles165Figure 37 : Amplification of Non-Human DNA0.96B0.73>;166corresponding primer set (j 0.96 kb for the ADA primers and 0.73 kb for the APRT primers)were not observed for any of the reactions utilizing non-human target DNAs (Figure 37) (Data forPALB, CAH, and LPL not shown). DNA which showed low levels of nonspecific amplification (Figure 37: Panel A-lanes 9 and 10, Panel B-lane 5) were reamplified with a 37 cycle reaction usingPALB primers. The absence of any observable DNA bands of the appropriate sizes even after 37cycles confirmed the previous results with 35 cycles (data not shown).To eliminate the possibility of inhibitors in the PCR which may have affected the activity ofthe Taq DNA polymerase, reactions containing non-human DNA were “spiked” with human DNA.Human DNA was successfully amplified in every case and reactions without spiked DNA weredevoid of amplified material (Figure 38). These results confirmed the specificity of theoligonucleotide primers for human DNA.167Figure 38: Amplification of Human DNA in Non-Human SamplesHuman DNA was spiked into non-human DNA samples and were amplified at the PALB locus.Lane m = molecular length standards1 = human DNA2 = dog DNA3 = cat DNA4 = chicken DNA5 = mouse DNA (cell line 3T3)6 dog DNA and human DNA7 = cat DNA and human DNA8 = chicken DNA and human DNA9 = mouse DNA and human DNA1.o1168CONCLUSIONSA DNA profiling system based on amplification of dimorphic restriction endonucleaserecognition sites offers many advantages. Minimal population data is required for each loci sincethere are only two allelic types (the restriction endonuclease recognition site is either present orabsent). Statistical problems with the Hardy-Weinberg test for equilibrium are absent as distinctfrom the complex statistical issues encountered when dealing with VNTRs. Interpretation of typingresults is simple as the size of the DNA fragment amplified is predetermined by the flankingprimers and the sizes of the digested fragments are also known. There are no requirements fortime consuming manual or computerized assessments of DNA band sizes or need for binningsystems.Incorporation of an internal control DNA fragment in each digest reaction enables theanalyst to monitor the activity of the restriction endonuclease which is critical to the wholedimorphic RES system. The validity and reliability of the system is also confirmed by the specificityof the oligonucleotide primers for human DNA at the designated (or appropriate) locus.Data compiled from an undefined population from the Greater Vancouver Areademonstrated that four (APRT, LPL, PALB and CAH) of the five loci, with dimorphic RES,analyzed in this study are in Hardy-Weinberg Equilibrium. Since the ADA locus is found to be notin Hardy-Weinberg Equilibrium, this locus cannot be used in the context of forensic discrimination.Deviation from equilibrium suggests that the Hardy-Weinberg formula is not applicable forcalculations of the probabilities of the occurence of specific genotypes in the population. Analysisof allelic frequencies at three loci (PALB, ADA, and CAH) of broadly categorized subpopulations(Orientals versus Non-Orientals), suggest that both subpopulations are in equilibrium at the PALBand CAH loci. Allelic frequency differences between subpopulations (even in the case of the ADAlocus), are not statistically significant except for the CAll locus. Hence subpopulation data basesshould be generated for each locus despite the presence or absence of Hardy-WeinbergEquilibrium.169Due to the possibility of name changes as a result of inter-racial marriages, categorizationof racial origin by surnames may not always reflect the true racial orgins of an individual.However, allelic frequency values should not be significantly altered because the number of suchname changes in the population study area is apparently very low and any frequency changes wouldbe small compared to the high allelic frequency values of these dimorphic RES loci.170SECTION IIDISCRIMINATION POWERThe value of this dimorphic RES system may be determined by calculating the probabilityof the system to discriminate between two samples. This ability or “discriminating power (DP) “isdefined as one minus the sum of the square of the phenotypic frequencies with the assumption thatthe RES loci are independent [146]. The sum of the square of the phenotypic frequencies can bedefined as the probability of identity (P1) or the probability of randomly selecting two individualswith the same phenotype (or genotype in this case) [141].1) Hypothetical Examples: (Analyses of 16 unlinked loci)I. Assuming that the frequencies of occurrence of the alleles at each of the 16 loci are 0.5(enzyme site present) and 0.5 (enzyme site absent).For 1 locus: DP = 1- Z(frequency of phenotypes)2= 1 - (0.25)2 + (0.50)2 + (0.25)2= 1 - 0.375 = 0.625For 16 loci : DP = 1 - (0.375)16 0.99999985P1 = Z(frequency of phenotypes)2= (0.375)16= 1.53 x 1W7Hence there would be an one in 6.5 million (1 / 1.53 x 1W7) chance of randomly selectingtwo individuals with the same phenotype (genotype) at these 16 loci.II. Assuming that the frequencies of occurrence of the alleles at each of the 16 loci are 0.7(enzyme site present) and 0.3 (enzyme site absent).For 1 locus: DP = 1 - (frequency of phenotypes)2= 1 - (0.49)2 + (0.42)2 + (0.09)2171= 1 - 0.4246 = 0.575For 16 loci: DP = 1 - (0.4246)16 = 0.99999989P1 (0.4246)16 = 1.116 x io.6Hence there would be an one in a nine hundred thousand chance (1 / 1.116 x 10-6) ofrandomly selecting two individuals with the same phenotype (genotype) at these 16 loci.2) Random DNA Profiling:In order to test the discriminatory potential of the dimorphic RES system, two groups of 10people were chosen at random (from the samples used to generate the allelic frequencies) andtheir DNA profiles were analyzed. In one group, 7/10 individuals had unique profiles (Table 22)and 8/10 individuals in the second group were distinguishable from one another (Table 23). Thishigh level of discrimination was achieved using only 4 loci containing dimorphic RES. Two alleleloci which have allelic frequencies of 50/50 will provide the highest probability of different alleles atthe same locus of two different individuals. As only 1/4 loci used here (LPL) have allelicfrequencies of 50-50, the feasibility of using loci with dimorphic RES is quite evident.3) Family DNA Profiling:Another important consideration was the ability of the proposed system to distinguishbetween close relatives. DNA profiles of individuals from two families were analyzed. UniqueDNA profiles were obtained for all (6/6) individuals from one family (Table 24) but only 4/6individuals were distinguishable in the second family (Table 25). For these loci, there are only twoalleles occurring at high frequencies so that the probability of two people possessing the sameallele is not too different regardless of whether they are related or not. However, once greaternumbers of loci are used, discrimination between relatives would be reduced since there would be agreater chance of more alleles being identical by descent.Table22:AllelicProfilesRi*=non-uniqueprofilefor thisgroupDiscrimination:70%%.Sample12*345*678*910!__________APRT÷1÷+1++/++1++1++1++1-+1-+1+-IPALB+1-+1++1--I-+1-+1-+1-+1-+1-+1+CAH+1+-I-+1-+1+-I-+1-+1-+1-+1-+1-LPL+1++1--I--I--I--I--I-+1+-I--I-ITable23:AllelicProfilesR2mp1e12*3456*78910APRT+1-+1++1++1-+1-+1++1++1++1++1+PALB+1++1-+1-+1-+1-+1++1-+1++1+‘CAH+1++1-+1++1++1-+1++1++1-+1-÷1LPL+1++1+//+1++1-1+1+÷1Discrimination:80%ITable24:AllelicProfilesFlDiscrimination:100%.epie123456Locus4%APRT+1++1++/++1++1++1+PALB+1-+1+-I--I--I-+1+CAH+1++1++1-+1-+1++1-LPL-I-+1-+1--I-+1+-I-tJTable25:AllelicProfilesF2%.1ip1e*123*456LocusAPRT+1-+1-+1-+1-+1+-IPALB+1-+1-+1-+1--I-+1-CAH+1-+1++1++1-+1-+1-LPL-I-+1--I--I-+1--I-*=non-uniqueprofileforthisgroupDiscrimination:67%1764) Discrimination Power With the Use of the Four RES Loci Studied Here:Z (phenotype frequencies)2of APRT = 0.4609LPL = 0.3699PALB = 0.4045CAl-I = 0.3817DP = 1- Z(frequency of phenotypes)2= 1 - (0.4609) x (0.3699) x (0.4045) x (0.3817)= 1 - 0.0263 = 0.9737P1 = Z (phenotype frequencies)2= 0.0263Hence there would be an one chance in thirty-eight (1 / 0.0263) of randomly selecting twoindividuals with the same phenotype (genotype) at these four loci.Although discriminating power with only 4 loci is not very high compared with VNTRanalysis, this can be increased with additional loci. Furthermore, the statistics for this system ismuch more reliable than for the VNTRs as there is no problem with differentiation of variousalleles and their corresponding frequencies. The ADA locus was not used in the above examplesbecause this locus was found to not be in Hardy-Weinberg Equilibrium.177CONCLUSIONSThe proposed dimorphic RES system is an extremely viable and attractive alternative to theVNTR system as a means for DNA profiling (genetic typing). Hypothetical calculations haveshown that if 16 unlinked loci (each with allelic frequencies of 50% - 50%), were analyzed, onlyone person out of approximately 6.5 million would have a particular allelic pattern at the 16 locitested (p170). Studies of DNA profiles from related and unrelated individuals demonstrated that ahigh level of discrimination is achievable with analysis of dimorphic RES at only 4 loci. The highnumbers of loci required (10-20) should not be a deterrent to using this system with respect toeither the availability of suitable loci or the amount of technical labour required.Although extensive preliminary work is necessary to locate 10-20 loci, obtain sequenceinformation, optimize PCR (for DNA from both blood and forensic specimens) and generate allelicfrequency data bases, this obstacle can be removed by work sharing between several laboratories.Each group would be responsible for a less daunting number of loci. Once the system is complete,standard procedures could be easily established. Primers with compatible reaction temperaturescould be put through the thermal cycler together (in different tubes) to improve time efficiency.By limiting the maximum size of the PCR products (prior to restriction endonuclease digestion) tono more than 2 kb, the separation of the DNA restriction fragments could be performed onpolyacrylamide gels which allows for the possibility of automation.In addition to the amount of labour involved, one must consider the feasibility of locatingsequence information around 10-20 dimorphic RES. Since the human genome has been estimatedto contain 1.4 x io heterozygous nucleotide sites due to neutral mutations [147], it should bepossible to locate at least 20 unlinked loci with dimorphic RES. DNA sequence information formany loci should be available in the near future as a result of the Human Genome Project. Thekey factor is the determination of linkage equilibrium (random combinations of alleles at variousloci) between 20 loci to ensure that the various loci are unlinked.178The five loci analyzed in this study are located on four different chromosomes to maximizethe chance of linkage equilibrium between the five loci. This is based on the assumption thatgenes on different chromosomes assort independently during meiosis and are thus in linkageequilibrium [148]. However, disequilibrium between different loci may still occur for loci ondifferent chromosomes [71]. Although simple methods are available to test for disequilibriumbetween as many as three loci [149-151], there is disagreement as to the best method formeasurement of multi-loci associations [150].Methods for analyses of multi-loci associations are complex and require in depth knowledgeof statistical principles and extensive computer time [E. Wijsman - personal communication]. Sincethese requirements are beyond the capability and accessibility of this author, the five loci studiedhere were not tested for disequilibrium. This testing must be performed prior to application of thissystem to forensic casework in a service-orientated role.179GENERAL CONCLUSIONSThe application of DNA genetic typing to human identification is undoubtedly the highlightof the past decade in the field of forensic science. Studies of apparently random variations at themolecular level in DNA certainly amplifies the amount of information retrieved thus, facilitatinggreater discrimination among individuals. The distinctive power of hypervariable regions consistingof variable number of tandem repeats (VNTR), is evident by the large number of observed allelesor bins (29) despite the conservative organization of the binning system used in this study.However, this tremendous diversity of alleles which affords VNTR loci their high discriminatingability, also renders them susceptible to complex technical and statistical problems. Statisticalproblems in association with possible racial and geographical differences with respect to allelefrequencies, demand that very conservative allele frequencies with high values to reduce chances offalse conclusions of guilt, be employed in the calculation of uniqueness. The fixed bin approached(used in this thesis) was designed to minimize differences between subpopulatons (both ancestraland geographical) [1061. Hence with the use of conservative allele frequencies and the highlycomparable results of data collected from defined and undefined populations suggest thatpopulation databases generated in one geographic location may be applicable to othergeographically distinct populations.The justified tendency towards conservatism leads to a reduction in the discriminativepower gap between VNTRs and dimorphic restriction endonuclease recognition sites (RES).Although this gap is still large, the decrease in discrimination when moving from VNTRs todimorphic RES, is not as great as it first appears. The disadvantage of the much higher frequencyvalues of the dimorphic RES system is offset by its reliability and simplicity. Reliability of results isensured as the amplified fragments and their digestion products are pre-determined so that typingerrors are detected. The smaller number of possible alleles (three) makes interpretation andstatistical analysis of results simple.180Allele frequency errors at one dimorphic RES locus would not dramatically alter theresultant profile uniqueness value (cf. errors at one VNTR locus). Since there are only twopossible alleles, each occurring at high frequencies, the extent of the error relative to the largefrequency value is small so that the level of inaccuracy is correspondingly small.In contrast to VNTRs, Hardy-Weinberg Equilibrium can be easily tested for the dimorphicRES system. Four of the five dimorphic RES loci presented here have demonstrated Hardy-Weinberg Equilibrium with both the undefined group and the loosely defined racial subpopulations(Orientals and non-Orientals). However, there is evidence of small allelic frequency differences inracial subpopulations at the carbonic anhydrase II (CAH) locus.The potential for generating characteristic DNA profiles by the analysis of a combination of10 to 20 dimorphic RES loci, is solidly confirmed by the demonstrated high level discrimination ofgroups of both related and unrelated individuals by testing with only 4 dimorphic RES loci. Thevolume of work required for the analysis of 10-20 loci is greatly reduced with the aid of thepolymerase chain reaction. Protocols for PCR can be developed and organised to handle this largenumber of analyses. In comparison with conventional VNTR (blot-hybridization) analysis, theDNA of amplified digestion products of dimorphic loci can be directly visualized on the gel withoutthe use of probes. Time spent on the analysis of banding patterns is minimal in contrast to that ofthe VNTR system.Use of the polymerase chain reaction technique in forensic science solves the importantproblem of sample size by increasing the quantity of DNA available for analysis . With thisproposed system, reliable results could be obtained from degraded DNA since only the RES and asmall portion of the flanking regions need to be intact. In situations where there is not enoughDNA for separate amplification of several loci, two or more different primer sets could be used inthe same reaction vial to amplify various regions simultaneously (“multiplex” amplification).Aliquots of the amplified products could serve as templates for secondary amplifications using onlyone specific primer set selected from the primary amplification. “Multiplex” amplification with 6 to1819 different primer sets have already been reported for the detection of Duchenne musculardystrophy [152,153]. Jeffreys et al. have also simultaneously amplified 6 single locus minisatelliteregions [154].Preliminary environmental studies which are discussed in the addendum, indicate thatforensic specimens can be successfully analyzed with both the VNTR (blot-hybridization) and thedimorphic RES (PCR) systems. Factors which affected the polymerase chain reactionamplification, with the exception of UV radiation, also affect the VNTR system. Samples exposedto the external environment are more likely to be successfully typed by the dimorphic RES system.Degraded DNA samples may be typed by this RES system because large sizes of intact DNAfragments are not required since the target fragments to be amplified are designed to be less than1.5 kb. The results obtained from this study do not provide any definitive conclusion as to the bestsystem for application to forensic specimens. The results of the dimorphic RES study areencouraging, however extensive research and development is still required for both systems.The protocol described here for the proposed dimorphic RES system, provides a simplepractical solution for individual identification in forensic science. Initial calculations showed that if16 such unlinked loci (with allele frequencies of 70% -30%) were tested, only one individual out ofnine hundred thousand individuals would have a particular allelic pattern at the 16 loci tested. Thishigh discriminating value can be further enhanced by utilizing loci with allele frequencies closer to50% - 50% and/or analyzing additional loci. Although, recent advancements in the amplification ofVNTR loci (AMP-FEP) may solve many of the problems with VNTR analysis such as higherresolution leading to discrete alleles and less complicated statistics [155], the dimorphic RES systemstill appears to be a promising alternative to VNTR analysis. Since data to establish identity arefrequently presented in court to non-scientists by non-scientist lawyers, judges, and juries, thetechnique herein described for dimorphic RES allows the data to be presented in a simple anddirect format that can be easily understood.182Recent criminal cases have highlighted the difficulties in the presentation of DNA evidencein court. Attempts have been made to discredit DNA evidence by making the technology appearcomplex, confusing, and unreliable. This dimorphic RES system makes possible the presentation ofDNA evidence in court in a simple, clear and precise way and should help remove these doubtsand difficulties that could inhibit valuable information necessary to convict or exonerate adefendant.Although this thesis has focussed exclusively on the use of DNA profiling in a forensiccontext, restriction fragment length polymorphisms (RFLPs) has other practical applications. Forexample, DNA analysis can be applied to: 1) determination of twin zygosity [156], 2) humanpaternity disputes (in cases of both palimony and immigration issues) [157,158,53], 3) identificationof domestic animals such as cattle and horses for breeding purposes and theft prevention,4) prevention of theft and poaching of wild animals, 5) enforcement of fishing regulations withrespect to specific species of fish [159], 6) selective fish breeding [160], 7) evolutionary studies ofhumans and animals, 8) screening of tissue compatibility for organ transplantations 9) assessementof clonality of tumors [161], and 10) epidemiology studies of microbial outbreaks. These numerousapplications of DNA profiling through RFLP analysis necessitates and ensures that continuousefforts will be made towards the development and refinement of this discriminating procedure wellinto the 21st century.S 183SUMMARY CONCLUSIONS1. VNTR allele frequency data bases collected by sampling an undefined populationof the Greater Vancouver Area are comparable to those collected by other laboratoriessampling Caucasian populations.2. Hypervariable regions in DNA consisting of variable number of tandem repeats exhibittremendous diversity and hence has the potential to provide high levels of discriminationsuitable for forensic profiling.3. A binning system is required to deal with the problems of a continuous distribution ofVNTR alleles, inadequate gel resolution of closely spaced alleles and inter/intra gelvariabilities.4. PCR amplification of DNA sequences containing dimorphic restriction endonuclease recognitionsites (RES) can be used to eliminate the use of DNA probes for detection of restrictionfragment length polymorphisms (RFLPs).5. Dimorphic RES allele frequency data bases collected by sampling an undefined population ofthe Greater Vancouver Area are comparable to those collected by other laboratories.6. Hardy-Weinberg Equilibrium was demonstrated for the 4/5 dimorphic RES loci analyzed in thisstudy, sampling either an undefined population (total) or two racial subpopulations(Orientals and non-Orientals).1847. Dimorphic RES allele frequency comparison of two racial subpopulations (Orientalsversus non-Orientals) indicated that there are no significant differences at the PALBand ADA loci.8. DNA profiling with only 4 dimorphic RES offers significant discrimination.9. Both the VNTR and dimorphic RES systems can be used with varying degrees of success toobtain DNA profiles from forensic casework specimens and/or laboratory simulatedforensic samples.1O.Although both DNA profiling systems described in this study still require extensiveresearch and development, the dimorphic RES system offers a much simpler and reliableapproach to DNA profiling.ADDENDUMInfluences of Environmental ConditionsOn DNA Profiling of Forensic Specimens185186INTRODUCTIONThe ultimate criteria of any system developed for forensic purposes, is that it be functional,reliable, and reproducible when applied to forensic specimens. As discussed in chapter 1 (sectionII), there has been extensive research into the types of specimens amenable to RFLP analysis.Correspondingly, studies have also been made to determine the types of forensic specimens suitablefor amplification by the polymerase chain reaction.Due to degradation of ancient DNA, amplification efficiency of aged samples was found tobe inversely related to the size of the fragment to be amplified [162]. Fragments with maximumlengths of 100-200 bp have been successfully amplified from mitochondrial DNA extracted from a7000 year old brain [163,164]. Amplification was found to be more efficient in partially degradedsamples than in undegraded samples [165]. This effect was attributed to more efficient strandseparation in samples that are partially degraded. Amplification of degraded DNA samples ispossible as long as there are DNA fragments containing the entire target sequence [165].The FBI has utilized the polymerase chain reaction to amplify a 242 bp region of the majorhistocompatibility complex (DQ-a), from DNA isolated from samples such as aged blood stains (1-9months old), 1 ul of semen stains, single hair roots and vaginal swabs containing semen [166]. PCRamplification of the DQcx locus from DNA isolated from either shed or plucked hairs has beenreported, with more reliable results from plucked hairs [167]. DNA from two forensic casesinvolving aged specimens (a > 10 year old semen stain and skull cap reported to be 18 months old)have also been shown to be amenable to PCR analysis [168]. Recently, Walsh et. al. has publishedprotocols for the extraction of DNA, suitable for PCR analysis, from a variety of forensic materials[169].In addition to types of specimens, one needs to determine environmental factors thatinfluence the success and validity of either the VNTR or the polymerase chain reaction systems.Several studies on the effects of ultraviolet light, humidity, soil, on bloodstains have already been187done [170-172]. The follow section serves as preliminary studies of environmental influences onPCR dimorphic RES analysis and makes comparisons with that of the VNTR (blot-hybridization)system.188METHODS1. PREPARATION OF DNAPreparation of BloodstainsOne hundred microliter aliquots of whole blood were pipetted onto a piece of 100% cottoncloth (washed and dried three times) and dried at room temperature. The dried bloodstains werethen treated as followsa) kept at room temperature without light (in a lab drawer) for varioustime periods up to 11 monthsb) hung on the external side of an east facing window (at 10th andHeather Street) in Vancouver, for various recorded time intervalsduring the period between October 17/90 and December 6/90(approximately 7 week period)* c) irradiated with varying amounts (0 - 21,600 J/M2) of short waveultra violet light using a Stratagene Linkerd) partially buried for up to 10 days in soil taken from an alley inVancouver*- The amount of UV irradiation was converted to an equivalent period of exposure to sunlight.The conversion factor used for this calculation was obtained by using a UV light meter (UVX-25Sensor) to measure the quantity of short wave (254 nm) UV irradiation from the sun. Readingswere made at 12:00 noon on three consecutive sunny days in mid January 1991 and the highestvalue was used in the calculation.Isolation of DNA From Bloodstains For VNTR Analysis (Protocol D)Deoxyribonucleic acid was extracted from each blood stain as described in a protocol usedby the FBI. The blood stain was cut into approximately 2 mm square pieces, placed into a 1.5 ml189microcentrifuge tube containing 400 ul stain extraction buffer and 10 ul of Proteinase K (20 mg/mi)and incubated overnight in a 56°C water bath. The liquid from the tube was then removed andsaved in another tube. Using a syringe needle head, a hole was made in the bottom of the originaltube which, at this point, only contained the damp pieces of stain. The tube with the hole andpieces of stain were placed on top of a new tube and spun for 2 minutes at 13,000 xg to draw outthe remaining fluid from the damp pieces of stain. The fluid collected in the new tube was pooledwith the fluid saved previously. This fluid was then extracted once with 500 ul ofphenol/chloroform/isoamyl alcohol (25:24:1) and once with 500 ul of chloroform/isoamyl alcohol(24:1). The DNA in the fluid was precipitated with 1.0 ml of cold absolute ethanol at -20°C for atleast 30 minutes. The precipitated DNA was collected by centrifugation at 13,000 xg for 15 minutesat -20°C and washed with 70% ethanol. The resulting DNA pellet was either air dried or dried ina vacuum desiccator and resolubilized overnight in 40 ul TE4 in a 56°C water bath. The entiresample was used for one VNTR typing analysis.Purification of DNA Extracted From Stains For VNTR Analysis:The resolubilized DNA was placed on a Millipore disc filter (0.05 uM) and dialysed for twohours in LTE.Isolation of DNA From Liquid Blood or Bloodstains With Chelex 100 For PCR Analysis(Protocol E)Deoxyribonucleic acid was extracted according to a protocol used by the FBI. Briefly, in a1.5 ml microcentrifuge tube containing 1.0 ml of dH2O, either a volume (50-100 ul) of liquidblood or a cut up piece of cotton blood stain (50-100 ul blood) was added. The tube with itscontents was incubated at room temperature for 30 minutes and then spun in a microcentrifuge for2-3 minutes at 13,000 xg. Without disturbing the pellet, as much as possible of the supernatant wasdiscarded and the pellet (including the pieces of fabric in the case of the bloodstain) was made upto a final volume of 400 ul with 5% (w/v) Chelex 100 (Biorad). This mixture was incubated for 30190minutes in a 56°C water bath (Blue M) and then vortexed at high speed for 10 seconds beforeboiling in a water bath for 8 minutes. The boiled mixture was once again vortexed at high speedfor 10 seconds and the supernatant (containing the DNA) was then collected by centrifugation at13,000 xg for 2-3 minutes.Purification of Chelex 100 Extracted DNA:The supernatant containing the extracted DNA was purified with GENECLEAN (Bio 101Inc.) according to the kit protocol. Approximately 2.5 volumes of Nal solution was added to theDNA solution (supernatant) and mixed before the addition of 8 ul of glass milk (suspension ofsilica matrix in water that binds single and double stranded DNA without binding DNAcontaminants). This mixture was incubated for 5 minutes in an ice bath to promote the binding ofthe DNA molecules to the silica matrix. Following this incubation, the mixture was spun for 5seconds at 13,000 xg and the supernatant containing the unwanted contaminants, was discarded.The silica matrix pellet (with bound DNA) was washed three times with ice cold NEW WASHbuffer (a mixture of Tris-acid and Tris-base with a pH of 7.0-8.5), the reagents of which wereprovided with the GENECLEAN kit. After the third wash, the DNA was eluted from the silicamatrix with dH2O and incubated for 3 minutes at 50°C. The suspension was spun for 30 secondsat 13,000 xg and the supernatant containing released DNA (80% of total) was collected. Thepellet was eluted once more to recover residual DNA (10-20%). The final total volume of theextracted stain DNA was 300 ul.2. PCR AMPLIFICATION OF DNA EXTRACTED FROM BLOODSTAINSPCR For the APRT, PALB, and ADA Loci From DNA Extracted From Bloodstains:Fifteen microliters (1/20) of Geneclean purified Chelex 100 extracted DNA from one bloodstainwas used for each amplification reaction. The polymerase chain reaction was carried out asdescribed previously for each loci except for changes in the MgC12 concentrations and the number191of cycles completed. The MgC12 concentrations were increased for each loci so that it was 2.0 mMfor the APRT locus, 2.5 mM for the PALB locus, and 3.3 mM for the ADA locus. Each PCR wasallowed to continue for 33 cycles.Control For PCR Amplification:As with the data base experiments a positive control (stock DNA sample with reliableamplification) was included in each PCR run. The absence of negative controls does not affect therecognition of amplification activity, the basic purpose of these experiments. However, negativecontrols should be included in actual forensic case analysis.3. VNTR ANALYSIS OF DNA EXTRACTED FROM BLOODSTAINSRefer to chapter 1 - methods section (VNTR analysis)192RESULTS AND DISCUSSIONTechnical Conditions of DNA Extraction For PCR Analysis of ForensicSamplesThe ability to amplify DNA isolated from forensic samples is a vital factor in assessing thefeasibility of employing this dimorphic RES system for forensic identification. At the onset of thisproject, PCR amplification of DNA from forensic samples was at its infancy. Therefore, despitethe success of other laboratories, it was important to demonstrate that amplification of forensicspecimens was also easily achievable and reproducible in this study with these specifically chosendimorphic loci.Initially, in the absence of other more suitable methods known to this laboratory, DNA wasextracted from stains using protocol D and used in the polymerase chain reaction (PCR).Unfortunately, amplification of these DNA samples was not successful, possibly due to the presenceof contaminants which inhibit the activity of the Taq DNA polymerase. The contaminant suspectedof causing this inhibition was sodium dodecyl sulfate (SDS) used in the extraction procedure. Thepresence of low levels (0.01%) of SDS has been reported to be inhibitory for Taq Polymerase I andthe addition of 0.1% of Tween 20 and 0.1% of NP4O completely reverses this inhibition [173].However, the addition of Tween 20/NP40 still did not result in amplification of the stain extractedDNA. The strategy was then to switch the extraction method to one which produces purifiedDNA.The next DNA extraction procedure attempt was that of the BRL DNA Capture Reagentprotocol. The DNA Capture Reagent (modified Sepharose beads) immobilizes DNA frombuffered DNA solutions or complex biological samples leading to its purification from undesiredmaterials by filtration or centrifugation. In product developmental studies, 80-100% of input DNA(in various buffer solutions) was immobilized by the Capture Reagent and 90% of this bound DNAwas subsequently recovered [174]. PCR amplification of stain DNA extracted with this BRLprotocol was not successful. Amplification was also not seen with a DNA solution recovered from193the DNA capture of 300 ng of DNA (in LTE). Since this “spiked” DNA (extracted from wholeblood using protocol A) had previously been shown to be amplifiable, it appears that in thissituation, the DNA Capture Reagent protocol is not as efficient as previously reported.Amplification of stain DNA was finally achieved with DNA extracted using Protocol Eobtained from the FBI and which has since been published by Walsh et al. [169]. This protocoldoes not require detergents but revolves around the use of high temperatures (100°C) in thepresence of an alkaline chelating ion exchange resin, Chelex 100 [175]. The alkalinity of Chelex100 (pH 10.8 at room temperature) not only facilitates the lysis of the cells but also assurescomplete DNA denaturation. Furthermore, this resin has a high affinity for polyvalent metal ionsand its presence during boiling of stain suspension is thought to prevent DNA degradation bychelating metals ions that serve as catalysts for such degradation.In order to simulate possible environmental conditions under which forensic specimensexist, bloodstains were exposed to the external elements for various time intervals and attemptswere made to amplify DNA extracted from such samples with three of the five dimorphic RESprimer sets. Only 1/20 of the purified DNA extracts were used because larger proportions resultedin lower amounts of amplified products. Low concentrations of red blood cell components such ashaemoglobin, hematin, protoporphyrin IX, FeC113 and FeSO4 have been demonstrated to inhibitPCR [176]. Inhibition by hematin occurs at a level lower (0.8 uM) than the other components (20uM), which is significant since this component is found in high levels in aged bloodstains and copurifies with DNA. It is suggested that hematin is a competitive inhibitor of Taq DNA polymerase.Technical Conditions of DNA Extraction For VNTR Analysis of Forensic Samples: Initially,protocol C (pZ7) was used for the isolation of DNA from bloodstains, blood swabs, and cellsattached to hair shafts. The change to protocol D coincided with the change from protocol A to Bfor DNA extraction from whole blood (chapter 1). Protocols B and D were obtained together fromthe FBI laboratory and it became apparent that these protocols were more suitable for forensicVNTR analysis. For example, with protocol C, forensic specimens are initially incubated overnight194at 37°C and in protocol D, the temperature of incubation is 56°C. Cellular nucleases releasedduring cell lysis or external nucleases (contaminants of the specimen) may still be active at 37°Cand not at 56°C. Hence, incubation at the higher temperature would inhibit possible DNAdegradation. To allow for comparisons between the two profiling systems, simulated forensicbloodstains identical to those analyzed by PCR were also typed by probing of 4 VNTR loci. Inorder to save time, two probe combinations (j 3’HVR a-globinIYNH24 and phins 310/CMM1O1),were used per hybridization reaction to probe simultaneously at two loci.Effects of Aging On DNA Genetic Typing of Bloodstains:Deoxyribonucleic acids extracted from bloodstains stored in a laboratory drawer for up to11 months were successfully amplified by the polymerase chain reaction (Figure 39). VNTR (blot-hybridization) band patterns were also obtained from such stains up to 8 months old (Figure 40).(Older stains for VNTR analysis were unavailable.) The degree of amplification or intensity ofDNA bands was decreased with increased age of bloodstains. This suggests that the amount ofDNA fragments smaller than 1 kb increases with increased age of sample. A value of 1 kb issuggested since that is the approximate length of the target fragment amplified with thesedimorphic RES loci and the entire fragment must be intact if amplification is to occur.However, DNA degradation may not be a sole cause of decreased levels of amplification orband intensity (in the case of VNTR blot-hybridization system) in aged stains. Although Pããboet.al. found, in ancient tissue samples, an inverse relationship between age and amplificationefficiency [163], the amount of DNA degradation did not correlate with age of specimen [164].High levels of modified thymine in ancient samples, suggest that significant amounts of oxidativedamage had occurred [164]. Hence, DNA in dehydrated tissues may be protected from hydrolyticdamage and remain relatively intact, but it is still susceptible to oxidative damage.Effects of Environmental Factors Particular to Vancouver, on DNA Genetic Typing of BloodstainsPossible causes for DNA degradation of forensic samples exposed to the external195Figure 39 : Amplification of DNA From Aged StainsDNA extracted from human bloodstains kept at room temperature without light for various timeperiods were amplified at the ADA locus.Lane m = molecular length standards1 = lab stock DNA from blood sample2 = DNA from 2 day old stain3 = DNA from 3 month old stain4 = DNA from 5 month old stain5 = DNA from 11 month old stain6 = negative control (no DNA)O.87 O.96196Figure 40 : VNTR Analysis of DNA from Aged StainsDNA extracted from human bloodstains kept at room temperature without light for varioustime periods were simultaneously probed with 3’HVR a-globin and YNH24 to detect allelesat the D16S85 and D2S44 loci.Lane M = molecular length standards1 = DNA from 2 day old stain2 = DNA from 5 month old stain3 = DNA from 8 month old stainLong arrows indicate position of alleles for the two loci.3’HVR /YNH24233 day exp197environment would be: temperature, humidity, air pollutants, ultraviolet light, and soil. Hencebloodstains were subjected to these various conditions prior to analysis by either the VNTR (blothybridization) or dimorphic RES (PCR) system.During the period between October 17 (1990) and December 6 (1990), the amount ofprecipitation (as measured at the Vancouver International Airport), was slightly higher than normalfor this time of the year [177]. Daily temperatures were also slightly higher than normal with amean temperature of not greater than 9.9°C for any of the 3 months. However the total hours ofsunshine during October to December, was 57 hours (25%) less than normal. Humidity was highwith relative humidity of not less than 87% during the entire 7 week period.Deoxyribonucleic acids were extracted from bloodstains hung for up to 7 weeks (duringOctober to December) on the external side of an east facing window (at 10th and Heather St.) inVancouver. All samples were successfully amplified by the polymerase chain reaction. Undigestedamplified products of APRT, PALB, and ADA from blood DNA (left panel), 4 hour old stain(middle panel) and 7 week old stain (right panel), are shown in Figure 41. Digestion products ofPALB amplified samples are shown in Figure 42. Not surprisingly, the degree of amplificationdecreased with age of stain (Figure 42). This conclusion appears to be in contradiction with theresults seen for DNA from a bloodstain only 4 hour old (Figure 42 - lane 1). Previousamplifications of DNA aliquots from the same sample demonstrated amplification levelscomparable to that of DNA from stains 1 day old. The low level observed in this one reaction isthe result of inconsistent amplification. This explanation is confirmed by variable levels ofamplification of DNA aliquots from identical samples (Figure 43 - lanes 1/A, 2/B, 3/C and 4/D).Amplification inconsistencies may be a consequence of minute differences in the amount of startingDNA which leads to great differences as a result of exponential accumulation of DNA. Anotherpossibility may be the failure of amplification during the first few of cycles such that amplificationdoes not actually begin until cycle 2 or 3, resulting in a 2 or 4 fold decrease in the amount ofproduct.198DNA extracted from liquid blood (Left panel), 4 hour old bloodstain middle panel) and7 week old external bloodstain (right panel) were amplified at the APRT, PALB and ADAloci.Figure 41 Amplification of DNA From Bloodstain Exposed to ExternalEnvironment (APRT/PALB/ADA)Lane m = molecular length standards1 = APRT locus (0.73 kb)2 = PALB locus (1.01 kb)3 = ADA locus (0.96 kb)1.35>O.87m123—/. 1.01I0.96199Figure 42 : Amplification of DNA From Bloodstain Exposed toExternal Environment (PALB)DNA extracted from human bloodstains either kept at room temperature without light(internal) or exposed to the external environment for various time periods, wereamplified at the PALB loci. The amplification products were digested with Fnu 4H1which resulted in digestion products of 0.65 and 0.36 (represented by long arrows).Lane m = molecular length standardsL = lab stock DNA from bloodB = DNA from blood extracted as per stains1 = DNA from 4 hour old stain (internal)2 = DNA from 1 day old stain (internal)3 = “ “ (external - environment)4 = DNA from 1 week old stain (internal)5 H H H H H (external)6 = DNA from 2 week old stain (internal)7 = “ “ (external)8 = DNA from 3 week old stain (internal)9 = “ “ “ (external)10 = DNA from 4 week old stain (internal)11 = “ “ (external)12 DNA from 5 week old stain (internal)13 = “ (external)14 = DNA from 6 week old stain (internal)15 = “ “ I’ (external)16 = DNA from 7 week old stain (internal)17 = “ “ (external)Figure42:AmplificationAfterExposuretoExternalEnvironment(PALB)mLB123456789m1011121314151617————©Figure 43 : Inconsistent AmplificationIdentical samples of DNA were amplified at the PALB locus.Lane m = molecular length standardsLane 1 and A lab stock DNA extracted from blood2 and B = blood DNA extracted per stain3 and C = 4 hour old stains (internal)4 and D = 1 day old stains (internal)201202Comparisons of PCR products of DNA from stains kept in a laboratory drawer (controlsample) and stains exposed to the external environment (test sample) for a corresponding period oftime, indicated no consistent difference in the level of amplification (Figure 42). This suggests thatthe polymerase chain reaction is not affected by the environmental factors prevalent during thattime period, but are influenced by the effects of aging.In contrast, observable differences were found with VNTR analysis of these two types ofstains. A seven week old stain stored in a laboratory drawer (Figure 44 - lane 5) produced oneband with 3’HVR a-globin, no bands with YNH24 or phins 310 and two bands with CMM1O1(results with phins310/CMM1O1 not shown). A bloodstain exposed to environmental conditions for7 weeks (Figure 44 - lane 6), failed to produce a pattern with any of the four probes even afterprolonged exposure of the autoradiograph. The 7 week old control and test stains (Figure 44 -lanes 5 and 6) compare poorly with the more fresh stain seen in Figure 44 - lane 1. This may bedue to insufficient quantities of high molecular length DNA molecules. This was concluded fromthe observation of only a faint low molecular length DNA smear in each lane (especially in the caseof the test sample - lane 6), after agarose gel electrophoresis of the restriction nuclease digestedsamples (data not shown). As the degree of DNA degradation increases, the intensity of highmolecular length bands decreases relative to the lower molecular length fragments [178]. DNAsamples in the adjacent lanes which were successfully typed demonstrate a DNA smear through outthe length of the gel after digestion with Hae III. These results suggest that both age of sampleand environmental conditions influence the quality and/or quantity of extractable DNA.This observation that environmental factors affect VNTR analysis confirms the results of aFBI validation study. In that study, VNTR analysis of 8 weeks old bloodstains (equivalent volumeof blood unknown) exposed to ambient outdoor temperatures (average daily high of 21°C) duringMarch through May, was also unsuccessful due to DNA degradation [179]. However, DNAextracted from control bloodstains kept in the dark, were of sufficient size and quantity to producea VNTR banding pattern. The difference in the degrees of success of VNTR analysis observed203Figure 44 : VNTR Analysis of DNA Extracted FromExternal/UV/Soil BloodstainsDNA extracted from bloodstains exposed to the external environment, or ultravioletradiation, or soil contaminants, were simultaneously probed with 3’HVR a-globin andYNH24 to detect alleles at the D16S85 and D2S44 loci.Lane M = molecular length standards1 = DNA from 4 hour old stain (internal)2 = DNA from stain exposed to 3600 JIM2 of UV3= ,, ,t ,, ,, H 9000 “4 = I? H H It H 144005 = DNA from 7 week old stain (internal)6 = “ “ “ “ “ “ (external)7 = DNA from 10 day old bloodstain buried in soilFigure 44 : VNTR Analysis of DNA From External/U.V./Soil Bloodstains3’HVR /YNH24204O.6Olit002345c:*t*aSIL3 day exp205between the control stains (no external exposure) of this study and that of the FBI, may be a resultof non-comparable volumes of blood in stains. Additionally, the partial VNTR band patternobtained with the 7 week (less than 2 months) old non-exposed bloodstain in this experiment is notconsistent with the greater success achieved with 5 and 8 months old non-exposed stains discussedpreviously (page 194: Figure 40). The only difference between the stains used in the twoexperiments is the environment of the stains during their drying period. The 5 and 8 month oldstains were dried during the month of February (1990) and the 7 week old stain was dried duringOctober. Differences in humidity levels and temperature may have played a role in the variableability to type aged stains. As mentioned previously, the humidity was high during October and thismay have prolonged the drying period of the stain.Since there are indications of an environmental role in the quality and/or quantity of DNAextracted from forensic samples, it was necessary to attempt to identify the environmental factorsinvolved.A) Weather -RFLP (VNTR) patterns consistent with those obtained from liquid blood samples of thesame origin, were obtained from bloodstains incubated at 37°C for up to 5 days [170], and 20weeks [179]. Hence prolonged high humidity, rather than temperatures levels, was a more likelycause of DNA degradation in the test sample exposed to the external environment. The largeamount of precipitation could also have influenced the amount of DNA available for extraction assome of the blood in the stain may have been washed away.B) Air Pollution -The contribution of air pollutants such as nitrogen oxides (nitric oxide and nitrogendioxide) and sulfur dioxide to DNA damage or degradation was also questioned. Nitrous acid(HNO2)formed from the reduction of nitrogen dioxide (NO2), reacts with primary amino groupsto oxidatively deaminate adenine to hypoxanthine, cytosine to uracil and guanine to xanthine [180].206Aqueous sulfur dioxide (SO2), which is the equivalent of bisulfite (HSO3), also catalyses thedeamination of cytosine to uracil [1811. Such changes in DNA bases could affect DNAamplification and restriction endonuclease digestion but would probably not promote DNAdegradation.Nitrogen oxides (as with sulfur dioxide) are produced by the combustion of fossil fuels[182]. Levels of nitrogen oxides are higher than for sulfur dioxide as NO are discharged fromsteam boilers, building heating systems and internal combustion engines. Although the meanhourly levels of NO (0.027 ppm) [183] during the period of stain exposure did not exceed themaximum acceptable air quality levels (0.210 ppm) [182], DNA damage could have occurredwithout any apparent affect on DNA amplification.Sulfur dioxide (SO2) is formed by the combustion of fossil fuels containing sulfur [182].Except for oil refineries and cement plants, natural gas rather than oil is used in almost all heatrequiring installations in the Greater Vancouver Regional District, thus the level of sulfur dioxide islow relative to most other cities. The mean hourly concentration of sulfur dioxide during Octoberto December was 0.006 ppm [183] which is well within the maximum acceptable air quality level of0.340 ppm [182]. Due to such low levels of SO2, it appears that any possible DNA damageattributable to sulfur dioxide would be negligible.C) Ultraviolet Irradiation -Another possible factor affecting the state of DNA molecules is ultraviolet irradiation.DNA damage can occur as a result of UV-induced pyrimidine dimers, interstrand DNA crosslinksand DNA-protein crosslinks. McNally et al. found that bloodstains exposed to ultravioletirradiation (from a UV light box), for periods up to 5 days, produced consistent RFLP (VNTR)patterns but the intensity of the radiolabeled DNA bands decreased with increased UV exposuretime [170]. Stains exposed to sunlight for 10 days failed to produce a VNTR (blot system) bandpattern while identical stains stored in darkness at the same temperature provided a bandingpattern [179]. RFLP analysis using VNTR probes produced DNA profiles from bloodstains207irradiated with up to 14,400 JIM2 of short wave (254 nm) ultraviolet light (Figure 44 - lane 4).This amount of UV light is equivalent to at least 4 hours of direct sunlight. (UV crosslinking ofDNA to membranes only require 1,200 JIM2 or the equivalent of 20 minutes of direct sunlight).Irradiated stain DNA produced consistent VNTR bands (Figure 44 - lanes 2,3,4) but weresomewhat less intense than bands from the untreated 4 hour old stain (Figure 44 - Lane 1).Although band intensities decreased slightly with increased doses of UV irradiation, there does notappear to be a simple direct relationship. Bloodstains irradiated with UV light at quantities whichdiffer by a factor of 4, demonstrated only minor differences in band intensities (Figure 44 - lane 2and 4). Since forensic samples are often not in direct sunlight for prolonged periods i.e. foldedand/or partially covered or shaded, the amounts of UV irradiation per time interval received bysuch samples, are probably much less than the values suggested here. Hence in terms of actualforensic casework stains, the amount of UV irradiation received from short term (weeks asopposed to months) exposure to sunlight, should not be a major cause of DNA damage and/ordegradation.As previously discussed (p202) amplification of stains exposed to the external environmentdid not appear to be influenced by environment in this situation, but were affected by age. Despitethis observation, it was necessary to determine whether high levels of UV irradiation would affectamplification. Bloodstains irradiated with various doses of ultraviolet light were all successfullyamplified. However, the level of amplification was much less than that obtained from untreatedsamples and did not correlate with the dosage of UV irradiation received (Figure 45 - panel A andB). Amplification at the APRT locus was possible only with stains irradiated with up to 14,400J/M2 (Figure 45 - panel B) whereas, amplification was successful at levels up to 21,600 J/M2 at theADA and PALB loci (Figure 46 - panel A and B, lane 4). Furthermore, amplification of all 3 lociwas not constant. Amplification occurred in one experiment and not in another and there was noapparent pattern to this inconsistency which occurred more frequently than in all other nonirradiated samples in the entire project.208Figure 45 : Amplification of DNA Extracted FromUV Irradiated BloodstainsDNA extracted from bloodstains exposed to ultraviolet radiation were amplified at thePALB (Panel A) and APRT (Panel B) loci.Panel A:Lane m = molecular length standards1 = lab stock DNA from blood2 = DNA from 4 hour old stain (no treatment)3 = DNA from stain irradiated with 600 JIM2 UV4 = H H 12005 = I? H H H H 2400 H H6 = H H H II II 36007 = H fl H H 90008 = 14400 “Panel B:Lane m = molecular length standards1 = lab stock DNA from blood2 = DNA from 4 hour old stain (no treatment)3 = DNA from stain irradiated with 2400 J/M2 UV4 = H H H H 14400 H•: 1.35— (0.87209Figure 45 : Amplification of DNA From U.V. Irradiated BloodstainsABO.730.87(0.60210Figure 46 : Amplification of DNA Extracted FromUV Irradiated and Soil BloodstainsDNA extracted from bloodstains exposed to ultraviolet radiation or soil contaminantswere amplified at the ADA (Panel A) and PALB (Panel B) loci.Lane m = molecular length standards1 = lab stock DNA from blood2 = DNA from 4 hour old stain (no treatment)3 = DNA from 10 day old bloodstain buried in soil4 = DNA from stain irradiated with 21600 JIM2 UVO.96 A1B211An example of such variation can be seen in Figure 45 (panel A - lane 8), where DNAfrom a stain irradiated with 14,400 JIM2 of UV light was not amplified with PALB in this PCRexperiment but amplification was possible with a dosage of 21,600 JIM2 in another experimentshown in Figure 46: panel B-lane 4.The results observed in this set of experiments do not correlate with the results of thestains exposed to the external environment. A stain irradiated with an equivalent of only 10minutes of sunlight (600 J/M2) exhibited a noticeable decrease in the level of amplification. Theconclusion reached earlier that the absence of an influence exerted by the environment onamplification may be erroneous. The affects of environment may have been observed but werewrongly attributed to only the age of the sample. Since, degradation was occurring with age, UVdamaged was not be separately distinguished as the effects of the two are not apparently additive.Buoncristiani et al. found that short wave UV exposure of naked DNA in solution of morethan 200 J/M2 reduced amplification and at levels exceeding 2000 J/M2, no amplification wasobserved [172]. It was further determined that short wave UV radiation causes pyrimidine dimers(which blocks Taq polymerase) and a good distribution of such dimers was evident at doses of 120J/M2. However, amplification of liquid whole blood, liquid semen and bloodstains exposed to highlevels of short wave UV radiation, was not significantly affected. The bloodstains results ofBuoncristiani et al. are not in agreement with the results of the present study. This discrepancymay be partially explained by differences in base sequence of the amplified DNA.Studies on the origin of ultraviolet damage in DNA by Becker and Wang had shown thatthere are sequence and structural requirements for ultraviolet photoproduct formation in DNA[184]. Becker and Wang found that the sequences surrounding purines or pyrimidines and theflexibility of DNA at that site will influence the ability of any base to form ultraviolet photoproducts(dimers). For double-stranded DNA, UV photoreactivity (predominantly dimerizations) occursbetween adjacent pyrimidines and very infrequently in non-adjacent pyrimidine residues. Incontrast, purine photoreactivity will only occur readily if the purines are flanked on the 5’side by at212least two contiguous pyrimidines which transfer the neccessary activation energy. The flexibility ofDNA refers to presence or absence of torsional constraints which may inhibit the adoption of ageometrical configuration neccessary for the photoreaction.Furthermore, effects UV irradiation on amplification of bloodstains by the first group wasnot observed possibly because of the size of their target fragments (110 bp and 242 bp). The sizeof the smallest target fragments in the present study is 730 bp and hence there is a greater chancefor the photoreactivity (i.e. DNA dimerization and strand breaks) within the target fragment.D) Soil -Forensic specimens are often exposed to soil and its contaminants as a result of thephysical location of the materials under investigation. RFLP (blot-hybridization) analysis of soiledstains (case samples) by McNally et al. indicated that partial degradation of DNA in soiled stainsdoes occur suggesting that soil has a negative effect on DNA integrity [171]. This result issubstantiated by two other studies in which RFLP analysis (blot-hybridization) was unsuccessfulwith four and 5 day old blood-soil samples [170,179]. However, there was bacterial contaminationof some of the blood-soil samples which was detected with bacterial probe pAC267 [170].In contrast with the three previous studies, this laboratory obtained an intense RFLP (blot-hybridization) band pattern with DNA extracted from bloodstains partially buried in soil for 10 days(Figure 44 - lane 7). Two of these bands are slightly distorted and the background is higher thanthe untreated control sample (lane 1). However, DNA from this soiled stain does not appear tohave degraded as a result of aging or the effects of soil and/or soil contaminants. This conclusioncan be drawn since the intensity of its DNA bands are at least comparable to, if not more than, thatof the 4 hour old untreated control sample.Amplification of the 10 day old soiled bloodstain was also successful at the ADA and PALBloci (Figure 46 - lane 3) but was not possible at the APRT locus. Of the five loci, the ADA andPALB loci are the most efficiently amplified. Hence, the inability to amplify the APRT mayeventually be resolved with further development of this system. The specificity of these primers for213human DNA eliminates the problem of bacterial contamination found in soil samples [179]. Inretrospect, a bacterial probe could have been used to probe the amplified products of the soiledbloodstain to ensure that the products were of human origin. However, this writer is confident thatthe amplified products are of human origin because the size of the amplified DNA agrees with thatof theoretical calculations, and the specificity experiments discussed earlier in chaper 2 (p161) areconclusive.214CONCLUSIONSThe application of the VNTR (blot-hybridization) and the dimorphic RES (polymerasechain reaction) systems, to forensic specimens i.e. bloodstains, has met with varying degrees ofsuccess. Protocols have been chosen and altered to provide the most accurate and reliable resultswith cotton bloodstains. The success of restriction fragment length polymorphism (RFLP) witheither system is largely dependent on the amount of high molecular length DNA and/or undamagedDNA extracted from such forensic specimens. The environmental exposure studies in this thesiswere not rigorously performed because the main purpose of these environmental exposure studieswas to determine the general effects of environment on DNA and precise quantitative andqualitative aspects of extracted DNA were not analysed. Since actual forensic specimens may be ofvariable quantity and subjected to an infinite variety of conditions, it is also valid to determineenvironmental assaults under partially defined conditions that would allow DNA typing analysis.Factors in this study which may be detrimental to the VNTR system include aging andhumidity. These same factors, with the addition of ultraviolet radiation, may also be influentialfactors for the dimorphic RES system. Despite the affects of these elements on both systems,varying degrees of positive results have been obtained with simulated forensic bloodstains. Hence,one does not know for certain whether a particular sample will provide a DNA profile. The valueof testing either real or simulated forensic specimens is to provide guidelines as to what may beexpected and to indicate direction for the further development of test system(s).215REFERENCES1. Stockdale R.E.Science Against CrimePub!. Marshall Cavendish Books Ltd., Londonp30-31 (1982)2. Gaensslen R.E., Desio P.J., and Lee H.C.Forensic Science: Genetic marker systems - Individualization ofBlood and Body fluidsEd. G. Daviesp209-40 (1986)3. Oepen I.Identification of characteristics in blood and semenstains - a reviewForen. Sci mt. 36,p183-91 (1988)4. Kuo M.C.Linking a bloodstain to a missing person by genetic inheritanceJ. Foren. Sci. 27:2,p438-44 (1982)5. Denault G.C., Takimoto H.H., Kwan Q.Y., and Pallos A.Detectability of selected genetic markers in dried blood on agingJ. Foren. Sci. 25:3, p479-498 (1980)6. Jenkins G.C., Brown J., Lincoln P.J., and Dodd B.E.The problem of the acquired B antigen in forensic serologyJ. Foren. Sci. Soc. 12, p597-603 (1972)7. Sensabaugh G.F., Blake E.T., and Northey D.H.Genetic markers in semen. III: Alteration of phosphoglucomutase isozyme patterns insemen contaminated with salivaJ. Foren. Sci. 25:3p470-8 (1980)8. Gaudette B.D.DNA Typing: A new service to Canadian policeRCMP Gazette 52:4, p1-7 (1990)9. Jeifreys A.J., Wilson V., Them S.L.Hypervariable “minisatellite” regions in human DNANature 314:7, p67-73 (1985)10. Jeffreys A.J., Wilson V., and Them S.L.Individual-specific “fingerprints” of human DNANature 316:4, p76-9 (1985)11. Gill P.O., Jeifreys A.J., and Werrett D.J.Forensic application of DNA “fingerprints”Nature 318:12,p577-9 (1985)21612. Kanter W., Baird M., Shaler R., and Balazs I.Analysis of restriction fragment length polymorphisms in deoxyribonucleic acid (DNA)recovered from dried bloodstainsJ. Foren. Sci. 31:2p403-8 (1986)13. Guisti A., Baird M., Pasquale S., et al.Application of deoxyribonucleic acid (DNA) polymorphisms to the analysis of DNArecovered from spermJ. Foren. Sci. 31:2, p409-417 (1986)14. Wainscoat J.S., Pilkington S., Peto T.E.A., Bell J.I.,and Higgs D.R.Allele-specific DNA identity patternsHum. Genet. 75, p384-387 (1987)15. Gill P.A new method for sex determination of the donor of forensic samples using a recombinantDNA probeElectrophoresis 8, p35-8 (1987)16. Gill P., Lygo J.E., Fowler S.J., and Werrett S.J.An evalution of DNA fingerprinting for forensic purposesElectrophoresis 8, p38-44 (1987)17. Fowler S.J., Gill P., Werrett D.J., and Higgs D.RIndividual specific DNA fingerprints from a hypervariable region probe: alpha-globin3’HVRHum. Genet. (1989)18. Haglund W.D., Reay S.T., and Tepper S.L.Identification of decomposed human remains by deoxyribonucleic acid (DNA) profilingJ. Foren. Sci. 35:3p724-9 (1990)19. Gusella J.F.DNA polymorphism and human diseaseAnn. Rev. Biochem. 55,p831-54 (1986)20. Fowler J.C.S., Burgoyne L.A., Scott A.C., and Harding H.W.J.Repetitive deoxyribonucleic acid (DNA) and human genomevariation- A concise review relevant to forensic biologyJ. Foren. Sci. 33:5, p1111-1126 (1988)21. Nei M.Molecular Evolutionary GeneticsColumbia University Press, N.Y.Chapter 6, p143 (1987)22. Alberts B., Bray D., Lewis J., et. al.Molecular Biology of CellGarland Publishing Company, New York,p214-215 (1983)[As reported by Fowler et al., JFS 33:5, p1111 (1988)]21723. Harris H., and Hopkinson D.A.Average heterozygosity per locus in man: an estimate based on the incidence of enzymepolymorphismAnn. Hum. Genet. 36, p9-20 (1972)24. Botstein D., White R.L., Skolnick M., and Davis R.W.Construction of a genetic linkage map in man using restriction fragment lengthpolymorphismsAm. J. Hum. Genet. 32,p314-31 (1980)25. Cooper D.N., Smith B.A., Cooke H.J., et a!.An estimate of unique DNA sequence heterozygosity in the human genomeHum. Genet. 69,p201-5 (1985)26. Jeifreys A.J.DNA sequence variants in the Ggamma, agamma, delta- and beta-globin genes of manCell 18, p1-0 (1979)27. Li W.H. and Graur D.Fundamentals of Molecular Evolutionp67-77Sinauer, Sutherland, Mass. (1991)28. Levitan M.Textbook of Human Genetics - 3rd EditionOxford University Press (1988)p14829. Sambrook J., Fritsch E.F., and Maniatis T.Molecular Cloning: A Laboratory Manual (2nd Edition)p5.3 (1989)Cold Spring Harbor30. Southern E.M.Detection of specific sequences among DNA fragments separated by gelelectrophoresisJ. Mo!. Bio!. 98,p503-17 (1975)31. Watkins P.C.Restriction fragment length polymorphism (RFLP): applications in human chromosomemapping and genetic disease researchBiotechniques 6:4,p310-2 (1988)32. Guse!!a J.F., Wex!er N.S., Conneal!y P.M., et al.A polymorphic DNA marker genetically linked to Huntington’s diseaseNature 306:17,p234-8 (1983)33. St. George-Hys!op P.H., Tanzi R.E., Polinsky R.J., et a!.The genetic defect causing familial Alzheimer’s Disease maps onchromosome 21Science 235, p885-890 (1987)21834. Ray P.N., Belfall B., Duff C., Ct al.Cloning of the breakpoint of an X;21 translocation associated with Duchenne musculardystrophyNature 318:19,p672-5 (1985)35. Tsui L.C., Barker D., Braman J.C., et al.Cystic fibrosis locus defined by a genetically linked polymorphic DNA markerScience 230, p1054-1057 (1985)36. Kan Y.W., and Dozy A.M.Antenatal diagnosis of sickle-cell anaemia by D.N.A. analysis of amniotic-fluid cellsThe Lancet ii, p910-1 (1978)37. Chang J.C., and Kan Y.W.A sensitive new prenatal test for sickle-cell anemiaNew Engl. J. Med. 307:1, p30-2 (1982)38. Woo S.L.C., Lidsky A.S., Guttler F., et al.Cloned human phenylalanine hydroxylase gene allows prenatal diagnosis and carrierdetection of classical phenylketonuriaNature 306:10,p151-5 (1983)39. Darras B.T., Harper J.F., and Francke U.Prenatal diagnosis and detection of carriers with DNA probes in Duchenne’s musculardystrophyNew Engl. J. Med. 316:16, p985-992 (1987)40. Oberle I., Camerino G., Heilig R., et al.Genetic screening for hemophilia A (Classic hemophilia) with a polymorphic DNA probeNew Engl. J. Med. 312:11, p682-686 (1985)41. Wyman A.R., and White R.A highly polymorphic locus in human DNAProc. Natl. Acad. Sci. USA 77:11,p6754-8 (1980)42. Bell G.I., Karam J.H., and Rutter W.JPolymorphic DNA region adjacent to the 5’ end of the human insulin geneProc. Nati. Acad. Sci. USA 78:9, p5759-5763 (1981)43. Bell G.I., Selby M.J., and Rutter W.J.The highly polymorphic region near the human insulin gene is composed of simpletandemly repeating sequencesNature 295:7, p31-5 (1982)44. Higgs D.R., Goodbourn S.E.Y., Wainscoat J.S., et al.Highly variable regions of DNA flank the human a-globin genesNucleic Acids Res. 9:17,p4213-24 (1981)45. Higgs D.R., Wainscoat J.S., Flint J., et al.Analysis of the human a-globin gene cluster reveals a highly informative genetic locusProc. Natl. Acad. Sci. 83,p51659 (1986)21946. Kobilinsky L. and Levine L.Recent application of DNA analysis to issues of paternityJ. Foren. Sci. 33,p1107-8 (1989)47. Waye J.S., Fourney R.M., Bowen J.H, et al.Forensic analysis of restriction fragment length polymorphism: theoretical and practicalconsiderations for design and implementationProceedings for The International Symposium on Human Identification1989-Promega p117-52 (1990)48. Pignatti P.F. and Mango M.Discussion of “Recent application to DNA analysis to issues of paternity”J. Foren. Sci. 35,p517-8 (1990)49. Jeffreys A.J., Brookfleld J.F.Y., and Semeonoff R.Positive identification of an immigration test-case using human DNA fingerprintsNature 317:31,p818-9 (1985)50. Gill P., and Werrett D.J.Exclusion of a man charged with murder by DNA fingerprintingForen. Sci. Internat. 35,p145-8 (1987)51. Wambaugh J.The BloodingPerigod Press/Bamtam Books (1989)52. Newmark P.DNA fingerprinting to be used for British immigrants?Nature 331:18, p556 (1988)53. McGourty C.Genetic tests made official by U.K. courtsNature 339:8 p408 (1989)54. Wong Z., Wilson V., Jeifreys A.J., and Them S.L.Cloning a selectedfragment from a human DNA “fingerprint”: isolation of an extremelypolymorphic minisatelliteNucleic Acids Res. 14:11, p4605-4616 (1986)55. Wong Z., Wilson V., Patel I., Povey S., and Jeffreys A.J.Characterization of a panel of highly variable minisatellites cloned from human DNAAnn. Hum. Genet. 51, p269-288 (1987)56. Jeifreys A.J., Royle N.J., Wilson V., and Wong Z.Spontaneous mutation rates to new length alleles at tandem-repetitve hypervariable loci inhuman DNANature 332:17, pZ78-28l (1988)57. Budowle B., Deadman H.A., Murch R.S., and Baechtel F.S.An introduction to the methods of DNA analysis under investigation in the FBI laboratoryCrime Lab. Digest 15:1 p8-22 (1988)22058. Wetton J.H., Carter R.E., Parkin D.T., and Walters D.Demographic study of a wild house sparrow population by DNA fingerprintingNature 327:14,p147-9 (1987)59. Burke T., and Bruford M.W.DNA fingerprinting in birdsNature 327:14,p149-52 (1987)60. Jeffreys A.J., Wilson V., Kelly R., et al.Mouse DNA “fingerprints”: analysis of chromosome localization and germ-line stability ofhypervariable loci in recombinant inbred strainsNucleic Acids Res. 15:7, p2823-2836 (1987)61. Hill W.G.DNA fingerprinting applied to animal and bird populationsNature 327, p98-99 (1987)62. Wells R.A., Green P., and Reeders S.T.Simultaneous genetic mapping of multiple human minisatellite sequences using DNAfingerprintingGenomics 5,p761-72 (1989)63. Waye J.S., and Fourney R.M.Identification of complex DNA polymorphisms based on variable number of tandem repeats(VNTR) and restriction site polymorphismHum. Genet. 84,p223-7 (1990)64. Lewis R.Witness for the prosecution: DNA fingerprintsDiscoverp44-52 (June 1988)65. Lifecodes CorporationBackground information for DNA-Print identification testTechnical bulletin66. Celimark DiagnosticsCourt cases in the United StatesTechnical bulletin67. Castonguay R.T.MessageCrime Lab. Digest 15:1, p1 (1988)68. Shutler G.G.A perspective on DNA fingerprintingRCMP Gazette 50:1, p22-7 (1988)69. Knoll P.J.An auspicious beginning for DNA genotyping in CanadaPaper presented at the Western Canada Crown SeminarBanff, April 11, 198922170. Thompson W.C., and Ford S.Is DNA fingerprinting ready for the courts?New Scientists 31, p38-43 (1990)71. Lander E.DNA fingerprinting on trialNature 339:15, p501-505 (1989)72. Neufeld P.J., and Colman N.When Science takes the witness standScientific American 262:5, p46-53 (1990)73. Lander E.Expert’s Report on People vs. CastroCourt report74. Ford S., and Thompson W.C.A question of identity: some reasonable doubts about DNA “fingerprints”Sciencesp37-43 (1989)75. Witkowski J.Fingerprint of the future?NEB Transcript 2:1, p2-7 (1989)76. Budowle B., Baechtel S., Giusti A.M., and Monson K.L.Applying highly polymorphic variable number of tandem repeats loci genetic markers toidentity testingClin. Biochem. 23, p287-293 (1990)77. Mudd J.L., and Presley L.A.Quality control DNA typing: a proposed protocolCrime Lab. Dig. 15:4, p109-113 (1988)78. Ferrara P.B.The forensic laboratory manager’s responsibility for quality assurance in DNA analysisCrime Lab. Dig. 17:2, P40-42 (1990)79. Statement of the working group on statistical standards for DNA analysisCrime Lab. Dig. 17:3, p53-58 (1990)80. Kearney J.J., and Mudd J.L.Guidelines for a proficiency testing program for DNA restriction fragment lengthpolymorphism analysisCrime Lab. Dig. 17:3, p59-64 (1990)81. Erickson D.Do DNA fingerprints protect the innocent?Sci. Amer., p18 (August 1991)82. Moss D.C.DNA - The new fingerprintsAm. Bar Asso. J., May 1,p66-70 (1988)22283. Wallace B.Basic Population GeneticsColumbia University PressNew York (1981) Chapter 5, and 884. Ayala F.J.Population and Evolutionary Genetics: A PrimerBenjamin/Cummings Company, Inc.Menlo Park, California 94025 (1982) p6185. Keeton W.T.Biological Science 3rd EditionW.W. Norton & Company, Inc.New York (1980) p755-786. Neufeld P.J., and Scheck B.C.Factors affecting the fallibility of DNA profiling: Is there less than meets the eye?Expert Evidence 1:4, 93-97 (1989)87. Baird M., Balazs I., Giusti A., et al.Allele frequency distribution of two highly polymorphic DNA sequences in three ethnicgroups and its application to the determination of paternityAm. 3. Hum. Genet. 39,p489-501 (1986)88. Balazs I., Baird M., Clyne M., and Meade E.Human population genetic studies of five hypervariable DNA lociAm. 3. Hum. Genet. 44, p182-190 (1989)89. Cohen J.E.DNA fingerprinting for forensic identification: Potential effects on data interpretation ofsubpopulation heterogeneity and band number variabilityAm. J. Hum. Genet. 46, p358-368 (1990)90. Mullis K.B., Faloona F., Scharf S., et al.Specific enzymatic amplification of DNA in vitro: The polymerase chain reactionCSH Quantitative Biol. 51,p263-7 (1986)91. Mullis K.B., and Faloona F.Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reactionMeth. Enzymol. 155,p335-50 (1987)92. Saiki R.K., Scharf S., Faloona F., et al.Enzymatic amplification of B-globin genomic sequences and restriction site analysis fordiagnosis of sickle cell anemiaScience 230,p1350-4 (1985)93. Budowle B. and Baechtel F.S.Modifications to improve the effectiveness of restriction fragment length polymorphismtypingAppl. Theor. Electrophoresis 1,p181-7 (1990)22394. Sambrook J., Fritsch EF., and Maniatis T.Molecular Cloning: A Laboratory Manual (2nd Edition)pE.1O-E.13 (1989)Cold Spring Habor95. Schaffer H.E., and Sederoff R.R.Improved estimation of DNA fragment lengths from agarose gelsAnalyt. Biochem. 115,p113-22 (1981)[Programing modifications by Jeffrey Bonnycastle]96. Southern E.M.Measurement of DNA length by gel electrophoresisAnalt. Biochem. 100,p319-23 (1979)97. Budowle B., Waye J.S. Shutler G.G., et al.Hae Ill-A suitable restriction endonuclease for restriction fragment length polymorphismanalysis of biological evidence samplesJ. Foren. Sd. 35:3,p530-6 (1990)98. Nakamura Y., Leppert M., O’Connell P., et al.Variable number of tandem repeat (VNTR) markers for human gene mappingScience 235,p1616-235 (1987)99. Elder J.K., and Southern E.M.Measurement of DNA length by gel electrophoresis II: comparison of methods for relatingmobility to fragment lengthAnalyt. Biochem. 128,p227-31 (1983)100. Zar J.H.Biostatistical Analysis - 2nd EditionPubi. - Prentice-Hall, Inc., Englewood Cliffs, New JerseyChapter 11, p162-184 (1984)101. Johnson P.H., Miller M.J., and Grossman L.I.Electrophoresis of DNA in agarose gels - II. Effects of loading mass and electroendosmosison electrophoretic mobilitiesAnalt. Biochem. 102,p159-62 (1980)102. Zar J.H.Biostatistical Analysis - 2nd EditionPubi. - Prentice-Hall, Inc., Englewood Cliffs, New JerseyChapter 2, p32 (1984)103. Devlin B., Risch N., and Roeder K.Estimation of allele frequencies for VNTR lociAm. J. Hum. Genet. 48, p662-676 (1991)104. Budowle B. and Monson K.L.A statistical approach for VNTR analysisProceedings of the International Symposium on the Forensic Aspects of DNA Analysis1989 - FBI Academy, Quantico, Virginiap121-35 (1989)224105. Promega Corporation GenePrint dataDNA tying products technical bulletinNo. 3 (1990-91)106. Budowle B., Giusti A.M., Waye J.S. et. al.Fixed-bin analysis for statistical evaluation of continuous distributions of allelic data fromVNTR loci, for use in forensic comparisonsAm. J. Hum. Genet. 48, p841-855 (1991)107. Budowle B., Baechtel S., Giusti A.M., et al.Data for forensic matching criteria for VNTR profilesProceedings for The International Symposium on Human Identification 1989 - Promegap103-115 (1990)108. Lebo R.V., Chakravarti A., Buetow K.H., et al.Recombination within and between the human insulin and B-globin gene lociProc. Nati. Acad. Sci. USA 80,p4808-12 (1983)109. Promega Corporation GenePrint dataDNA tying products technical bulletin,No. 10 (1990-91)110. Budowle B. and Stafford J.Response to expert report by D.C.HartlCrime Lab. Dig. 18:3, p101-108 (1991)111. van Eede P.H., Henke L., Fimmers R., et al.Size calculation of restriction enzyme HaeIII-generated fragments detected by probeYNH24 by comparsion of data from two laboratories: the generation of fragment-sizefrequenciesForen. Sci. mt. 49, p21-3 (1991)112. Statistics Canada: 1981 census of Canada, Vancouver, British ColumbiaOttawa, Ont.: Minister of Supply and Services (1983)113. Fourney R.M., Shutler G.G., Monteith N., et al.DNA typing in the Royal Canadian Mounted PoliceProceedings of the International Symposium on the Forensic Aspects of DNA Analysis1989 - FBI Academy, Quantico, Virginiap137-146 (1989)114. Action R.T., Harman L., Rodney M.S., et al.Comparison of VNTR allele frequencies in White and Black populationsProceedings for The International Symposium on Human Identification1989 - Promega p5-20 (1990)115. Brenner C. and Morris J.W.Paternity index calculations in single locus hypervariable DNA probes: validation and otherstudiesProceedings for The International Symposium on Human Identification1989 - Promega p21-53 (1990)225116. Devlin B., Risch N. and Roeder K.No excess of homozygosity at loci used for DNA fingerprintingSci. 249, p1416-1449 (1991)117. Technical CommentsForensic DNA tests and Hardy-Weinberg EquilibriumSci. 253, plOYl-lO’4l (1991)118. Stoney D.Statistics applicable to the inference of a victim’s blood type from familial testingJ. Foren. Sci. 24:1, p9-22 (1984)119. Tyler M.G., Kirby L.T., Wood S., et al.Human blood stain identification and sex determination in dried blood stains usingrecombinant DNA techniquesForen. Sci. Tnt. 31,p267-72 (1986)120. Yokoi T. and Sagisaka K.Sex determination of blood stains with a recombinant DNA probe: comparison withradioactive and non-radioactive labeling methodsForen. Sci. Tnt. 41,p117-24 (1989)121. Lee H.C., Pagliaro E.M., Berka K.M., et al.Genetic markers in human bone: T. Deoxyribonucleic acid (DNA) analysisJ. Foren. Sci. 36:2,p320-3 (1991)122. Pããbo S.Molecular cloning of Ancient Egyptian mummy DNANature 314:18, p644-645 (1985)123. Perry W.L., Bass W.M., Riggsby W.S., and Sirotkin K.The autodegradation of Deoxyribonucleic acid (DNA) in human rib bone and itsrelationship to the time interval since deathJ. Foren. Sci. 33:1,p144-53 (1988)124. Bar W., Kratzer A., Machier M., and Schmid W.Postmortem stability of DNAForen. Sci. Tnt. 39, p59-70 (1989)125. Saiki R.K., Gelfand D.H., Stoffel S., et al.Primer-directed enzymatic amplification of DNA with a thermostable DNA polymeraseScience 239,p487-91 (1988)126. Embury S.H., Scharf S.J., Saiki R.K.Rapid prenatal diagnosis of sickle cell anemia by a new method of DNA analysisNew Engl. J. Med. 316:11, p656-661 (1987)127. Kogan S.C., Doherty M., and Gitschier J.An improved method for prenatal diagnosis of genetic diseases by analysis of amplifiedDNA sequencesN. Engi. J. Med. 317, p985-990 (1987)226128. Claustres M., Williams C., and Williamson R.Amplification of DNA for detection of cystic fibrosis-linked polymorphismsJ. Ped. 115:5,p749-52 (1989)129. Rosa P.A., and Schwan T.G.A specific and sensitive assay for the Lyme disease spirocheteBorrelia burgdorferi using the polymerase chain reactionJ. Infect. Dis. 160:6, p1018-1029 (1989)130. Demmier G.J., Buffone FJ., Schimbor C.M., and May R.A.Detection of Cytomegalovirus in urine from newborns by using polymerase chain reactionDNA amplificationJ. Infect. Dis. 158:6 p1177-1184 (1988)131. Kasai K.K., Nakamura Y., and White R.Amplification of a variable number of tandem repeats (VNTR) locus (pMCT118) by thepolymerase chain reaction (PCR) and its application to forensic scienceJ. Foren. Sci. 35:5, p1196-1200 (1990)132. Horn G.T., Richards B. and Klinger K.W.Amplification of a highly polymorphic VNTR segment by the polymerase chain reactionNucleic Acids Res. 17:5, p2140 (1989)133. Boerwinkle E., Xiong W., Fourest E., and Chan L.Rapid typing of tndemly repeated hypervariable loci by the polymerase chain reaction:application to the apoliprotein B 3’hypervariable regionProc. Natl. Acad. Sci. USA 86,p212-6 (1989)134. Broderick T.P., Schaff D>A., Bertino A.M., et al.Comparative anatomy of the human APRT gene and enzyme: nucleotide sequencedivergence and conservation of a nonrandom CpG dinucleotide arrangementProc. Natl. Acad. Sci. USA 84,p3349-53 (1987)135. Sasaki H., Yoshioka N., Takage Y., and Sakaki Y.Structure of the chromosomal gene for human serum prealbuminGene 37,p191-7 (1985)136. Wiginton D.A., Kaplan D.J., States J.C., et al.Complete sequence and structure of the gene for human adenosine deaminaseBiochem. 25,p8234-4 (1986)137. Venta P.J.Personal communication 1990 (unpublished)138. Oka K., Tkalcevic G.T., Stocks J., Galton D.J., et al.Nucleotide sequence of Pvu II polymorphic site at the human lipoprotein lipase gene locusNucleic Acids Res. 17:16, p6752 (1989)139. Sensabaugh G.F.Consequences of nucleotide misincorporation during the polymerase chain reactionProceedings of the International Symposium on the Forensic Aspects of DNA Analysis1989 - FBI Academy, Quantico, Virginia227p263-264 (1989)140. Stambrook P.J., Dush M.K., Trill J.J., and Tischfield J.A.Cloning of a functional human adenine phosphoribosyl-transferase (APRT) gene:identification of a restriction fragment length polymorphism and preliminary analysis ofDNAs from APRT-deficient families and cell mutantsSom. Cell and Mo!. Genet. 10,p359-67 (1984)141. Sensabaugh G.Biochemical markers of individualityForensic Science HandbookEd. R. Saferstein (Prentice-Hall)p338-415 (chapter 8) (1982)142. Li S., Oka K., Galton D., and Stocks J.Pvu-II RFLP at the human lipoprotein lipase (LPL) gene locusNucleic Acids Res. 16:5, p2358 (1988)143. Yoshioka k., Sasali H., Yoshioka N., et al.Structure of the mutant prealbumin gene responsible for familial amyloidoticpolyneuropathyMol. Biol. Med. 3,p319-28 (1986)144. Tzall S., Ellenbogen A., Eng F., and Hirschhorn R.Identification and characterization of nine RFLPs at the adenosine deaminase (ADA) locusAm. 3. Hum. Genet. 44, p864-875 (1989)145. Lee B.L., Venta P.J., and Tashian R.E.DNA polymorphism in the S’flanking region of the human carbonic anhydrase II gene onchromosome 8Hum. Genet. 69,p337-9 (1985)146. Jones D.A.Blood samples : probability of discriminationJ.Foren. Sci. Soc. 12,p355-9 (1972)147. Kimura M.The Neutral Theory of Molecular EvolutionCambridge University Press, Cambridgep238 (1983)148. Keeton W.T.Biological Science 3rd EditionW.W. Norton & Company, Inc.New York (1980) p613-617149. Nei M.Molecular Evolutionary GeneticsColumbia University Press, N.Y.Chapter 7p176-5 (1987)228150. Weir B.S.Inferences about linkage disequilibriumBiometrics 35,p235-54 (1979)151. Thompson E.A., Deeb S., Walker D., and Motuisky A.G.The detection of linkage disequilibrium between closely linked markers: RFLPs at theAl-Clil apolipoprotein genesAm. J. Hum. Genet. 42,p113-24 (1988)152. Chamberlain J., Gibbs R.A., Ranier J.E., et al.Deletion screening of the Duchenne muscular dystrophy locus via multiplex DNAamplificationNucleic Acids Res. 16:23, p11141-11156 (1988)153. Gibbs R.A., Chamberlain J.S., and Caskey C.T.Diagnosis of new mutation diseases using the polymerase chain reactionIn PCR Technology: Principles and Applications for DNA Amplificationp171-9 (1989)Ed. H.A. ErlichStockton Press, New York, N.Y.154. Jeffreys A.J., Wilson V., Wong A., et al.Multilocus and single locus minisatellite DNA probes in forensic medicineAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginiap77-85 (June 1989)155. Budowle B., Chakraborty R., Giusti A.M., et al.Analysis of the VNTR locus D1S8O by the PCR followed by high-resolution PAGEAm. J. Hum. Genet. 48,p137-44 (1991)156. Cheung L.L.C., McNeil B.K., Ferris J.A.J. and Autor A.P.Determination of twin zygosity by DNA genetic typingAbstractCanadian Society of Forensic Sciences MeetingToronto, Ont., CanadaOct. 18-21, 1988157. Chimera J.A., Harris C.R. and Litt M.Population genetics of the highly polymorphic locus D16S7 and its use inpaternity evaluationAm. J. Hum. Genet. 45,p926-31 (1989)158. Akane A., Matsubara K. et. al.Paternity testing: blood group systems and DNA analysis by variable number of tandemrepeat markersJ. Foren. Sci. 35,p1217-25 (1990)159. Bartlett S.E. and Davidson W.S.Forensically informative nucleotide sequences (FINS) identify different species of tunaPoster presentationCanadian Society of Forensic Science229Ottawa, Ont., CanadaOct. 1-5, 1990160. Harris A.S., Bentzen P. and Wright J.M.Cloning of hypervariable loci from fish DNAPoster presentationCanadian Society of Forensic ScienceOttawa, Ont., CanadaOctober 1-5, 1990161. Fey M.F., Wells R.A., Wainscoat J.S. and Them S.L.Assesment of clonality of gastrointestinal cancer by DNA fingerprintingJ. Clin. Invest. 82:5,p1532-7 (1988)162. Päabo S.Amplil’ing ancient DNAPCR Protocols - A Guide to Methods and ApplicationsEd.- M.A.Innis, D.H.Gelfand, J.J.Sninsky, and T.J. WhiteAcademic Press, Inc., San Diego, CaliforniaChapter 20, p159-166 (1990)163. Paãbo S., Gifford J.A., and Wilson A.C.Mitochondrial DNA sequences from a 7000-year old brainNucleic Acids Res. 16:20,p9775-87 (1988)164 Päãbo S.Ancient DNA: Extraction, characterization, molecular cloning, and enzymatic amplificationPNAS 86,p1939-43 (1989)165. Reynolds R., von Beroldingen C., and Sensabaugh G.F.Effects of DNA degradation on amplification by the polymerase chain reactionAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginia, p257 (June 1989)166. Comey C.T.The use of DNA amplification in the analysis of forensic evidenceCrime Lab. Dig. 15:4,p99-103 (1988)167. VonBeroldingen C., Walsh S., Roby R., and Sensabaugh G.DNA in hairAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginiap265-6 (June 1989)168. Blake E.T., Pääbo S., and Stolorow M.D.DNA amplification and typing from aged biological evidenceAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginiap26’7-268 (June 1989)169. Walsh P.S., Metzger D.A., and Higuchi R.230Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensicmaterialBiotechniques 10:4, p506-513 (1991)170. McNally L., Shaler R.C., Baird M., et al.Evaluation of deoxyribonucleic acid (DNA) isolated from human bloodstains exposed toultraviolet light, heat, humidity, and soil contaminationJ. Foren. Sci. 34:5, p1059-1069 (1989)171. McNally L., Shaler R.C., Baird M., et al.The effects of environment and substrata on deoxyribonucleic acid (DNA): the use ofcasework samples from New York cityJ. Foren. Sci. 34:5, plO’7O-lO’7’7 (1989)172. Buoncristiani M., von Beroldingen C., and Sensabaugh G.Effects of UV damage on DNA amplification by the polymerase chain reactionAn International Symposium on the Forensic Aspects of DNA Analysis FBI Academy,Quantico, Virginiap259 (June 1989)173. Gelfand D.H.PCR Technology: Taq DNA polymerase-Chapter 2Ed. H.A. ErlichPub!. Stockton Press 1989p20174. Harding J.D., Gebeyehu 0., Bebee R. et al.Rapid isolation of DNA from complex biological samples using a novel capture reagentmethidium-spermine-sepharoseNucleic Acids Res. 17, 694758 (1989)175. Singer-Sam J., Tanguay R.L., and Riggs A.D.Use of Chelex to improve the PCR signal from a small number of cellsAmplifications : A forum for PCR usersIssue 3, p11 (September 1989)176. Walsh S. and Higuchi R.PCR inhibition and bloodstainsAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginiap281-282 (June 1989)177. Environment CanadaMonthly meteorological summarysTaken at Vancouver International Airport, B.C.October 1990 - December 1990178. Konzak K.C., Reynolds R., von Beroldingen C., et al.Effects of DNA damage and degradation of RFLP analysisAn International Symposium on the Forensic Aspects of DNA AnalysisFBI Academy, Quantico, Virginiap255 (June 1989)231179. Adams D.E.Validation of the FBI procedure for DNA analysis: a summaryCrime Lab Dig. 15:4, p106-108 (1988)180. Stryer L.BiochemistryW.H. Freeman and Company, San FranciscoSecond edition, p636 (1981)181. Blackburn G.M. and Gait M.J.Nucleic Acids in Chemistry and BiologyIRL Press, Oxford University Press, New Yorkp265 (1990)182. Greater Vancouver Regional DistrictAmbient Air Quality - Annual Report1989183. Greater Vancouver Regional DistrictAmbient Air Quality - Quarterly Report1990184. Becker M.M. and Wang Z.Origin of ultraviolet damage in DNAJ. Mol. Biol. 210,p429-38 (1989)


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items