Identity Assessment

Thomas M. Williams, M.D., Howard J. Baum, Ph.D., Victor W. Weedn, M.D., J.D. and Malek Kamoun, M.D., Ph.D.

Identity testing exploits variations present within the human genome to distinguish among individuals. Identity assessment has six basic uses: (1) to confirm or refute that a sample is from a specific person in forensic testing; (2) to identify unknown human remains or victims of a mass disaster; (3) to resolve questions regarding the identity of a clinical specimen; (4) to select donors for a planned transplant recipient to minimize rejection and improve graft survival via histocompatibility testing; (5) to assess whether hematopoietic cells are donor- or recipient-derived following stem cell transplantation; and (6) to identify the parents of a child.

Variation in the Human Genome

Identity testing began with the use of serologic methods to identify variations in proteins that differ among individuals. The discovery of the genetic basis for these protein differences and of genetic variability at loci not encoding proteins, coupled with technical advances, allowed the field to move to direct analysis of DNA. Genetic variation among individuals is extensive, with about one sequence difference for every 400 to 1250 nucleotides on autosomal chromosomes. Variants of a genetic locus in a population are referred to as alleles. A locus is said to be polymorphic when the least common allele has a frequency ≥0.01 in a population. Although several alleles may be found in a population for an autosomal locus, an individual may have at most two alleles at that locus. Individuals may have one or two alleles for X-linked loci.

Genetic Variation Useful in Identity Testing

Several classes of genetic variants are found in the genome; some are more useful than others for identity testing. Most of the variants used occur in the noncoding genetic regions, such as introns, regulatory domains, and regions between genes, whereas some variants occur in gene domains transcribed into RNA, that is, the exons. Highly repetitive sequence elements that contribute to the structure of centromeres and telomeres and hundreds of thousands of copies of transposable elements that move about the genome over time may vary among individuals. However, these repetitive sequence elements generally are not useful for identity testing. Several million single-nucleotide polymorphisms (SNPs) have been identified in the genome (see Chapter 38). A subset of SNPs can be identified based on the ability of a restriction endonuclease to digest double-stranded DNA at the site of the variation. These SNPs are referred to as restriction fragment length polymorphisms (RFLPs). SNPs and most RFLPs are not very useful for identity testing because they usually have only two alleles.

Variable numbers of tandem repeat loci (VNTRs) or minisatellite loci consist of repeated sequences of DNA. The core sequence is from 8 to 80 nucleotides long. The core is repeated from 4 to 40 times, thus forming 4 to 40 alleles. The allele size difference can be detected as an RFLP if the locus containing the VNTR is digested with a restriction endonuclease, which cuts the DNA outside of the VNTR, and the resulting restriction fragments are hybridized to a labeled DNA probe in Southern hybridization assays. VNTRs are attractive for identity testing, because the loci usually have a number of different alleles with relatively high allele frequencies. Minisatellite regions are commonly near the telomeric end of the chromosome and have core repeats of 8 to 80 base pair lengths, resulting in DNA fragment lengths of 0.5 to >20 kilobases.

Short tandem repeat (STR) or microsatellite loci consist of DNA sequence motifs that have core repeats of two to seven base pairs.6,15 Examples include the dinucleotide 5′ CACACACA 3′ and the tetranucleotide 5′ TTTATTTATTTA 3′. Thousands of STRs are scattered throughout the genome. Because they are flanked by unique sequences, each can be specifically amplified with the polymerase chain reaction (PCR) for analysis. In populations of individuals, multiple alleles may be present based on differences in the numbers of repeated motifs at the locus. STRs have many characteristics that make them ideal for identity testing: (1) they can be analyzed in fluorescent automated systems; (2) alleles can be assigned in a definitive manner following analysis; (3) STR loci are almost always transmitted in families in a Mendelian fashion; (4) the loci may have 10 or more alleles, often with substantial allele frequencies, making them highly informative and making it easier to resolve mixtures of DNA; and (5) extensive information is available about allele frequencies in many human populations for STRs commonly used in identity testing.⁷

Commercially available STR systems employ tetrameric and pentameric repeat loci, which produce fewer artifactual bands and are characterized by roughly equal amplification of both alleles within an individual (Figure 41-1). Fragments can be labeled during PCR amplification with fluorescently tagged primers that facilitate multiplexing.

Figure 41-1 Identity testing via short tandem repeat (STR) analysis. An STR locus can be specifically amplified with the polymerase chain reaction (PCR) using primers binding to unique sequences adjacent to the repeat motif in the genome. In the example shown, the polymorphic repeat is the tetramer TTTC. Three example alleles with 6, 7, and 9 repeats at this locus are shown. The genotype for an individual or for an evidence sample can be determined by performing PCR with a fluorescently labeled primer and sizing the products on a high-resolution gel with laser detection. PCR products should differ in size from each other by multiples of 4 nucleotides (nt). Use of several fluorescent dyes and PCR products of different sizes allows multiplexing for simultaneous assessment of a number of independently segregating STR loci.

Apart from STRs distributed across the whole genome, two special genetic regions with sufficient sequence variability for identity testing include the human leukocyte antigen (HLA) loci within the major histocompatibility complex (MHC) and mitochondrial DNA. The HLA loci described in the “Transplantation Testing” section later in this chapter are interesting in that the polymorphisms are densely packed and are preferentially located in the exons rather than the introns of these genes on chromosome 6. Mitochondrial genome variation is also useful in forensic identity testing and is described in that section.

Forensic Dna Typing

DNA testing has revolutionized criminalistics.19,33 Only fingerprint evidence can sometimes rival the ability of DNA as trace evidence left at a scene to identify a perpetrator. As a general rule, other trace evidence merely links an article, instrument, or material to a scene. The origin of DNA-based identity testing is generally traced to a 1985 article in Nature by Alec Jeffreys.¹⁸ He coined the term “DNA fingerprint” and suggested that the hybridization of DNA probes to polymorphic genetic loci could be exploited for forensic purposes. Jeffreys first applied his techniques to civil and criminal cases in England. In the United States, DNA-based identity testing was introduced via commercial laboratories and later the Federal Bureau of Investigation (FBI). Today approximately 200 forensic DNA typing laboratories have been established in the United States, along with many other DNA laboratories around the globe. Forensic DNA testing is also used to identify decomposed unknown human remains through kinship analysis and can be used to identify victims of a mass disaster.⁴

Forensic Applications

Forensic testing differs from clinical laboratory testing in several ways: (1) the forensic question is usually one of identity rather than one of presence or absence of a trait or analyte quantification, as is done in most clinical laboratory analyses; (2) specimens received by forensic laboratories are much more diverse than the typical blood, fluid, and tissue samples handled by clinical laboratories; (3) clinical samples are collected under controlled circumstances, while evidence from which DNA must be isolated may be exposed to the environment in a variety of ways. This can lead to degradation of the sample. Experiments may be necessary to validate testing for a particular case; (4) forensic samples may include a mixture of DNA such as a vaginal swab containing female epithelial cells and sperm from one or more donors. In addition, a surface may contain more than one biological fluid such as blood from a victim and saliva from someone else; (5) evidentiary material cannot be replenished and may be present in only trace amounts. Testing may consume the sample, and thus complete or repeat testing may be impossible; and (6) forensic identity testing is scrutinized in a judicial environment, requiring complete accounting for chain of custody following its collection and strict validation of procedures.

Most other laboratories perform routine analyses of samples collected in defined ways. Forensic identity testing must contend with much greater variability in samples and testing conditions.

Genetic Systems Used in Forensic Identification

Numerous genetic systems that are employed by forensic laboratories are summarized in Table 41-1.^3,24

TABLE 41-1

DNA Typing Systems and Their Characteristics

mtDNA, Mitochondrial DNA; PCR, polymerase chain reaction; RFLP, restriction fragment length polymorphism; SNPs, single-nucleotide polymorphisms; STRs, short tandem repeats; VNTRs, variable numbers of tandem repeat loci.

VNTR Analysis by RFLP

Jeffreys described a method to create a barcode-like DNA fingerprint based on VNTRs. Because the probes used bind to several loci in Southern hybridization experiments and produce numerous bands per probe, they are termed multilocus probes. In the United States, laboratories have preferred single-locus probe (SLP) systems that hybridize only to a single genetic locus. However, Southern hybridization–based RFLP analysis is expensive, labor intensive, difficult to automate, and less sensitive than PCR-based methods. Finally, RFLP is sensitive to DNA degradation—an important issue with environmentally exposed forensic specimens. RFLP tests have been largely abandoned in favor of the more efficient PCR-based assays discussed later.⁵

Short Tandem Repeats

Most identity testing performed today relies on the PCR. PCR testing is inherently sensitive, allowing routine analysis of nanogram quantities of genomic DNA and often successful testing of picogram quantities (one cell contains 5 to 10 pg of DNA). Low copy number (LCN) STR analysis ⁸ detects quantities of DNA down to the single cell level PCR, which underlies the characterization of STR and other loci for forensic identity testing loci described in this and later sections.

STR testing is quick, less expensive, more forgiving with respect to technical skills needed, less sensitive to DNA degradation, and more amenable to automation in comparison with the Southern hybridization methods described earlier. Although less discriminating than RFLP genetic markers, STR analysis can be made as powerful as Southern RFLP analysis through the use of large numbers of informative loci.

The National Institute of Justice provided funding for the initial application of STRs in forensics. STRs were used in forensic casework during the first Persian Gulf War and were widely adopted for testing by forensic laboratories in the United Kingdom and the United States in the mid to late 1990s.

The FBI laboratory’s combined DNA index system (CODIS) blends forensic science and computer technology into an effective tool for investigating violent crimes. CODIS enables federal, state, and local crime laboratories to exchange and compare DNA profiles electronically, thereby linking crimes to each other and to convicted offenders. The FBI convened a panel of forensic scientists in 1998 that chose a panel of 13 STR loci for use in the National DNA Index System. These 13 core loci have become the standard for casework and databanking for most forensic laboratories around the world (Table 41-2). They have been commercialized as kits in a variety of formats. STRs are now routinely used in crime laboratories globally and typically yield discriminatory values of one in trillions to sextillions.

TABLE 41-2

List of Short Tandem Repeat (STR) Core Loci

The 13 STR core loci are listed against the STR systems present in currently available commercial STR kits. Identifiler and MiniFiler are available from Applied Biosystems, Inc. (Foster City, Calif), and Powerplex 16 is available from Promega Corporation (Madison, Wis).

Low copy number STR testing is currently being used in the United Kingdom, Australia, New Zealand, and the United States. Its extreme sensitivity is both an advantage and a disadvantage. Very small samples can be detected, but because of the small amounts of DNA analyzed, the system suffers from stochastic effects. Contamination and secondary transfer are potential problems with this system. Another approach to increase the sensitivity of degraded DNA samples is the use of mini STRs, which essentially are the same tandem repeats used in the commercial kits described previously, but the flanking PCR primers are moved closer to the tandem repeats. This results in amplification of smaller fragments, which decreases the likelihood that degradation will affect an STR. Greater sensitivity has been seen with mini STRs.

Gender Markers and Y Chromosome Markers

Amelogenin is a low molecular weight protein found in tooth enamel. The amelogenin gene is useful as a gender marker. The X chromosome amelogenin gene differs from its homolog on chromosome Y by a six–base pair polymorphism, allowing the distinction between individuals with 46,XY and 46,XX karyotypes. Males will display amelogenin locus heterozygosity; females will exhibit homozygosity. Reagents for assessing the amelogenin locus are incorporated into commercially available STR kits.

Y chromosome polymorphic loci can be used as identifying loci found only in males. In this way, male-specific DNA obtained from a vaginal swab can be typed without the usual differential extraction, in which the DNA from spermatozoa is released after the female fraction has been isolated from epithelial cells. Y chromosome loci useful for identity testing include STRs or SNPs (described later in this chapter). Laboratories typically employ commercially available panels of 12 to 17 Y chromosome STRs for analyses (Table 41-3). Y chromosome SNPs are in development.

TABLE 41-3

Y Chromosome Markers *

^*Y chromosome markers are listed in their major haplotype groups.

Y chromosome polymorphic loci are linked, resulting in discriminatory power that is significantly less than that of a panel of independently segregating somatic STR loci. Discriminatory values can be increased by using a large panel of Y chromosome markers in conjunction with a large database of typed individuals.

Mitochondrial DNA

Mitochondrial genomes are circular double-stranded DNA molecules that are 16,569 bp long and are present as one or more copies within the mitochondria of a cell. Thus mitochondrial DNA (mtDNA) is present in hundreds to thousands of copies per cell. mtDNA, unlike chromosomal DNA, does not undergo meiosis and does not participate in genetic recombination events. mtDNA remains stable over generations, except for the acquisition of mutations at a rate 10 to 20 times that of nuclear DNA.

mtDNA is transmitted to children via oocytes. Although it is generally thought that mitochondrial DNA is exclusively derived from the mother, a minor contribution from the father is occasionally present, particularly in disease states. The normal state of mitochondria is generally thought to be one of homoplasmy, in which all the mtDNA has the same sequence. However, because of mutational events, a state of heteroplasmy, in which more than one mtDNA sequence is present in the same tissue, may exist. High-level heteroplasmy is generally on the order of 30% of the mtDNA sequence before it is reported. Unrecognized low-level heteroplasmy is common. Heteroplasmy appears to be somewhat tissue-specific rather than uniform throughout the body. Thus two shed hairs may show discrepant mtDNA sequences. Because of heteroplasmy, one or two nucleotide mismatches between two individuals are not an absolute basis for exclusion of a tested individual.

In the human mitochondrial genome, only approximately 1200 bases in the region of transcription origin (15971-579), known as the displacement loop (D-loop) or the control region, are noncoding. This D-loop consists of two hypervariable regions that contain the majority of polymorphisms useful for identity (HVI: 16024-16365; HVII: 73-340). Polymorphisms outside this region can also be employed for testing. mtDNA polymorphisms are typically identified for forensic testing via DNA sequencing of hypervariable regions.¹⁵ This method is expensive, labor intensive, and highly sensitive to contamination.

The mtDNA sequence obtained from a specimen is compared with a reference sequence (revised Cambridge sequence, www.mitomap.org/MITOMAP, accessed May 24, 2010). As in the case with Y chromosome STRs, because mtDNA polymorphisms are linked, individual polymorphism frequencies cannot be multiplied together to generate a likelihood of identity such as by independently segregating chromosomal locus allele frequencies. Instead the mtDNA haplotype identified in a sample is compared with those deposited in a database to derive a frequency statistic. Many mitochondrial haplotypes in the database are unique. Because the database has more than 6000 entries, it can be fairly stated that many mitochondrial haplotypes have a discriminatory value greater than 1 in 6000. However, 18 common haplotypes have population frequencies greater than 0.5%, including a haplotype present in 7% of the population. In aggregate, these common haplotypes account for 20% of all haplotypes.

mtDNA is useful primarily for identity testing in four contexts. First, a sample may be available that contains mitochondrial but not nuclear DNA. For example, shed hairs that do not have roots generally contain only mtDNA. Second, when the DNA within a specimen, such as skeletal remains, is substantially degraded, the high copy number and small size of mtDNA make it more likely to yield a result than nuclear DNA. Third, mtDNA analysis may become essential when only a distant relative is available for a reference specimen. In this example, nuclear DNA requires samples from multiple close kindred, but mtDNA matching would require only a distant maternal relative. Fourth, in database searches of unidentified human remains or missing person relatives, the algorithms for the search often produce several matches, and mitochondrial DNA analysis is needed to identify the true match.