Mitochondrial DNA (mtDNA) provides a valuable locus for forensic DNA typing in certain circumstances. The high number of nucleotide polymorphisms or sequence variants in the two hypervariable portions of the non-coding control region can allow discrimination among individuals and/or biological samples. The likelihood of recovering mtDNA in small or degraded biological samples is greater than for nuclear DNA because mtDNA molecules are present in hundreds to thousands of copies per cell compared to the nuclear complement of two copies per cell. Therefore, muscle, bone, hair, skin, blood and other body fluids, even if degraded by environmental insult or time, may provide enough material for typing the mtDNA locus. In addition, mtDNA is inherited from the mother only, so that in situations where an individual is not available for a direct comparison with a biological sample, any maternally related individual may provide a reference sample.

A mtDNA analysis begins when total genomic DNA is extracted from biological material, such as a tooth, blood sample, or hair. The polymerase chain reaction (PCR) is then used to amplify, or create many copies of, the two hypervariable portions of the non-coding region of the mtDNA molecule, using flanking primers. Primers are small bits of DNA that identify and hybridize to or adhere to the ends of the region one wishes to PCR amplify, therefore targeting a region for amplification and subsequent analysis. Care is taken to eliminate the introduction of exogenous DNA during both the extraction and amplification steps via methods such as the use of pre-packaged sterile equipment and reagents, aerosol-resistant barrier pipette tips, gloves, masks, and lab coats, separation of pre- and post-amplification areas in the lab using dedicated reagents for each, ultraviolet irradiation of equipment, and autoclaving of tubes and reagent stocks. In casework, questioned samples are always processed before known samples and they are processed in different laboratory rooms. When adequate amounts of PCR product are amplified to provide all the necessary information about the two hypervariable regions, sequencing reactions are performed. These chemical reactions use each PCR product as a template to create a new complementary strand of DNA in which some of the As, Ts, Cs, and Gs (nucleotide bases) that make up the DNA sequence are labeled with dye. The strands created in this stage are then separated according to size by an automated sequencing machine that uses a laser to "read" the sequence, or order, of the nucleotide bases. Where possible, the sequences of both hypervariable regions are determined on both strands of the double-stranded DNA molecule, with sufficient redundancy to confirm the nucleotide substitutions that characterize that particular sample. At least two forensic analysts independently assemble the sequence and then compare it to a standard, commonly used, reference sequence. The entire process is then repeated with a known sample, such as blood or saliva collected from a known individual. The sequences from both samples, about 780 bases long each, are compared to determine if they match. The analysts assess the results of the analysis and determine if any portions of it need to be repeated. Finally, in the event of an inclusion or match, the SWGDAM mtDNA database, which is maintained by the FBI, is searched for the mitochondrial sequence that has been observed for the samples. The analysts can then report the number of observations of this type based on the nucleotide positions that have been read. A written report is provided to the submitting agency.

While mtDNA is useful for forensic examinations, it has also been used extensively in two other major scientific realms. First, there are a number of serious human diseases caused by deleterious mutations in gene-coding regions of the mtDNA molecule, which have been studied by the medical profession to understand their mode of inheritance. In addition, molecular anthropologists have been using mtDNA for almost a decade to examine both the extent of genetic variation in humans and the relatedness of populations all over the world. Because of its unique mode of maternal inheritance it can reveal ancient population histories, which might include migration patterns, expansion dates, and geographic homelands. Recently mtDNA was extracted and sequenced from a Neanderthal skeleton. These results allowed anthropologists to say with some conviction that modern humans do not share a close relationship with Neanderthals in the human evolutionary tree. While all the applications of mtDNA, including forensic, are relatively recent, the general methods for performing a mtDNA analysis are identical to those used in molecular biology laboratories all over the world for studying DNA from any living organism. There have been over a thousand published articles regarding mtDNA.

MtDNA has advantages and disadvantages as a forensic typing locus, especially compared to the more traditional nuclear DNA markers that are typically used. As mentioned above, mtDNA is maternally inherited, so that any maternally related individuals would be expected to share the same mtDNA sequence. This fact is useful in cases where a long deceased or missing individual is not available to provide a reference sample but any living maternal relative might do so. Because of meiotic recombination and the diploid (bi-parental) inheritance of nuclear DNA, the reconstruction of a nuclear profile from even first degree relatives of a missing individual is rarely this straightforward. However, the maternal inheritance pattern of mtDNA might also be considered problematic. Because all individuals in a maternal lineage share the same mtDNA sequence, mtDNA cannot be considered a unique identifier. In fact, apparently unrelated individuals might share an unknown maternal relative at some distant point in the past.

At the present time the available forensic database of human mitochondrial DNA sequences has around 4800 sequences available for a search of a casework sequence. The current convention in the event of an inclusion (a match between questioned and reference sample sequences) is for the analyst to report the number of times the observed sequence is present in the database to provide some idea of its relative frequency in the database. A frequency statistic may also be used, and a 95% or 99% confidence interval is placed around the calculated frequency to account for the inherent uncertainty in the frequency calculation. While most types appear to be rare or at least infrequent in each of the racial databases (African or African-origin, Asian or Asian-origin, Caucasian or European-origin, and Hispanic), there is one type which is seen in around 7% of Caucasians. However, almost two thirds of newly-typed samples have novel sequences, so we have not yet uncovered all the variation present in the general human population. For novel types, a 95% or 99% upper bound frequency calculation may be performed. In general, the pattern observed in most populations around the world, with the exception of a few populations of anthropological interest, is that the vast majority of sequences is uncommon, and relatively few types present at frequencies greater than 1% in the databases. Because of this fact, it will be possible to exclude greater than 99% of the population as potential contributors of a sample in most cases, except where one is dealing with a more "common" type.

In contrast, a multilocus nuclear DNA typing profile provides vastly superior discriminatory power, such that we can now approach the possibility that a typed individual has a unique profile with respect to any other person in the world. Therefore, mtDNA can never provide the resolution of individuality that nuclear typing can. For this primary reason, it should be reserved for cases or samples for which nuclear typing is simply not possible. Candidates for mtDNA typing analyses would most likely be: 1) shed hairs with no follicle, tissue, or root bulb attached, 2) hair shaft fragments, 3) bones or teeth which have been subjected to long periods of high acidity, high temperature, or high humidity, 4) stain or swab material that has been previously unsuccessfully typed for nuclear markers, and 5) tissue (skin, muscle, organ) that has been previously unsuccessfully typed for nuclear markers. Hair roots, when available, should be removed from the shaft and processed separately for nuclear DNA markers prior to attempting mtDNA analysis on the hair shaft. Hair shafts or fragments are only suitable for mtDNA analysis as they can contain fewer than 100 copies of the mtDNA molecule and virtually no nuclear DNA. The same is generally true for older skeletal remains. While mtDNA typing of blood stains is possible, it is more likely that mixtures will be obtained, due to the extreme sensitivity of this form of typing in samples that unlike hairs and bones are difficult to clean before DNA extraction.

Finally, it must be noted that mtDNA analyses are the most rigorous and time-consuming of DNA forensic analyses. Based on informal statistics available from all laboratories performing these typings, the rate of throughput is approximately 1-2 cases/analyst/month. The reasons for this include: 1) small/degraded samples requiring numerous PCR reactions to obtain sufficient DNA template for sequencing, 2) exhaustive procedures to control for contamination, and 3) sequencing analyses of both strands of DNA in both hypervariable regions. In addition, for some types of samples, especially hairs, mtDNA analysis is more likely to consume the whole sample than nuclear DNA typing. For example, a single mtDNA analysis could be performed on a 0.5-2 cm hair fragment. A 4 cm fragment could have duplicate testing for confirmation of the sequence. In both cases the fragment would be totally consumed. However, a root ball, follicle, or skin tissue attached to a hair would also be consumed in a nuclear typing effort. For both mtDNA and nuclear DNA testing there is a possibility that sufficient extracted DNA might remain for duplicate testing in another lab. Swatch, swab, stain, bone, and tooth analyses are less likely to consume all material, as these samples can often be divided, although the difficulties of obtaining enough DNA for analysis could result in consumption of these materials as well. For the reasons above, pre-analysis documentation (microscopy, photography) is desirable.

Most importantly, mitochondrial DNA testing should only be performed by laboratories with considerable experience in handling the unusually difficult samples that require this form of testing. The primary reason for this is that experienced labs can extract minimal amounts of mtDNA from difficult samples. In the event of a sample failure, an inexperienced lab would never know whether their extractions and PCRs were simply not sensitive enough, or whether the sample lacked non-degraded DNA altogether. In addition, contamination controls are heightened in a mitochondrial DNA laboratory, where working at the limits of sensitivity is standard operating procedure.

Because of the advantages and in spite of the limitations mentioned above, mtDNA analysis has found a place in the forensics arena. Several dozen cases have been tried in US courtrooms using mtDNA evidence to augment more traditional forms of evidence, and several post-conviction exonerations have been obtained in cases where microscopically examined hairs have recently been analyzed for mtDNA. All appellate decisions handed down to date have upheld mtDNA testing and written decisions may be viewed at www.denverda.org. MtDNA forensic testing should be utilized primarily in situations where nuclear DNA typing is not an option, or in the event that nuclear typing has been attempted and is unsuccessful. In these cases, mtDNA typing can provide additional information about the relationship of an individual to a biological sample heretofore unavailable.

— Copyright 2002 by Terry Melton, Mitotyping Technologies, LLC