Books have been written on the subject of ancient DNA and the community has organized itself. An informal newsletter is distributed and two international conferences on ancient DNA have been held (Nottingham, England, 1991; Washington, D.C. 1993). Jurassic Park was written by Michael Crichton from the machinations of the Extinct DNA Study Group at the University of California-Berkeley, a founding group of this community.

There is a background of important scientific work from which to draw in performing DNA analyses of ancient skeletal remains. The common ground between all scientific efforts in the field of ancient DNA and military skeletal remains identification of the Vietnam and Korean conflicts is the extraction of DNA information from samples in which the DNA is so extensively broken into small fragments and extensively damaged. The techniques and concerns of molecular anthropologists in the analysis of ancient DNA are applicable to military skeletal remains identification efforts.

G. Polymerase Chain Reaction

DNA from skeletal remains is degraded or broken down into very small fragments. This is particularly true of ancient DNA, where the average fragment size may be less than one hundred bases. Furthermore, most of the original DNA has been destroyed or washed away so that very low quantities of DNA are present and recovering DNA sequence information from old skeletal remains poses a significant technical challenge.

The polymerase chain reaction (PCR) is a method to amplify a target region of DNA. PCR is an enzymatic reaction resulting in the exponential production of copies of a given DNA segment, where the copies produced can be themselves copied. A million-fold increase in the number of copies of target DNA is often accomplished by PCR, permitting further analysis. Accordingly, PCR-based testing is exquisitely sensitive. This PCR amplification technique has revolutionized DNA testing and all molecular biology. PCR is of such significance that Dr. Kary

Mullis was awarded the 1993 Nobel Prize in Chemistry for its discovery. DNA testing of minimal and degraded DNA of ancient skeletal remains is made possible by the PCR method of amplification.

H. Military Skeletal Remains Identification

Substantial medical and/or dental records and nonDNA evidence exist to identify the majority of remains recovered from Southeast Asia using traditional methods; mtDNA testing would not be required in most cases where intact remains can be recovered. It is estimated by CMAOC that approximately 500 cases would require DNA tests, however, the number is unavoidably speculative. The AFDIL is currently resourced to perform mtDNA testing at a rate of ten cases per month. These skeletal remains involve cases with name associations and living family members are known in nearly all cases. The AFDIL will normally obtain the bone mtDNA sequences from the remains and then compare them to family reference blood specimens. Many families have already donated blood specimens in hopes that they may be useful for identification of their family member. The identification efforts using mtDNA have been highly successful and are anticipated to continue for the next five years.

The United Nations Command Military Armistice Commission (UNCMAC), representing the 16 nations that supported the South Korean government in the Korean conflict, has continued to press for the repatriation of the remains of UNC servicemembers since the termination of hostilities in 1953. During Operations Little Switch and Big Switch in 1953, a total of 3,748 U.S. POWs (out of a UNC total of 13,457) were repatriated. In 1954, the remains of 1,868 U.S. servicemen (out of a UNC total of 4,023) were repatriated in Operation Glory. Of these remains, 866 are declared unknown and 865 were buried in the National Memorial Cemetery of the Pacific (Punchbowl), Honolulu, Hawaii.

The total of over 8,100 U.S. servicemembers (out of a UNC total of over 10,200) includes those remains that have not been recovered, and were buried in known UNC cemeteries in North Korea, lost or buried at sea, and others who were unaccounted for with the body not recovered. Also included are 389 personnel (out of a UNC total of 2,233) about whom the Korean People's Army (KPA) and the Chinese People's Volunteers (CPV) should have knowledge. Information gathered from intelligence sources and

POW debriefings suggested that these were individuals who were possibly captured and died under KPA/CPV control.

In May 1990, for the first time since 1954, the KPA repatriated remains. Over 200 remains have been turned over in recent years. However, the change of power in North Korea has brought uncertainty into the future of remains repatriation.

The poor condition of remains recovered from Korea, the lack of records necessary to make an identification, and the inability to make joint recoveries have largely impeded CILHI's ability to identify these remains using traditional methods. In 1973, about 800-. of the necessary medical and personnel records were destroyed in a fire at the National Personnel Records Center in St. Louis, Missouri. Existing records of that era provide a dearth of medical and dental information. Recent repatriations include many cases in which the midfacial portion of the skull is missing and thus the most useful anthropologic features for identification are absent. Accordingly, mtDNA testing is anticipated to be used in most Korean cases.

The Casualty Data section of CILHI has established preliminary figures for the number of recoveries through joint United States/KPA recovery operations and investigations in North Korea. There is sufficient information currently existing to establish a known location, where the recovery of remains is possible for 2,400 of 6,000 remains located in North Korea. These include 1,612 reported burials in former POW camps with known locations, 181 reported interments in known temporary cemeteries north of the demilitarized zone (DMZ), and 633 known aircraft loss sites north of the DMZ. In an additional 568 incidents, sufficient information is known which may lead to remains recovery. These include 535 reported POW camp burials with no known location, 1 reported interment in a cemetery with no location, and 32 aircraft losses without a fixed or general location. Most cases represent burials by American personnel during advances before reoccupation by the KPA. Accordingly, CILHI estimates the upper limit of the reasonable prospect of the number or individuals that may be recovered through joint investigations and recoveries is about 3,000.

I. NonDNA Identification Evidence

For Korean cases, nonDNA evidence, such as location of recovery

and forensic anthropologic and dental examinations, can provide identifying information, but due to the paucity of antemortem medical and dental records, these will serve primarily to limit the servicemembers considered for a potential mtDNA match. CILHI scientists estimate that they will be able to narrow the possible name associations to an average of 25 from the overall pool of the approximately 8,100 cases based on age, race, and dental characteristics in the vast majority of cases and in some cases to within five individuals.

J. Current Methodology

The process of mtDNA testing ancient remains is technically difficult and demanding, but AFDIL now has considerable experience. The first steps involve the preparation of the sample. A portion of bone (about two grams per extraction) is cleaned to prevent cross contamination and to remove any mineralization that inhibits mtDNA testing. The sample is pulverized; from the bone powder, DNA is extracted. A quick analysis of the quality and purity of the DNA recovered is performed at that stage. The region of the mtDNA to be analyzed is then amplified using PCR. Multiple copies of specific segments of the mtDNA that are to be analyzed are generated until sufficient material for the sequencing reaction is produced. Once this process is completed automated sequencing instrumentation can then read the exact sequence of the mtDNA molecule. Finally, the mtDNA sequence from the family reference specimen is compared to that of the skeletal samples. The amplification and sequencing scheme employs two overlapping sets of primer sets which allow for further confirmation of sequencing results (Figure 7).

The Task Force finds that identification of so-called ancient skeletal remains by a program of mtDNA testing is possible, particularly in association with other information. A few specimens may remain unresolved. Although contenders may emerge; at this time, mtDNA sequencing technology is the most appropriate technology.

II. FACTORS

DSB TOR: To evaluate factors that might influence the effectiveness of using mtDNA techniques. [What factors might effect the utility of mtDNA testing for identification purposes, e.g. age of the skeletal remains, manner of interment, environmental temperature, acidity of soil, limitations of sample availability, commingling, etc? Do environmental factors affect the mtDNA sequence results obtained? How does somatic mutation and heteroplasmy affect mtDNA typing efforts? To what extent may the lack of family reference samples impede mtDNA typing efforts? What is the likely average number of bases obtained from Southeast Asian and Korean remains? if this data are not available, what size sampling would be sufficient for establishing reasonable estimates? How will these limitations of mtDNA sequence information available from Southeast Asian and Korean cases affect the ability to identify unassociated remains?]

Many factors may affect the ability of the military to perform accurate and successful mtDNA identifications on the skeletal remains.

A. Skeletal Remains

Recovered remains from Southeast Asia vary drastically in their quantity and quality. Some skeletal remains are virtually complete, where others have completely disintegrated, dissolving in the acidic soil. The integrity of the remains is affected by the manner of interment, length of interment, environmental temperature, and the acidity of the soil. In some cases, a hot environment can preserve by permitting thorough desiccation; in some cases, total submersion can permit preservation. Anthropologists note that even within the same burial site remains may demonstrate great differences in their degree of preservation. Thus, small changes in the environment can cause substantially different rates of degeneration.

Bone samples from Southeast Asia have demonstrated that they harbor only small amounts of mtDNA, and that it is severely fragmented. The variance in the quality of mtDNA within the bone samples is significant and unpredictable. The quality of the bone and the ease of obtaining genetic information from the bone

varies between bones from the same individual. Skeletal remains of the last Russian Tsar, Nicholas II, tested 73 years after his death, yielded relatively abundant quantities of large mtDNA fragments despite their age. The determination of whether or not a sample of bone will yield mtDNA information is simply a matter of trial. A second sampling of bone may yield mtDNA when the first sample did not. Teeth may yield mtDNA when bone will not.

There are indications that extraction of mtDNA from skeletal remains from North Korea may be significantly easier and more productive than from skeletal remains from Southeast Asia. The climate of Korea is far cooler and drier than Southeast Asia. Visual inspection of the quality of remains from Korea reveal that they are in reasonable condition relative to remains recovered from Southeast Asia. The AFDIL experience with Korean remains suggests they may not be any more problematic than those from Southeast Asia.

Recent repatriation of Korean remains disclose an 80% rate of admixture of different remains or "commingling". This is often easily recognized when, for instance, two right femurs are submitted; other times commingling is more difficult to diagnose. Joint recovery operations, which could include the use of more experienced U.S. recovery teams could substantially decrease the rate of commingling. Commingled remains may require multiple mtDNA samplings.

B. Family Reference Specimens

A lack of family reference samples may prohibit effective mtDNA identification. If necessary, family references could at least theoretically be obtained from exhumed remains of deceased family members. Occasionally, pre-mortem reference specimens such as locks of hair, paraffin-embedded material from prior biopsies, and neonatal bloodstains can be obtained to compare with the deceased individual. The ability of the military to find appropriate kindred is a factor in the success of mtDNA identifications. Unfortunately, the immediate kindred of servicemembers from the Vietnam and particularly the Korean conflicts are an aging population and hence the availability of family references is rapidly diminishing. There are indications

of interest by many families, suggesting they will cooperate by donating blood specimens (section V., Family Reference Database, p. 29).

C. Discriminatory Potential

The full discriminatory potential of the D-loop sequence is not yet known, since the region has been sequenced from relatively few people. Samples of sequences from large Caucasian populations have shown no sequence occurring with a frequency of greater than 3%. Sequences occurring with higher frequency would have almost certainly been detected. More frequent sequences have been found within some native African tribal populations. Coincidental matches between unrelated members are therefore likely to arise only a few percent of the time.

The amount of information obtained from family members will quickly allow greater precision to be attached to statements of discriminatory power, as will the sequences from remains. The latter sequences may, however, be of less than the complete hypervariable regions. Once 500 sequences are available from a population, a particular sequence can either be assigned its observed frequency or said (with over 99-. confidence) to have a frequency of less than 1% if it is not observed in the sample. However, sufficient data are available to conclude that a high discriminatory power is achievable by current methods and can be used for reasonable estimates of mtDNA sequence population frequencies.

Although, the DNA information from hypervariable regions I and II will permit great discrimination, the sequence information actually obtained from ancient samples may be limited, reducing the discriminatory power. The full mtDNA sequence may not be obtained due to amplification failure by one or more sets of primers, fading of terminal sequence signal, or from internal sequence ambiguities. The AFDIL is very conservative in determination of sequence or "base calling" and hence may not call an ambiguous, though informative base. Accordingly, the lack of full sequence information may result in limitation of the full potential discriminatory power in a given case.

Another potential limiting factor is the occurrence of multiple servicemembers of the same lineage. For example, brothers and maternal first cousins will have the identical mtDNA sequence.

The rate of maternal kindred relations among servicemembers is not known. The rate for servicemembers serving in Korea is likely somewhat higher than current frequencies, due to the larger families during that time period. Regardless, families themselves should typically have some idea of the existence of other kindred servicemembers who might potentially confound an identification.

D. Mutations

The mtDNA sequence is identical throughout the body unless a somatic mutation arises. Investigations by Monnat and Reay demonstrated that the mtDNA control region does not differ among the various tissues of the body. An investigation of 83 retinal cell clones, resulting in 32,000 bases of DNA sequence (from both the D-loop and coding regions), revealed only one mutation (in a tRNA gene).

More extensive age-related changes have however been found in other post-mitotic cells (Annex F). Bone and teeth have not yet been extensively studied. while somatic and germline mutations should be vigilantly looked for, they should not vitiate mtDNA identification in the majority of cases.

The polymorphic nature of mtDNA, especially within the control region, is evidence of mutation, and surveys support a widely held belief that mutations occur at a higher rate in the mitochondrion than in the nucleus. Whatever the mechanism for this elevated rate, it makes the mitochondrion useful for human identification. The disadvantage with high mutation rates is that sequences within the same maternal lineage, especially when separated by several generations, may exhibit occasional differences.

Although no comprehensive studies have been performed, evolutionary studies have estimated that the average fixed mutation rate for the mtDNA control region is one nucleotide difference per every 300 generations, or one difference every 6,000 years. Consequently, one would not expect to observe many examples of nucleotide differences between maternal relatives.

The AFDIL has observed at least two examples of fixed sequence differences between mother and child in approximately 30 maternal relative comparisons evaluated by AFDIL during routine casework.

In one case, there were two nucleotide differences between a mother and her daughter, but not in her two sons. The maternal relationship was verified using 12 DNA markers, including 8 RFLP loci.

The AFDIL surveyed 46 mother-child comparisons in a controlled study of family specimens from the Centre d'Etude du Polymorphisme Humaine (CEPH), showing no sequence differences. The AFDIL is completing an extensive study across three hundred generations which will help to further define the fixed mutation rate. Preliminary data suggest that the observed rate will be closer to one nucleotide difference every 50 generations. Dr. Mark Stoneking has found similar results in a recent study of the resident population of Tristan da Cunha, a small island off the Atlantic coast of South Africa.

This higher than expected rate of mutation may be explained by so-called "hot spots" which have a higher mutation rate than other regions even within HVl and HV2.

E. Heteroplasmy

The condition in which only a single mtDNA sequence (albeit in many copies) is present in a cell is termed homoplasmy. The possibility exists that a subpopulation of mitochondria could harbor a different mtDNA sequence due to mutation or paternal contribution; this is termed heteroplasmy. Heteroplasmy is most often seen in disease states, in which a mtDNA sequence is defective (e.g. deletions, duplications) and hence at a competitive disadvantage. Stable heteroplasmy in humans although rare has been described.

Measuring the fixed mutation rate does not take into account the rate of heteroplasmy during the process of fixing a particular mutation. It is not known whether most mutations are manifested through many generations via heteroplasmy or if they are fixed during a single generation. Examples of each have been observed (i.e., the Russian Tzar was reported to be heteroplasmic, despite the lack of heteroplasmy in other family members). The consensus of the Task Force Committee is the level of heteroplasmy within the control region of humans is generally low.

F. Environmental Damage

Environmental changes do not alter the mtDNA sequence. Long term exposure has resulted in depurination in which Guanines result in Adenines after PCR, but this has been observed only in material thousands of years old. This does not appear to be a problem with the identification of skeletal remains from the Vietnam and Korean conflicts. If environmental damage could be assessed in a rapid and cost effective manner, then that information could be used as indication of the quality of the data obtained or the need for repair of the DNA before amplification.

G. Contamination

The mtDNA present in ancient skeletal remains is minimal. The mtDNA that had existed has largely been enzymatically or chemically broken down into small fragments and much has been leached away. The AFDIL reports that only a few dozen target mtDNA fragments are obtained by their extraction techniques on skeletal remains from Vietnam. Therefore, the methods of mtDNA typing must be very sensitive. The enhanced sensitivity required for this type of testing (PCR-based) also means that it is capable of detecting trace contaminants.

MtDNA contaminants can come from a variety of sources. Contamination through proximity to other remains and through specimen handling in the field is not thought to be a significant source, especially when a laboratory removes the outer surfaces of bone samples prior to mtDNA extraction. Shed skin or exhaled droplets from individuals performing the extraction is another potential source of contaminating mtDNA. The mtDNA sequence of laboratory personnel must be determined so that their sequence types will be recognized as a possible contaminant. The most serious contamination concern is cross-contamination by PCR product within a laboratory. Millions of copies of the mtDNA target are generated during the PCR process. If a minute quantity of this product is allowed to contaminate another mtDNA extract, the contaminating product may mask the true mtDNA type of the extract. When all appropriate precautions are taken, random sporadic contaminants, not attributable to any known source, may be encountered.

Controlling contamination is vital to the success of a PCR-based mtDNA testing program, especially when targeting old skeletal remains. A range of precautionary measures tailored to the specific laboratory are necessary and when anticipating high