DNA Identification in Mass Fatality Incidents
Kenneth Kidd
Because the World Trade Center identification effort presented challenges not found in ordinary forensic work, it was not self-evident how to best evaluate the DNA data statistically. The KADAP discussions and recommendations were, I think, essential to reaching the final set of criteria.
Kinship analysis is necessary when known reference samples are not available for the reported missing. In kinship analysis—well-established in the field of human genetics—close relatives provide DNA samples. Assignment of remains to a particular family is made by assessing the relative likelihood ratios.
“Indirect” DNA-based identification, or kinship analysis, involves two steps. The first is a screening process, in which DNA profiles from the remains are compared with family reference samples. The process allows families to be eliminated and “candidate families” to be isolated. The second step determines the likelihood that a particular remains sample fits into a particular candidate family. [Note: When only more distant relatives are available to provide reference samples, kinship analysis cannot provide robust estimates of relationships.]
Kinship analysis is not always successful for making unambiguous identifications of victim remains. Its limitations include difficulty in identifying remains belonging to same-sex full-siblings, particularly monozygotic twins. Also, close relatives are sometimes victims of the same disaster, and differentiating such relationships in the absence of very complete pedigrees may not be possible. It is helpful to be aware of the number of related victims prior to assigning identity, but this may not be possible in an open-system mass fatality event. In fact, this may be difficult even in a well-documented closed disaster, in which a high level of statistical confidence may not be reached without a sufficient number of close relatives to provide reference samples.
Generally, the more close relatives available as references, the better the chances for attaining the identification threshold. For example, when standard paternity trios (mother, father, child) are available and numerous polymorphic loci are typed, sufficiently high likelihood ratios typically result. However, if only one sibling is available to provide a family reference sample, it is unlikely that a statistical threshold for a DNA-based identification would be met.
Because a personal reference item does not always contain sole-source DNA (or contain the DNA of the person to whom it is attributed), it is advisable to confirm direct-comparison DNA by indirect kinship analysis. Although it may seem like additional work to first examine the profile from the personal effect against the profiles from relatives to establish the fit in the pedigree, verifying the source of the DNA from the personal effect actually reduces the time needed to finalize DNA-based victim identification. When verifying the origin of the personal effect within the kindred of reference samples, the kinship analysis by itself does not need to meet the threshold requirement for identification, as it serves only to confirm the direct comparison to personal effects.
When multiple family members—or, particularly, an entire family—are victims of an incident, identifying at least one of them with high confidence improves the ability to achieve certainty for the others; the caveat, here, is that same-gender siblings may not be distinguishable by kinship analysis unless direct reference samples are used. Although fragmented remains of same-sex siblings usually can be distinguished from each with DNA analysis, families would not be able to be told which remains came from which sibling unless direct reference samples (from one or the other) are used to differentiate the remains. Without direct references from a known personal effect, same-gender siblings usually cannot be distinguished when remains are fragmented or otherwise lacking in traditional anthropological and other phenotypic characteristics.
