Some mathematical problems in the DNA identification of victims in the 2004 tsunami and similar mass fatalities

Abstract

DNA is a major and essential identification tool for mass fatality incidents including the hundreds of thousands of victims of the 2004 Indian Ocean tsunami. Mathematical complications characteristic of this sort of mass fatality include prevalence of related victims, the many races represented among the victims, and various identification modalities in tandem with DNA. Four mathematical problems of interest are discussed in this paper. (1) Other quantifiable factors (i.e. geography) can be formally accounted for by including a likelihood ratio that can be thought of as reducing the “effective number of victims.” (2) When a victim is found and tentatively identified as V, but then it comes to light that the victim has a relative W who is also missing, confidence in the identity is depressed. To account for the existence of W, increment the effective number of victims by the likelihood ratio supporting W as the identity of the victim. (3) When several apparently related victims are found, their mutual identities should be calculated simultaneously. Compared to one-at-a-time, serial identifications, this is both logical and may lead to much more confidence in the identities. (4) Although there may be many different population groups represented among the missing, it is generally sufficient to consider population statistics for only a few of them in deciding whether to declare an identification.