유전자검사를 이용한 대량 희생자 개인식별에서의 통계적 해석

Abstract

DNA profiles have been increasingly used as the most reliable means to identify remains from war or mass disaster. To establish the identity with such a large set of victims, special care should be taken to correlate remains with correct family references while avoiding coincidental match between non-relatives. Therefore we address here relevant statistical and combinatorial issues in the DNA identification of mass victims. A simple and general formula for the likelihood ratio governing any potential kinship between two DNA profiles was presented, and for that purpose, the probabilities that a given relative and an individual share autosomal identical-bydescent alleles were calculated. In addition, a method dealing with the allele drop-out in kinship analysis and the estimation of a cold hit were discussed.