× Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø

Simple consequences of Mendelian genetics imply that relatives share more alleles in their DNA profiles as compared to unrelated individuals, and further, individuals of the same population are expected to share alleles more frequently than pairs of individuals chosen from different populations. While this had been the logic of detecting and eliminating the inadvertent presence of duplicate DNA profiles in large forensic databases, several court proceedings have seen allegations suggesting that forensic laboratories have mistakenly purged databases, forcing elimination of multiple locus matches in their databases. In this presentation, we show that such allegations are scientifically flawed, and the detection and elimination of profiles that show multiple locus matches are legitimate and valid. Through the use of the ITO method of allele sharing statistics, we present an algorithm of computing the distribution of allele sharing between individuals of any arbitrary relationships, which in turn results in a likelihood ratio criterion for testing any hypothesized relationship between a pair of individuals. The application of this method to several DNA profiling databases shows that it is perfectly legitimate and scientifically accurate to delete profiles that show multiple locus matches from databases, since they can not occur by any reason other than being inadvertent duplicate records of the same individual. Furthermore, we demonstrate that should such duplicate records remain undetected, the resultant allele frequently estimates from the databases can only over-estimate the estimated profile frequencies. The presence of duplicate samples in a database, however, may erroneously suggest multilocus dependence of alleles among loci that are inherently independent. These results also argue that the direct counting method for computing multilocus genotype frequencies from databases is scientifically inefficient, if not totally invalid. (Research supported by US Public Health Service research grants from the National Institutes of Health and the National Institute of Justice.)