The assignment of the weight of DNA evidence depends on a number of factors (allele probability estimates, the population genetic model used, the value of the coancestry coefficient, etc.). One of these factors is the allele probability estimates from a database. The appropriateness of any particular database to any given case is a matter of judgement. This judgement is often based on relevant background information such as the location of the crime. In 1996 the NRC Committee on DNA report1 stated that,
empirical studies show that the differences between the frequencies of the individual profiles estimated by the product rule from adequate subpopulation databases (
) are within a factor of about 10 of each other.... This statement has proven valuable as a gauge on variability caused by the database. However it was developed at a time before STR multiplexes and is overdue for an update. The present study examines the validity of the factor of 10 method for the interpretation of DNA typing results today. It compares the match probabilities obtained using hundreds of different allele frequency databases from around the world. More specifically, the authors first simulated sets of genotypes from a database of interest based on the allele frequencies of that database. Second, match probabilities were obtained using the allele frequency data from the original allele frequency database used for simulating the set of genotypes, and match probabilities were obtained using allele frequency data from each of the other databases in the study corresponding to the same ethnic group. The match probabilities obtained using the data from the other allele frequency databases were then compared with the match probabilities obtained using the original allele frequency database used for simulating the set of genotypes. The results of these comparisons indicate a variation among the match probabilities that is greater than a factor of 10.
DNA interpretation, allele frequencies, subpopulations.