The effect of population subdivision on estimates of the likelihood ratio in criminal cases using single-locus DNA probes

Abstract

The interpretation of single-locus DNA profile evidence in criminal cases in normally performed using an assumption of independence of frequencies of alleles at different loci. The critical parameter is the probability of a suspect matching a criminal's DNA profile by change. Many authors have been concerned that the assumption of linkage equilibrium may be inaccurate and that the strength of evidence against suspects may be thereby exaggerated. Some have calculated the maximum size of this effect. What matters, however, is not the upper bound of what the true probability might be, but what the actual probability is in a subdivided population. By calculating this, one can show that the effect of population subdivision on the strength of DNA profile evidence is very minor.