How to assess the number of contributors to a DNA mixture: a Bayesian approach

Forensic scientists are routinely faced with the problem of assessing the total number of contributors while examining DNA mixtures. Inferential procedures can be biased, in particular when assessments of the number of contributors are not made prior to considering the genotypic configuration of potential donors. Moreover, the number of contributors may be incompatible with the set of alleles observed in the profile of a mixed stain, given the genotype of a potential contributor. In such situations, procedures leading to a fixed number contributors as output can not be appropriate. Different strategies to infer the number of contributors in a mixed stain are analyzed by considering mixtures in a qualitative way. A particular emphasis is devoted to the Bayesian approach providing a posterior probability distribution that encapsulates the uncertainty about the number of contributors to a mixed stain. Simulation studies and graphical models will be implemented and the performances of alternative procedures will be analyzed and discussed.