Familial database search on two-person mixture

This work presents how the familial database search can be applied to forensic investigations in which the DNA evidence is a mixture. The forensic identification is based on a ranked list of the likelihood ratios for each candidate in the DNA database of being the relative of the contributor of the forensic sample. A simulation study using Hong Kong Chinese data demonstrated the comparable effectiveness of our approach that in 90% of the cases, the sibling of the contributor can be found in the top 50 profiles, and the parent/child can be found in the top 10 profiles. We also discuss the prediction on the performance of the search that can guide the decision of the police force to initiate the investigation.