Searching a DNA databank with complex mixtures from two individuals

Abstract In 2000, Canada established a CODIS-based National DNA Databank (NDDB) composed of two indices: a Convicted Offender Index (COI) and a Crime Scene Index (CSI). Our laboratory populates the CSI index with single-source and mixed profiles on the 13 P+/COfiler ® loci. These mixtures are coded based on the number of mixed loci (loci with more than 2 alleles). We first compared the behaviour of several hundreds low complexity mixtures (≤3 mixed loci), moderate complexity mixtures (4–6 mixed loci) and single-source profiles uploaded over the past 9 years, when run against 180000 COI profiles. We demonstrated the value of searching a databank with mixtures as hundreds of matches were reported to investigators, while the proportion of rejected matches was similar across the 3 types of samples. Based on these results, we designed a new study to evaluate 130 complex (7–13 mixed loci), two-contributor mixtures which were run against the COI. Results show that such mixtures return a manageable number of candidate matches in spite of their complexity: 70% returned no candidates while only a few generated more than one candidate. Databanking strategies coupled with mixture interpretation guidelines and review of original electropherograms maximise the use of a databank while minimising the risk of adventitious hits.