Facilitating complex DNA mixture interpretation by sequencing highly polymorphic haplotypes

Abstract

Interpretation of complex DNA mixtures is an ongoing challenge in the field of forensic genetics. Commonly used STR markers are quite polymorphic, enabling very high statistical association between a single source DNA profile from a crime scene and a matching suspect. STR typing of low order mixtures with two and three contributors also commonly produces high statistical association for a contributor, using current interpretation software. However higher order mixtures, with four contributors or more, are more challenging. Shared alleles among the many contributors may complicate the correct assessment of the number of contributors to a mixture and decreases the statistical support for inclusion of contributors. Recently, there is a rising use of massively parallel sequencing for forensic applications, and markers such as SNPs and short haplotypes are receiving more attention. However, these markers are even less polymorphic than autosomal STRs and are not suitable for complex mixture interpretation. We propose to use a different panel of haplotype markers, which contain many SNPs and are very polymorphic. These markers can be sequenced either by standard sequencing technologies or by a new instrument called MinION that sequences DNA as it passes through a nanopore. This instrument is very small and can sequence long stretches of DNA, but suffers from high error rate. We present a method for calling haplotype alleles facing high error rate, and use simulation to test its robustness. We also calculate likelihood ratio (LR) between propositions of contribution and non-contribution for individuals that do, and do not, contribute to various complex DNA mixtures. Our results indicate that the correct alleles can be identified in a mixture, despite the high sequencing error rate, and that contributors get high LR (>10⁹) even in complex mixtures with up to five contributors. Non-contributors receive a very small LR, below 1 in most cases (>98%), which support their exclusion as possible donors to the complex DNA mixtures.

Keywords: Complex mixtures; Forensic DNA analysis; Haplotypes; MinION; Nanopore.