Massively parallel sequencing of extended SNP panels with applications in relationship inference

Short tandem repeat (STR) markers are the current golden standard in forensic genetics, whereas single nucleotide polymorphisms (SNPs) have emerged as an alternative. In this thesis it was investigated if using SNP panels as a supplement to STR panels could lead to more conclusive results in complex kinship cases. 16 samples from eight complex kinship cases with inconclusive results were analysed with two STR panels and two supplementary SNP panels. The results were evaluated by comparing the likelihood ratio (LR) calculated based on the two STR panels, with the LR calculated based on the combination of both STR- and SNP panels. The STR analyses were performed with traditional capillary electrophoresis, while the SNPs were analysed with massively parallel sequencing using the Ion TorrentTM Personal Genome MachineTM (PGMTM) System (Thermo Fisher Scientific). Sequencing on the PGMTM is not a part of the accredited routine at REFA, and this method was thoroughly evaluated and reviewed in this project. Several software and tools were used in the evaluations in the project. For instance, the statistical software Familias was used to calculate LR for the real cases and the simulations, and to perform a blind search. Results from both the real cases and the simulations showed a notable decrease in the number of inconclusive cases when SNPs were included in the analyses. However, throughout this project it has been revealed that there are several important aspects that can affect the final conclusion in complex kinship cases, especially when a large number of markers are applied. These have been reviewed in terms of further work with constructing a SNP panel that can be used in routine work, and it was suggested that the markers should be ancestry-insensitive, not in linkage disequilibrium with each other, and that linkage should be calculated and included in the analyses. It was also suggested that larger- or better suited panels were needed to solve cases where the proposed relationship was half sibling of parent or equivalent. Additionally, it was shown in this project that the use of the correct allele frequency databases in the calculations was crucial, especially if ancestry- sensitive markers were applied. Finally, it was concluded that SNPs are well suited as supplement to STRs in complex kinship cases, but that further investigations should be performed in respect to construct a panel and a procedure that is suitable for routine work