Evaluation of gain in statistical power for kinship analysis using sequence-based versus length-based STR genotyping.

Abstract

In this study, 137 pairwise relationships representing four major relationship categories involving 49 Turkish individuals from four families were analyzed to evaluate the potential gain in the statistical power associated with likelihood ratios (LR) when using sequence-based versus length-based genotyping methods over the same STR loci coverage. To this end, the MPS Precision ID GlobalFiler NGS STR panel Kit and CE GlobalFiler™ PCR Amplification Kit were used. MPS-based analysis revealed the presence of 37 STR DNA sequence variations and / or the presence of 26 STR DNA sequence flanking region SNPs compared to the 150 unique alleles obtained with CE-based genotyping. Considering that most kinship LR calculation software do not readily take into consideration STR DNA sequence variants and STR DNA sequence flanking region SNP data that becomes available during MPS-based genotyping, an alphanumeric allele re-coding system was implemented to incorporate such additional STR isoallelic data to the already available allele calls. Over all the four major relationship categories analyzed, a significant increase in the mean combined LR (cLR) was observed when going from CE-based to MPS-based typing, whereby a 78.08 to 7,864,630.60-fold increase was noted. More specifically, in 134 out of the 137 pairwise relationships analyzed, MPS-based cLR values were higher than those calculated using CE-based data. While the mean cLR was >1,000 for three out of the four major relationship categories when using CE, the only exception being the third degree relationships, the mean cLR was >1,000 for all the four major relationship categories when using MPS. Notably, the mean cLR obtained for the third degree relationships was 47.61 with CE and 3,717.31 with MPS. In comparison with CE-based genotyping, when fully taken into account as proposed in the current study, the DNA sequence variation data afforded by MPS-based genotyping led to a statistically significant gain in terms of cLR values obtained. The use of MPS for cLR calculations had the most impact for both the second and third degree relationships, the two complex / distant type analyzed, hence further underscoring the prospects for MPS in kinship analysis. While the current study demonstrated that cLR is likely to increase substantially upon going from CE to MPS genotyping over the same loci coverage for a given case, when the additional DNA sequence variances are also taken into consideration, further increases are expected due to the more diverse type of forensic markers and even wider loci coverages used by MPS kits.