Evaluating the Impact of Dropout and Genotyping Error on SNP-Based Kinship Analysis With Forensic Samples.

Technological advances in sequencing and single nucleotide polymorphism (SNP) genotyping microarray technology have facilitated advances in forensic analysis beyond short tandem repeat (STR) profiling, enabling the identification of unknown DNA samples and distant relationships. Forensic genetic genealogy (FGG) has facilitated the identification of distant relatives of both unidentified remains and unknown donors of crime scene DNA, invigorating the use of biological samples to resolve open cases. Forensic samples are often degraded or contain only trace amounts of DNA. In this study, the accuracy of genome-wide relatedness methods and identity by descent (IBD) segment approaches was evaluated in the presence of challenges commonly encountered with forensic data: missing data and genotyping error. Pedigree whole-genome simulations were used to estimate the genotypes of thousands of individuals with known relationships using multiple populations with different biogeographic ancestral origins. Simulations were also performed with varying error rates and types. Using these data, the performance of different methods for quantifying relatedness was benchmarked across these scenarios. When the genotyping error was low (<1%), IBD segment methods outperformed genome-wide relatedness methods for close relationships and are more accurate at distant relationship inference. However, with an increasing genotyping error (1-5%), methods that do not rely on IBD segment detection are more robust and outperform IBD segment methods. The reduced call rate had little impact on either class of methods. These results have implications for the use of dense SNP data in forensic genomics for distant kinship analysis and FGG, especially when the sample quality is low.

skater: an R package for SNP-based kinship analysis, testing, and evaluation.

Motivation: SNP-based kinship analysis with genome-wide relationship estimation and IBD segment analysis methods produces results that often require further downstream process- ing and manipulation. A dedicated software package that consistently and intuitively imple- ments this analysis functionality is needed. Results: Here we present the skater R package for SNP-based kinship analysis, testing, and evaluation with R. The skater package contains a suite of well-documented tools for importing, parsing, and analyzing pedigree data, performing relationship degree inference, benchmarking relationship degree classification, and summarizing IBD segment data. Availability: The skater package is implemented as an R package and is released under the MIT license at https://github.com/signaturescience/skater. Documentation is available at https://signaturescience.github.io/skater.

vcferr: Development, validation, and application of a single nucleotide polymorphism genotyping error simulation framework.

Motivation: Genotyping error can impact downstream single nucleotide polymorphism (SNP)-based analyses. Simulating various modes and levels of error can help investigators better understand potential biases caused by miscalled genotypes. Methods: We have developed and validated vcferr, a tool to probabilistically simulate genotyping error and missingness in variant call format (VCF) files. We demonstrate how vcferr could be used to address a research question by introducing varying levels of error of different type into a sample in a simulated pedigree, and assessed how kinship analysis degrades as a function of the kind and type of error. Software availability: vcferr is available for installation via PyPi (https://pypi.org/project/vcferr/) or conda (https://anaconda.org/bioconda/vcferr). The software is released under the MIT license with source code available on GitHub (https://github.com/signaturescience/vcferr).