Evaluating alignment and variant-calling software for mutation identification in <i>C</i>. <i>elegans</i> by whole-genome sequencing

<div><p>Whole-genome sequencing is a powerful tool for analyzing genetic variation on a global scale. One particularly useful application is the identification of mutations obtained by classical phenotypic screens in model species. Sequence data from the mutant strain is aligned to the reference genome, and then variants are called to generate a list of candidate alleles. A number of software pipelines for mutation identification have been targeted to <i>C</i>. <i>elegans</i>, with particular emphasis on ease of use, incorporation of mapping strain data, subtraction of background variants, and similar criteria. Although success is predicated upon the sensitive and accurate detection of candidate alleles, relatively little effort has been invested in evaluating the underlying software components that are required for mutation identification. Therefore, we have benchmarked a number of commonly used tools for sequence alignment and variant calling, in all pair-wise combinations, against both simulated and actual datasets. We compared the accuracy of those pipelines for mutation identification in <i>C</i>. <i>elegans</i>, and found that the combination of BBMap for alignment plus FreeBayes for variant calling offers the most robust performance.</p>