Abstract
Genome sequencing of mutants is one of the most widely used techniques to identify genes that control traits of interest including human diseases. Traditionally, variants are called against reference genomes and various filtering techniques are applied to reduce the number of candidate mutations. However, if the genetic background of the mutant is different from the reference genome, the number of background variants may exceed the number of true mutations by several orders of magnitude resulting in candidate lists that cannot be effectively reduced.
We introduce the problem of differential variant calling in mutants from diverse genetic backgrounds. In the example of a mutant strain of the nematode Pristionchus pacificus, where the genetic background is not identical to the reference genome (≈1% genome-wide divergence), we show that simple intersection filtering does not effectively reduce the list of candidate mutations due to the combined effect of high number of background mutations, missing coverage in the wildtype sample, and problematic regions in the genome assembly. Although restriction to sites with coverage in mutant and wildtype sample greatly reduced the number of candidate sites, we further improved this candidate set by implementing a customized variant calling procedure. This takes the mutant sample and the control sample of same genetic background as input and calls variants exhibiting strong discriminative signals across the two samples. Intersecting the candidate mutations with an interval identfied from mapping by RAD-seq revealed a likely splice-site mutation in the P. pacificus dpy-1 gene, which has been previously shown to cause the associated morphological phenotype.
Our study shows that combined analysis of mutant and wildtype samples drastically increases the potential to find true mutations. We hope that these results may be helpful for other model systems, where the identification of candidate mutants is complicated by assembly errors, lab-derived mutations, or different genetic backgrounds.
