Figure 3

The effect of increased sequencing of the normal genome on Seurat’s somatic mutation detection. Legend: Demonstration of the effect of increased sequencing of the normal genome in a matched normal/tumor analysis using Seurat. We present three common scenarios: A ) a locus with a true somatic variant, but presented with low frequency of the variant allele, because of mapping difficulty, low purity of the tumor biopsy or because of the variant being present only in a minor sub-clonal population. B ) a locus with a potential false-positive call, because of erroneously-aligned variant evidence. C ) a locus with a variant genotype in the normal genome, but with a coincidental lack of evidence causing it to appear as a tumor-only variant. In all three scenarios, the increase in sequencing data available for the normal genome updates the expectation of variant evidence (by altering the shape of the conjugate beta distribution) and consequently amplifies Seurat’s capability to correctly reject the last two cases and accept the first case.