Fig. 1

cisMuton calls. a False-positive SNV calls in negative control data where tumor FFPE (for foreground data) and tumor frozen samples (for background data) were taken from the same tissue block. The numbers were normalized by target region size (477 k bp). b Sensitivity estimation using semi-simulated data. We mixed reads from a cell line and reads from an unmatched normal sample to mimic decreasing tumor purity. Variants genotyped by SNP arrays in the pure cell line sample were used as answers. c Isolated calls, i.e., variants called by each given tool, and those not called by the given tool but by all the others, in 70 FFPE samples. Target regions were the same between all the tools and the numbers were normalized by the target region size. d SC-FPs and SC-FNs evaluated by mass spectrometry for variants from the datasets of panel c. The sample size (n) is indicated below the x-axis. Variants with ≥ 5% VAFs were selected. e Integrative Genomics Viewer (IGV) [33] screenshot of an SNV that was called by both cisCall and mass spectrometry but missed by Mutect