Fig. 6

The performance measurements of variant-calling tools in different datasets. a, b Performance curves showing the median TPRs (a) and FDRs (b) according to the minimal read depths at SNP loci in high-confident coding regions for the SMART-seq2 hepatocellular carcinoma dataset. c Boxplots showing the log-transformed counts of detected SNVs using different sequencing platforms. d, e Performance curves showing the median TPRs (d) and FDRs (e) according to the minimal read depths at SNP loci in high-confident coding regions for the 10X hepatocellular carcinoma dataset