Figure 1.
Coverage and mutation burden in samples from selected genomics studies. (A) The assay type, experimental design, and DNA source is shown for each study selected for comparison. WGS studies in the darker shade of pink also performed sequencing on matched normal samples and the remaining studies shown all (or primarily) sequenced only tumors. (B) Box-whisker plot showing the effective nonredundant coverage across the target space or genome (MeanCorrectedCoverage), respectively for the samples from WES (top) or WGS (bottom) studies. Samples from fresh frozen or FFPE tissue are shown separately where that information was available. Individual points showing the coverage of each sample are overlaid. The average depth reported in each study or, when not reported, the targeted coverage is indicated. (C-D) All samples from 8 studies were subjected to the same in-house variant calling pipeline to determine the number of coding variants we could detect. These values are plotted as a function of coverage and shown separately for WES and WGS studies. Insets on the right-hand side show the number of cases in each study with MeanCorrectedCoverage of at least 25×. ∗Data from the ICGC MALY-DE project were used in a series of studies. We use ICGC to refer to all the cases available through EGA. ∗∗Panea et al12 described WGS, but the data deposited in EGA contained both WGS and WES data combined. The original bam files containing both data types were used here, and thus we present them as WES data for comparison. FFPE, formalin-fixed paraffin-embedded.