Figure 4

Motivation for the pseudo- and the RNA-seq validation methods. For two READ patients (columns), the upper row shows the signed GATK variant quality scores based on the tumor-normal Illumina exome-seq pairs, and the lower row shows the histogram of the variant allele fraction in the tumor Illumina RNA-seq. Note that for the READ samples, mutation calling was done by three callers (Caller H, I, and J) using SOLiD exome-seq pairs. For the upper panels, we first compiled all the variant sites reported in any VCF file based on SOLiD exome-seq pairs. Then, for each such site, using the Illumina exome-seq data, we obtained the GATK variant call quality score for each tumor (y-axis) and the normal sample (x-axis). When no variant allele was detected by the GATK UnifiedGenotyper, we flipped the sign. Mutations that were validated by the 454 sequencing technology are colored: red (somatic), blue (germline) and green (wildtype). Among the validated mutations, those with the RNA-seq depth ≥ 5x were further examined for the variant allele fraction in the tumor RNA-seq data (lower panels).