Fig. 2

Amount of variants for all 29 breast cancer cell lines based on RNA-seq data. a Filtering by read depth and quality (pass) reduced variant numbers markedly, whereas RNA-edit sites (edit) and low complexity regions (lcr) affected less variants. b Further filtering of dbSNP (dbsnp) data lowered numbers per sample substantially. Finally, focussing on variants in protein coding regions (snv_indel) and variants in less than 20% of samples resulted in about 1000 variants per sample (snv_indel_20). Mutations included single nucleotide variants (SNVs), insertion and deletions (InDel). c Correlation between mapped million reads and filtered variants decreased with every filtering step