Autologous reference types can confound the detection of somatic mutation in solid cancers

Vast number of somatic mutations has been proved to be affected by the factors of sequencing methods, analysis pipelines and validation methods. We here showed the effect of autologous reference types on the detection of cancer-associated somatic mutations with the somatic single nucleotide variations (SNVs) and clinical data of solid tumors from the Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC). The distribution of somatic SNVs was significantly different among groups of autologous references in 6 cancers detected by whole genome sequencing (WGS) and 5 cancers detected by the random sequencing of exonic regions selected from the genome (WXS), especially in protein coding region of 5 cancers with age, gender and TNM adjusted. In addition, only 60.24% (95% CI: 49.65%-70.83%) of the somatic SNVs called from normal blood by WXS were found in those called from normal solid tissue tested by WXS / WGS, while 31.78% (95%CI: 4.14%-59.42%) of the somatic SNVs called from normal tissue adjacent to primary by WXS were found in those from normal blood tested by WXS / WGS. These findings suggested that more representative types of normal tissues should be included in detection of cancer-associated somatic mutations.