Revisiting variation in the somatic mutation landscape of non-small cell lung cancer.

Abstract

Non-small cell lung cancer (NSCLC) is driven by a diverse array of somatic mutations. The vast majority of literature on NSCLC is based on targeted assays or small sample sizes, limiting the ability to provide a comprehensive view of NSCLC mutation profiles. Here, we analyzed genome-wide screen data (including whole genome sequencing and whole exome sequencing) from 1,874 NSCLC subjects to identify molecular subtypes and putative driver genes and to explore the effect of intrinsic and extrinsic factors on somatic mutation profiles. We showed that genome-wide screen data support existing knowledge, such as the TP53:KRAS mutation co-occurrence pattern as a key distinctive feature, but do not reveal additional broad molecular subtypes. In contrast, we demonstrated that low-frequency molecular subtypes or potential driver genes continue to be identified. Using driver gene identification algorithms, we found 50 potential driver genes including ANG, CDK10, CTDSP2, HOXA5, RBP4, and SPHK2, which show evidence of positive selection in NSCLC. Finally, we provided insights into the intrinsic and extrinsic covariates associated with the NSCLC somatic mutation landscape, while confirming associations with ethnicity (TP53 and EGFR), NSCLC subtype (14 genes including KRAS, NFE2L2, and STK11), and smoking history (KRAS, CSMD3, and TP53), we dismissed gene-level associations with sex when other covariates are controlled for. The results presented here represent a concise up-to-date summary of variation in the somatic mutation landscape and carry importance for NSCLC geneticists, medical practitioners, and drug discovery scientists.