Abstract
Genomic alterations constitute crucial elements of colorectal cancer (CRC). However, a comprehensive understanding of CRC genomic alterations from a global perspective is lacking. In this study, a total of 2,778 patients in 15 public datasets were enrolled. Tissues and clinical information of 30 patients were also collected. We successfully identified two distinct mutation signature clusters (MSC) featured by massive mutations and dominant somatic copy number alterations (SCNA), respectively. MSC-1 was associated with defective DNA mismatch repair, exhibiting more frequent mutations such as ATM, BRAF, and SMAD4. The mutational co-occurrences of BRAF-HMCN and DNAH17-MDN1 as well as the methylation silence event of MLH-1 were only found in MSC-1. MSC-2 was linked to the carcinogenic process of age and tobacco chewing habit, exhibiting dominant SCNA such as MYC (8q24.21) and PTEN (10q23.31) deletion as well as CCND3 (6p21.1) and ERBB2 (17q12) amplification. MSC-1 displayed higher immunogenicity and immune infiltration. MSC-2 had better prognosis and significant stromal activation. Based on the two subtypes, we identified and validated the expression relationship of FAM83A and IDO1 as a robust biomarker for prognosis and distant metastasis of CRC in 15 independent cohorts and qRT-PCR data from 30 samples. These results advance precise treatment and clinical management in CRC.