Abstract
BACKGROUND: Colorectal cancer (CRC) is among the most common cancers diagnosed worldwide. Although genome-wide association studies have effectively identified the genetic basis of CRC, there is still unexplained variability in genetic risk. Transcriptome-wide association studies (TWAS) integrate summary statistics from CRC genome-wide association studies (GWAS) with gene expression data to prioritize these GWAS findings and uncover additional gene-trait correlations. METHODS: First, we carried out a post-GWAS analysis using summary statistics from a large-scale GWAS of CRC (n = 4,562 cases, n = 382,756 controls). Second, combined with the expression weight sets from GTEx (v7), susceptibility genes were identified with the FUSION software. Colocalization, conditional and fine-mapping analyses, phenome-wide association study (pheWAS), and Mendelian randomization were employed to further characterize the observed correlations. RESULTS: In the post-GWAS analyses, we first identified new genome-wide significant associations: three genomic risk loci were identified at 8q24.21 (rs6983267, P = 6.98 × 10(-12)), 15q13.3 (rs58658771, P = 1.40 × 10(-10)), and 18q21.1 (rs6507874, P = 1.91 × 10(-14)). In addition, the TWAS also identified four loci statistically significantly associated with CRC risk, largely explained by expression regulation, including six candidate genes (DUSP10, POU5F1B, C11orf53, COLCA1, COLCA2, and GREM1-AS1). We further discovered evidence that low expression of COLCA2 is correlated with CRC risk with Mendelian randomization. CONCLUSIONS: We discovered novel CRC risk loci and candidate functional genes by merging gene expression and GWAS summary data, offering new insight into the molecular processes underlying CRC development. This makes it easier to prioritize potential genes for follow-up functional research in CRC.