A distinct class of pan cancer susceptibility genes revealed by alternative polyadenylation transcriptome wide association study

Alternative polyadenylation (APA) plays an important role in cancer initiation and progression; however, current genome- and transcriptome-wide association studies (GWAS and TWAS, respectively) mostly ignore APA when identifying putative cancer susceptibility genes. Here, we performed a pan-cancer 3'untranslated region (UTR) APA TWAS (3'aTWAS) by integrating 80 well-powered (n>50,000) GWAS datasets across 23 major cancer types with APA quantification from 17,330 RNA sequencing samples across 49 tissue types and 949 individuals. We found that genetic variants associated with APA represent around 24.4% of cancer GWAS variants and are more likely to be causal variants explaining a large portion of cancer heritability. We further identified 413 significant APA-linked cancer susceptibility genes. Of these, 77.4% have been overlooked by traditional expression- and splicing-studies, given that APA may regulate translation, protein localization, and protein-protein interactions independent of the expression level of the genes or splicing isoforms. As proof of principle validation, modulation of four novel APA-linked breast-cancer susceptibility genes significantly altered cancer cell proliferation. Our study highlights the significant role of APA in discovering new cancer susceptibility genes and provides a strong foundational framework for enhancing our understanding of the etiology underlying human cancers.