COMPREHENSIVE CHARACTERIZATION OF ALTERNATIVE POLYADENYLATION IN HUMAN CANCERS AND APPLICATIONS IN NEOANTIGEN PREDICTION

Abstract

Alternative polyadenylation (APA) plays an important role in post-transcriptional regulation of eukaryotic gene expression. To understand APA regulatory mechanism across cancers, I developed two bioinformatics methods for profiling APA landscape and discovering APA regulators. By analyzing 10,631 cancer samples, 1,372 genes with recurrent tumor-specific APA events were identified and potential APA regulators were uncovered. Applying my pipeline to 12 major cancers collected in TCGA, I further focused on identification of intronic APA (IPA), which could lead to truncated protein isoforms and may serve as a source of neoantigens. Widespread tumor-specific IPA transcripts and neoantigens were observed across cancers. The predicted IPA-derived neoantigens were evidenced to be truly expressed in cancer cells by proteomics data. Importantly, the frequency of IPA-derived neoantigen was comparable to previously focused neoangtien sources, i.e. somatic mutation and intron retention. Together, my study suggests IPA could serve as a novel source for the development of immunotherapeutic vaccines.