Abstract
Deciphering the post-transcriptional mechanisms (PTM) regulating gene expression is critical to understand the dynamics underlying transcriptomic regulation in cancer. Alternative polyadenylation (APA)-regulation of mRNA 3'UTR length by alternating poly(A) site usage-is a key PTM mechanism whose comprehensive analysis in cancer remains an important open challenge. Here we use a method and analysis pipeline that sequences 3'end-enriched RNA directly to overcome the saturation limitation of traditional 5'-3' based sequencing. We comprehensively map the APA landscape in lung cancer in a cohort of 98 tumor/non-involved tissues derived from European American and African American patients. We identify a global shortening of 3'UTR transcripts in lung cancer, with notable functional implications on the expression of both coding and noncoding genes. We find that APA of non-coding RNA transcripts (long non-coding RNAs and microRNAs) is a recurrent event in lung cancer and discover that the selection of alternative polyA sites is a form of non-coding RNA expression control. Our results indicate that mRNA transcripts from EAs are two times more likely than AAs to undergo APA in lung cancer. Taken together, our findings comprehensively map and identify the important functional role of alternative polyadenylation in determining transcriptomic heterogeneity in lung cancer.