A PAN-CANCER ANALYSIS OF ALTERNATIVE PROMOTERS USING RNA-SEQ DATA

Abstract

Most human protein-coding genes transcribe multiple isoforms via alternative splicing and are regulated by multiple, distinct promoters. The wide availability of RNA-seq data allows the examination of alternative splicing, yet study of alternative promoters mostly relies on experimental approaches such ChIP-Seq and CAGE with limited availability. Therefore, the large-scale comprehensive analysis of promoter landscape is still not feasible experimentally and computationally. Here, we developed a computational method called proActiv to infer promoter activity by using widely available RNA-Seq data. Furthermore, by utilizing our promoter activity estimates, we developed a statistical framework to identify alternative promoters displaying context-dependent regulation. We find that promoters are deregulated across tissues, cancer types, and patients; and showed that promoter activity estimates could help us interpret noncoding promoter mutations. Our study suggests that promoter activity can be robustly and efficiently estimated using RNA-Seq data and thew dynamic landscape of promoters shapes the tissue and cancer transcriptome.