Analysis of alternative cleavage and polyadenylation by 3' region extraction and deep sequencing.

Abstract

Alternative cleavage and polyadenylation (APA) generates diverse mRNA isoforms. We developed 3' region extraction and deep sequencing (3'READS) to address mispriming issues that commonly plague poly(A) site (pA) identification, and we used the method to comprehensively map pAs in the mouse genome. Thorough annotation of gene 3' ends revealed over 5,000 previously overlooked pAs (∼8% of total) flanked by A-rich sequences, underscoring the necessity of using an accurate tool for pA mapping. About 79% of mRNA genes and 66% of long noncoding RNA genes undergo APA, but these two gene types have distinct usage patterns for pAs in introns and upstream exons. Quantitative analysis of APA isoforms by 3'READS indicated that promoter-distal pAs, regardless of intron or exon locations, become more abundant during embryonic development and cell differentiation and that upregulated isoforms have stronger pAs, suggesting global modulation of the 3' end-processing activity in development and differentiation.