APASdb: a database describing alternative poly(A) sites and selection of heterogeneous cleavage sites downstream of poly(A) signals

Abstract

Increasing amounts of genes have been shown to utilize alternative polyadenylation (APA) 3′-processing sites depending on the cell and tissue type and/or physiological and pathological conditions at the time of processing, and the construction of genome-wide database regarding APA is urgently needed for better understanding poly(A) site selection and APA-directed gene expression regulation for a given biology. Here we present a web-accessible database, named APASdb (http://mosas.sysu.edu.cn/utr), which can visualize the precise map and usage quantification of different APA isoforms for all genes. The datasets are deeply profiled by the sequencing alternative polyadenylation sites (SAPAS) method capable of high-throughput sequencing 3′-ends of polyadenylated transcripts. Thus, APASdb details all the heterogeneous cleavage sites downstream of poly(A) signals, and maintains near complete coverage for APA sites, much better than the previous databases using conventional methods. Furthermore, APASdb provides the quantification of a given APA variant among transcripts with different APA sites by computing their corresponding normalized-reads, making our database more useful. In addition, APASdb supports URL-based retrieval, browsing and display of exon-intron structure, poly(A) signals, poly(A) sites location and usage reads, and 3′-untranslated regions (3′-UTRs). Currently, APASdb involves APA in various biological processes and diseases in human, mouse and zebrafish.

A comprehensive analysis of 3′ end sequencing data sets reveals novel polyadenylation signals and the repressive role of heterogeneous ribonucleoprotein C on cleavage and polyadenylation

“…Many experimental protocols to capture transcript 3 ′ ends and enable studies of the dynamics of polyadenylation have been developed (for review, see de Klerk et al 2014), and consequently, a few databases of 3 ′ end processing sites are available (Lee et al 2007;Derti et al 2012;You et al 2014). However, none of these databases has used the entire set of 3 ′ end sequencing data available to date, and thus, their coverage is limited.…”

“…By comparing the 3 ′ end processing sites from two recent genome-wide studies (Derti et al 2012;You et al 2014), we found that a substantial proportion was unique to one or the other of the two studies (Supplemental Table 1). This motivated us to develop a uniform and flexible processing pipeline that facilitates the incorporation of all published sequencing data sets, yielding a comprehensive set of high-confidence 3 ′ end processing sites.…”

“…4B, 5B), probably indicating that the canonical poly(A) sites of constitutively expressed transcripts are identified by the majority of protocols, whereas poly(A) sites that are only used in specific conditions were captured only in a subset of experiments. Although in constructing our catalog we used most of the reads generated in two recent studies (>95% of the reads that supported human 3 ′ end processing sites in these two data sets mapped within the poly(A) site clusters of our human catalog) (Derti et al 2012;You et al 2014), only 61.82% (You et al 2014) and 41.38% (Derti et al 2012) of the unique processing sites inferred in these studies were located within poly(A) clusters from our human catalog. This indicated that a large fraction of the sites that were cataloged in previous studies is supported by a very small number of reads and lacks canonically positioned poly(A) signals.…”

“…Sequencing of mRNA 3 ′ ends takes advantage of the poly(A) tail, which can be captured with an oligo-dT primer. More than 4.5 billion reads were obtained with several protocols from human or mouse mRNA 3 ′ ends in a variety of cell lines (Shepard et al 2011;Lin et al 2012), tissues (Derti et al 2012;You et al 2014), developmental stages Ulitsky et al 2012), and cell differentiation stages (Hoque et al 2013), as well as following perturbations of specific RNA processing factors Jenal et al 2012;Martin et al 2012;Almada et al 2013;Ji et al 2013). Although some steps are shared by many of the proposed 3 ′ end sequencing protocols, the studies that employed these methods have reported widely varying numbers of 3 ′ end processing sites.…”

Subcellular RNA profiling links splicing and nuclear DICER1 to alternative cleavage and polyadenylation

“…The relative positions of these CPA sites in the 3 ′ UTR of a gene and the sequence elements contained between them may result in the differential regulation of APA isoforms through RNA stability, translational efficiency, or subcellular localization (Sandberg et al 2008;Weill et al 2012). The realization that APA occurs in most genes, and thereby potentially regulates widespread gene expression, triggered a flurry of investigations to annotate APA profiles in different tissues and cellular states (Ozsolak et al 2010;Derti et al 2012;Lin et al 2012;You et al 2015). RNA-seq-based global analysis extended the proportion of human genes that contain multiple CPA sites to ∼70% and demonstrated that APA not only occurs throughout tissues but may also be evolutionarily conserved (Tian et al 2005;Derti et al 2012;Miura et al 2013).…”