REDIportal: a comprehensive database of A-to-I RNA editing events in humans

Abstract

RNA editing by A-to-I deamination is the prominent co-/post-transcriptional modification in humans. It is carried out by ADAR enzymes and contributes to both transcriptomic and proteomic expansion. RNA editing has pivotal cellular effects and its deregulation has been linked to a variety of human disorders including neurological and neurodegenerative diseases and cancer. Despite its biological relevance, many physiological and functional aspects of RNA editing are yet elusive. Here, we present REDIportal, available online at http://srv00.recas.ba.infn.it/atlas/, the largest and comprehensive collection of RNA editing in humans including more than 4.5 millions of A-to-I events detected in 55 body sites from thousands of RNAseq experiments. REDIportal embeds RADAR database and represents the first editing resource designed to answer functional questions, enabling the inspection and browsing of editing levels in a variety of human samples, tissues and body sites. In contrast with previous RNA editing databases, REDIportal comprises its own browser (JBrowse) that allows users to explore A-to-I changes in their genomic context, empathizing repetitive elements in which RNA editing is prominent.

Figure 1.

Computational workflow used to load RNA editing sites in REDIportal. Raw data in Fastq format are quality checked by FASTQC and aligned onto the reference human genome by STAR (14). REDItools (15) are then used to interrogate multiple read alignments using a large collection of known RNA editing sites from ATLAS repository (6) and RADAR database (11). WGS data are finally included in REDItools tables and, in turn, stored in REDIportal.

Figure 2.

RNA editing distribution along human tissues and a graphical overview of sites stored in REDIportal. A-to-I events collected in REDIportal derive from RNAseq data encompassing 55 human body sites grouped in 30 different tissues. The distribution of RNA editing events detected according to our computational workflow, depicted in Figure 1, are shown in (A) along tissue specific events (in blue). Here, tissue specific events are defined as sites edited at non-zero level in a unique tissue. RNA editing sites stored in REDIportal are classified in three main categories, shown in (B), depending on their location: (i) ALU, residing in Alu repetitive elements, (ii) REP, located in non-Alu repetitive elements and (iii) NONREP, placed in non repetitive regions. The vast majority of A-to-I changes occur in protein coding genes (73%) as shown in (C) and especially in intronic regions (D). In untranslated regions, RNA editing changes are more abundant in 3′ UTRs than in 5′ UTRs (D).

Figure 3.

Query and retrieval in REDIportal. Editing sites can be searched entering genome coordinates or gene symbols in combination with additional filters. In figure we show a query for editing positions occurring in non-repetitive elements (NONREP) of BLCAP gene. A-to-I sites found in REDIportal are displayed in a dynamic table with a variety of extra information such as: (i) a link to UCSC genome browser; (ii) the genomic position; (iii) the reference and edited nucleotide; (iv) the strand; (v) the dbSNP accession (if any); (vi) the editing location; (vii) the repeated element (if any); (viii) the gene symbol according to Gencode v19; (ix) the genic region; (x) the number of edited samples; (xi) the potential amino acid change; (xii) the PhastCons conservation score across 46 organisms; and (xi) a flag indicating in which database (ATLAS, RADAR or DARNED) is reported.

Figure 4.

Extra information triggered on demand. Users can access detailed information on demand triggering the content in child rows. As shown in figure, each child row comprises four panels, indicated by red arrows and highlighted in red rectangles, and consists of: (i) an interactive heat map to explore RNA editing levels across body sites; (ii) an interactive box plot to look at editing level variation per body site; (iii) alternative gene/transcript annotations according to RefSeq and UCSC databases and (iv) editing details with the number of edited samples/tissues/body sites and a link to a further table containing RNA editing levels and the exact number of RNAseq and WGS reads supporting each event.

Figure 5.

RNA editing details. For each editing position, users can request detailed information displayed in dynamic and filterable tables. Such tables include valuable biological evidence such as the tissue or body site of origin, the number of edited or unedited reads, the WGS support and the detected editing levels.

Figure 6.

JBrowse in REDIportal. RNA editing sites can be easily inspected in their genomic context though JBrowse. Left click triggers a pop-up window with genomic data. Right click opens a new window with editing levels, tissues and body sites, and supporting RNAseq and WGS reads.