A Comprehensive Survey of Non-Canonical Splice Sites in the Human Transcriptome Guillermo E

A Comprehensive Survey of Non-Canonical Splice Sites in the Human Transcriptome Guillermo E

10564–10578 Nucleic Acids Research, 2014, Vol. 42, No. 16 Published online 14 August 2014 doi: 10.1093/nar/gku744 A comprehensive survey of non-canonical splice sites in the human transcriptome Guillermo E. Parada†, Roberto Munita*†, Cledi A. Cerda and Katia Gysling* Nucleus Millennium in Stress and Addiction, Department of Cellular and Molecular Biology, Faculty of Biological Sciences, Pontificia Universidad Catolica´ de Chile, Alameda 340, Santiago, Chile Received June 11, 2014; Revised July 31, 2014; Accepted August 1, 2014 Downloaded from ABSTRACT Proper intron recognition and removal rely on consen- sus sequences located at the intron/exon boundaries. Dinu- We uncovered the diversity of non-canonical splice cleotide sequences at these boundaries have been found to sites at the human transcriptome using deep tran- be strongly conserved and relevant for proper splicing (3–5). scriptome profiling. We mapped a total of 3.7 bil- Nearly all introns belong to the so-called U2-type, which are http://nar.oxfordjournals.org/ lion human RNA-seq reads and developed a set of spliced by the major spliceosome and are flanked by GT– stringent filters to avoid false non-canonical splice AG splice site dinucleotides. The most frequent exception site detections. We identified 184 splice sites with to this rule are the U2-type GC–AG splice sites, compris- non-canonical dinucleotides and U2/U12-like con- ing ∼0.9% of human splice sites (6). On the other hand, sensus sequences. We selected 10 of the herein iden- about 0.4% of the human splice sites belong to the U12- tified U2/U12-like non-canonical splice site events type. These introns are processed by the minor spliceosome and successfully validated 9 of them via reverse and even though they were first described to have AT–AC dinucleotides at the intron/exon boundaries, the vast ma- at Pontificia Universidad Cat�lica de Chile on May 19, 2016 transcriptase-polymerase chain reaction and Sanger / jority of them contain GT–AG sites (7). Indeed, the AT–AC sequencing. Analyses of the 184 U2 U12-like non- sites comprise only ∼0.09% of the splice sites (6). canonical splice sites indicate that 51% of them Despite the disruptive splicing effects that have mutations are not annotated in GENCODE. In addition, 28% of of splice site dinucleotides (3–5), introns with non-canonical them are conserved in mouse and 76% are involved splice sites (that is, with sequences other than GT–AG, GC– in alternative splicing events, some of them with AG or AT–AC at the intron/exon boundaries) have been tissue-specific alternative splicing patterns. Interest- reported to be efficiently removed (6,8–12). These reported ingly, our analysis identified some U2/U12-like non- non-canonical splice sites have U2/U12-like splice site con- canonical splice sites that are converted into canon- sensus sequences (U2/U12-like non-canonical splice sites). ical splice sites by RNA A-to-I editing. Moreover, the For instance, evolutionary conserved U2-like introns with U2/U12-like non-canonical splice sites have a differ- GA–AG splice sites have been identified in FGFR genes (8,9) and a functional GT–TG splice site has been found ential distribution of splicing regulatory sequences, in the GNAS gene (10,11). Although the first global analy- which may contribute to their recognition and regula- sis of splice sites in the human transcriptome, conducted 14 tion. Our analysis provides a high-confidence group years ago, did not find confident evidence for non-canonical of U2/U12-like non-canonical splice sites, which ex- splice sites (13), most recent analyses based on expressed hibit distinctive features among the total human sequence tag (EST) sequences have reported U12-like non- splice sites. canonical splice sites (6) and more examples of U2-like GT– TG introns (12). The advent of high-throughput sequencing technolo- INTRODUCTION gies has provided an unprecedented opportunity to explore the complexity of mammalian transcriptomes (14). For in- Most genes in higher eukaryotes are interrupted by non- stance, analyses of RNA-seq data have resulted in the dis- coding sequences, called introns, which are precisely excised covery of thousands of new splice sites and alternative splic- from pre-mRNAs during splicing. Nuclear pre-mRNA in- ing events in the human transcriptome (15–17). However, trons are processed by the spliceosome, a complex macro- the high resolution power of high-throughput sequencing molecular machine composed of five small nuclear RNAs and numerous proteins (1,2). *To whom correspondence should be addressed. Tel: +562 23542654; Fax: +562 23542660; Email: [email protected] Correspondence may also be addressed to Roberto Munita. Tel: +562 23542657; Fax: +562 23542660; Email: [email protected] †The authors wish it to be known that, in their opinion, the first two authors should be regarded as Joint First Authors. C The Author(s) 2014. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Nucleic Acids Research, 2014, Vol. 42, No. 16 10565 has not been used to generate a non-canonical splice site junctions were extracted from unique gapped alignments catalog on the human transcriptome. that had an anchor ≥15 without mismatches. The splice To make a comprehensive analysis of non-canonical junctions that were supported by three or more cDNA/EST splice sites present in the human transcriptome, we have sequences were considered as present in cDNA/EST align- processed nearly 3.7 billion RNA-seq reads from 16 human ments. tissues and a lymphoblastoid human cell line (GM12878). As the alignments to the reference genome can span over Our systematic analysis provides a list of high-confidence indels and single nucleotide polymorphisms (SNPs), false non-canonical splice sites and an insight into their charac- non-canonical splice junctions could be detected. For this teristic features. Our comprehensive identification of non- reason, we discarded the gapped alignments to the refer- canonical splice sites will improve the human transcriptome ence genome which had non-canonical splice junctions that annotation. Further understanding of the mechanism un- spanned over indels or SNPs reported on SNPdb135. derlying the recognition and processing of non-canonical splice sites could expand our knowledge of the splicing Downloaded from process. We provide the full annotation and quantification of the entire list of high-confidence canonical and non- Library-guided detection of splice junctions canonical splice junctions for each analyzed human tissue All the splice junctions identified in RNA-seq and (available as a UCSC Hub at http://54.214.245.35/Tracks/ cDNA/EST alignments were used to create a library of Splicing/hub.txt). splice junctions (Figure 1A). We mapped the previously http://nar.oxfordjournals.org/ used RNA-seq data, in addition to tissue-specific RNA-seq MATERIALS AND METHODS data from 16 human tissues, to the library of splice junctions Processing of RNA-seq data using Bowtie (25), configured to align with a SOAP-like ‘-v 2’ mode allowing two mismatches with a seed length of 50 nt We used the used RNA-seq data of GM12878 cell line pro- (Figure 1B). We checked that the new splice junction map- vided by ENCODE project (18) and RNA-seq data of a ping reads found did not map to other genomic location and mixture of 16 human tissues generated by Illumina Body did not have repetitive or low complexity sequences. More- Map 2.0 project (for additional information see Supple- over, to ensure the reproducibility of the splice junction de- mentary Data). The reads were processed in order to re- tection, we selected the splice junctions that were present in at Pontificia Universidad Cat�lica de Chile on May 19, 2016 move the adapters and low-quality sequences (PHRED at least two different data sources (tissues extracted from the score ≤ 10) with FASTX toolkit (http://hannonlab.cshl.edu/ same individual were considered as one data source) (Figure fastx toolkit/index.html) and sickle (https://github.com/ 1C). najoshi/sickle) for trimming of paired-end reads. After this, only the reads ≥50 nt long were kept. RNA-seq reads of GM12878 provided from the ENCODE project (18) (Sup- plementary Data) were aligned to the diploid genome of Non-canonical splice junction processing GM12878 (19) and the RNA-seq of the 16 human tissues mixture to the reference genome (hg19) using MapSplice Direct repeat sequences located at the splice junctions (20). MapSplice was configured to detect canonical and generate alignment ambiguity and multiple putative non- non-canonical splice junctions with anchor length ≥8, a canonical splice sites. This phenomenon can lead to false minimal intron size of 1 nt and allowing two mismatches non-canonical splice site annotation (13). For this reason, 25 nt each. when a non-canonical splice junction was flanked by direct repeats, we checked that it did not have ambiguous canoni- cal splice sites (13). Moreover, based on the position weight Ab initio detection of splice junctions matrices (PWMs) reported for canonical human splice junc- Splice junctions were extracted from unique gapped align- tions (6), we selected the alignment that yields the non- ments that have an anchor ≥8 nt. To avoid false non- canonical splice junction with the best fit into the canonical canonical introns derived from alignment errors, the read PWMs. alignments that have non-canonical splice junctions were Based on the canonical splice junction PWMs, we clas- corroborated by BLAT (21) and were discarded if they had sified the non-canonical introns /as U2 U12-like or non- repetitive sequences present at Repbase (22) or low com- U2/U12.

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    15 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us