Global and unbiased detection of splice junctions from RNA-seq data

Adam Ameur,Anna Wetterbom,Lars Feuk,Ulf Gyllensten

doi:10.1186/gb-2010-11-3-r34

Abstract

SplitSeek can be used to detect novel splicing events in SOLiD RNA-seq data without the need for a pre-defined library.

Highlights

High-throughput sequencing of mRNA opens unprecedented opportunities to identify the spectrum of splice events in a sample on a global scale
By applying our method to available highthroughput sequencing of RNA (RNA-seq) data from mouse cells [6], we showed that splice junctions can be identified at almost nucleotide precision and with a very low false-discovery rate (FDR)
If a splice junction is located in the gap between the two anchors, the two parts are matched to different genomic positions

Summary

Introduction

High-throughput sequencing of mRNA opens unprecedented opportunities to identify the spectrum of splice events in a sample on a global scale. The typical approach for detecting splicing in RNA-seq experiments has been to map the reads to a junction library consisting of predefined exon-exon boundaries [1,2,3,4,5,6]. These strategies can successfully recover many splice events, they do not analyze splicing from a truly global and unprejudiced perspective. A genome with 100,000 (105) exons, which is a low estimate for mammalian genomes, would yield 1010 combinations To address this problem, the size of the junction library must be reduced dramatically, and most methods consider only the candidates involving known exons within the same gene. This type of analysis is restricted to the relatively small number of species in which coordinates of genes and exons have been found

Methods

Results

Discussion

Conclusion