Prediction of alternatively skipped exons and splicing enhancers from exon junction arrays
- Equal contributors
1 Department of Biostatistics and Informatics, Colorado School of Public Health, University of Colorado Denver, 4200 East 9th Avenue, B-119, Denver, CO 80262, USA
2 School of Mathematics and Statistics, Sydney Bioinformatics, Carslaw Building F07, University of Sydney, NSW 2006, Australia
3 Department of Epidemiology and Biostatistics, University of California San Francisco, Box 0560, San Francisco, CA 94143, USA
BMC Genomics 2008, 9:551 doi:10.1186/1471-2164-9-551Published: 20 November 2008
Alternative splicing of exons in a pre-mRNA transcript is an important mechanism which contributes to protein diversity in human. Arrays for detecting alternative splicing are available using several different probe designs, including those based on exon-junctions. In this work, we introduce a new method for predicting alternatively skipped exons from exon-junction arrays. Predictions based on our method are compared against controls and their sequences are analyzed to identify motifs important for regulating alternative splicing.
Our comparison of several alternative methods shows that an exon-skipping score based on neighboring junctions best discriminates between positive and negative controls. Sequence analysis of our predicted exons confirms the presence of known splicing regulatory sequences. In addition, we also derive a set of development-related alternatively spliced genes based on fetal versus adult tissue comparisons and find that our predictions are consistent with their functional annotations. Ab initio motif finding algorithms are applied to identify several motifs that may be relevant for splicing during development.
This work describes a new method for analyzing exon-junction arrays, identifies sequence motifs that are specific for alternative and constitutive splicing and suggests a role for several known splicing factors and their motifs in developmental regulation.