Hi Yijing,
In version 2.0.5 we fixed some issues that were causing problems related to exon analysis. Are you using this version.
If you are using this version, then it could be an issue with the bed files or something else. To test this, you can try the sample data in our tutorial:
http://code.google.com/p/altanalyze/wiki/Tutorial_AltExpression_RNASeq
If this works fine for you then it could be the bed file or the way the samples are named. For more on creating the exon bed file see:
http://code.google.com/p/altanalyze/wiki/BEDTools
bamToBed -i accepted_hits.bam -split| coverageBed -a stdin -b /home/user/BAMtoBED/hESC_differentiation_exons.bed > /home/user/RNASeqStudy/Sample1/day0_s1__exons.bed
For more on the file naming conventions see:
http://code.google.com/p/altanalyze/wiki/TopHat
Finally, you can simply assess the quality of the bed file by looking at the read counts associated with exons in genes you know should be highly expressed and genes you know are not expressed based on the junction data (you can search for the Ensembl gene IDs in this file).
Best,
Nathan
On Dec 7, 2011, at 8:21 AM, Zhang Yijing <yz...@partners.org> wrote:
> Dear Dr Salomonis,
> I work in Mass General Hospital. I'm using altanalyze to detect alternative splicing events from my RNA-seq samples. Using only junctions.bed, I could find hundreds of splicing events, most of them seems true from IGV, but when I add exon.bed to the bed file, no splicing event could be discovered. The same thing happens for all my comparisons.
> For exon.bed calculation, I used the exon file generated to BAMtoBED folder, and used Bedtools command:
> coverageBed -abamaccepted_hits.bam -b Mm_cere_exons.bed -split >exons.bed
> Is there anything wrong with my process?
> Also, could alanalyze take pair-end RNA-seq data into account?
> Thanks!
> yijing
