junctions usage

[fadista]

Dear all,

I would like to know if there is a tool out there that can handle the junctions file coming out from Cufflinks and treat them as for instance the DEXSeq Bioconductor package, but instead of exon usage it would do junctions usage analysis (or transcript isoform deconvolution).

Thanks.

[sudders]

Hi, I've done this using DEXSeq, I just used DEXSeq and treated the junciton counts as exon counts and then did DEXSeq as usual.

I found that the results as WAY more conservative than DEXSeq on exon counts (on exon counts, differentail genes: 426, on junction counts differential genes: 50), but that when I looked at the resuts(in for e.g. IGV) I believed far more of them .

I found that the problem with doing DEXSeq analysis on exon read counts is that almost all my differentially used exons were 3' UTRs becuase the were the longest. Using junctions gets around this because all junctions are the same 'length'.