Without genome reference, how to find exon-exon junctions given isoforms

Question

[lwc628]

Is there any program that detects the exon-exon junctures when provided with isoforms without using the genome reference sequence?

[spiderdijon]

I'm not sure exactly what you're trying to do. You want to detect non-canonical exon-exon junctions correct? What do you mean when provided with isoforms - you want to use the isoform's exons as a basis for the spice-junctions? Why do you need to do this without a reference?

[lwc628]

Thanks for your comments. I generated a denovo transcriptome assembly using Trinity, and in it I have multiple isoforms for a gene model. Without reference, I don't know exon-exon junctions in those isoforms, which is why I asked this question.

[spiderdijon]

What organism are you working with?

Answer 1

[spiderdijon]

I'm going to assume your organism has a reference genome, otherwise this is a different question all together. From a scan of the Trinity documentation, it looks like it gives you the assembled fasta sequences of your genes and transcripts. So to find the exon-exon junctions, you're going to need to compare your transcript sequences to your reference organism's sequence. Try using something like BLAT to align your transcripts, which will give you the coordinates that each "exon" maps to. Depending on how many transcripts you need to do this for, you may need to run the command-line version of BLAT, then convert the output to something like BED format, assuming you need to visualise your transcript exon boundaries.