Alignments for DEXseq

[MDonlin]

Hi,

I'm just ready to try out DEXseq, but have a question about the alignment files. I ran the alignment using TopHat with my standard transcript.gtf file as the reference. Can I use this alignment file with the flattened exon file produced by dexseq_prepare_annotation.py or do I need to do the alignment against the flattened exon file first?

To make it clearer:
Working with single-end reads
Alignment done with h99_transcript.gtf using TopHat
BAM -> SAM using samtools
h99_transcript.gtf -> h99_exons.gff using dexseq_prepare_annotation.py

The question is, do I need to redo the alignment against h99_exons.gff or can I use the alignment generated against h99_transcript.gtf?

Thanks,
Maureen

[areyes]

Hi Maureen,

I have no much experience with tophat, but what I can say its that in the flattened file, the exons are broken to define non-overlapping "exonic regions". These are defined base on the same regions from the real exons. So I would not expect a lot of changes by redoing the alignment in junction reads. That´s my feeling, but somebody correct me if I am wrong!

Alejandro

[MDonlin]

I was able to test it yesterday and TopHat does not recognize the flattened exon file as having mapped junctions. I did the alignment using the starting gtf file and was able to use the dexseq_counts.py to get counts from the resulting SAM and flattened exon file. I've not yet had a chance to use the DEXSeq to determine if there is any alternative exon usage.