Problems with cufflinks GTF annotated files and DEXSeq

[ml439]

I'm trying to use a Cufflinks GTF that includes novel isoforms as an input for DEXSeq.
My RNA-Sseq libraries were stranded and my cufflinks command included fr-secondstrand, but now when i run the phyton script i get the following error:

Traceback (most recent call last):
File "dexseq_prepare_annotation.py", line 55, in <module>
*** exons[f.iv] += ( f.attr['gene_id'], f.attr['transcript_id'] )
*File "_HTSeq.pyx", line 514, in HTSeq._HTSeq.GenomicArray.__getitem__ (src/_HTSeq.c:10576)
KeyError: 'Non-stranded index used for stranded GenomicArray.'


Here’s the line 514 of my gtf file:

1 Cufflinks exon 6113070 6113253 988 - . gene_id "CUFF.74"; transcript_id "ENSRNOT00000061956"; exon_number "2"; FPKM "0.0020481810"; frac "0.496199"; conf_lo "0.000000"; conf_hi "0.011202"; cov "0.023020";

I’m lost and I don’t know what to do. Anybody knows what the problem is and how can it be solved? I would appreciate any help i could get from the DEXSeq experts. Thank you

[dpryan]

Do a quick:

cut -f 7 something.gtf | uniq | sort | uniq

If you see a ".", then that's the source of the problem.

[ml439]

Thanks, I've used your command and this is what i've got:

+
-
.

Does that mean that i have unstranded data in my GTF? how could that be if my libraries are stranded and i used the fr-secondstrand in my cufflinks? any idea how to solve the issue?
Would it be possible to remove those lines from the GTF? or what would be the optimal way of dealing with this issue?

thanks!

[dpryan]

I haven't a clue why that's happening, but something like:

awk 'BEGIN{OFS="\t"}{if($7!=".") print $0}' original.gtf > filtered.gtf

should filter things.

[ml439]

Thank you very much for your help.
I've managed now to remove those lines (6) and run dexseq_prepare_annotation.py without an error. You've saved my day