Cufflinks 2.0.2 segmentation fault on long fragments

[plger]

I have a systematic segmentation fault problem with cufflinks (v.2.0.2), which is independent of the dataset analyzed and caused by the input parameters --frag-len-mean (m) and --frag-len-mean (s), arguably because I've got unusual fragment mean length and SD.
(This is not like the other cufflinks segmentation fault reported, which is at a different step in the run).

The process stops when estimating frag length distribution, right after the "Inspecting bundle ..." output:

cufflinks -q -u -p 1 -m 800 -s 200 -g $GTF -M $MASK -b $GENOME.fa data.bam
[10:14:40] Loading reference annotation.
[10:14:43] Loading reference annotation.
[10:14:43] Inspecting reads and determining fragment length distribution.
Processed 189391 loci.
> Map Properties:
>       Normalized Map Mass: 52735571.57
>       Raw Map Mass: 52735571.57
Segmentation fault

I've tried requesting lots of memory, but it didn't make a difference, and I also get the same result using cufflinks' test_data.sam with the S.cerevisiae genome, suggesting that insufficient memory is not the issue.
Using another GTF or changing cufflinks version (I've tried the last three) also didn't make a difference.

Any of the following changes independently allowed cufflinks to run smoothly:

• using -G instead of -g AND not providing a genome (-b)
• removing the -m and -s parameters
• reducing -m and -s to something more "normal", like 500 and 100 respectively.

Any help would be greatly appreciated,
Pierre-Luc

[sterding]

Did you solve the problem? I have the same error as you, but occurring in the next step of "Inspecting reads and determining fragment length distribution."

I found another post (http://seqanswers.com/forums/showthread.php?t=27951) where Anelda suggested to have the genome.fa and GTF in same order of chromosome. I'll like to try it.

[plger]

No, I couldn't solve it other than by making the changes I mention in the post.
In my case at least it was not a matter of sorting.

[gwilymh]

No, I couldn't solve it other than by making the changes I mention in the post.
In my case at least it was not a matter of sorting.

What do you mean specifically when you say 'sorting'?