gffread segmentation fault

[gokhulkrishnakilaru]

Hello All,

I donno what is going wrong with the gffread utility from cufflinks.

This is the command I have been using

gffread -w /path/test.fa -g /path/hg19_ucsc.fa /path/test.gtf

My GTF file is just 4MB with around 48K records. The output fasta file prints only around 42K records with a size of 480MB and then I see this error.

/path: line 16: 28068 Segmentation fault (core dumped) gffread -w /path/test.fa -g /path/hg19_ucsc.fa /path/test.gtf

I have another GTF file of size 12MB which is being used to generate a fasta file of size 3.7GB and then it generates the same segmentation error.

All helps are appreciated.

[swaraj]

Fragment your gtf file into five separate chunks and run each with gffread. it may help you unserstand the core of the fault.

[gokhulkrishnakilaru]

In the third file, I am getting the segmentation fault error. Could you please help me what it means?

[seb567]

In the third file, I am getting the segmentation fault error. Could you please help me what it means?

A program is represented as segments in a computer's memory. A segmentation fault means that the program tried to access an address that is not located in one of its segments.

Usually, this means that there is a bug in the software.

[gokhulkrishnakilaru]

A program is represented as segments in a computer's memory. A segmentation fault means that the program tried to access an address that is not located in one of its segments.

Usually, this means that there is a bug in the software.

I broke my large file into chunks of 10000 lines each which left me with 10 files. The program runs fine for 9 files and throws this error for the last file.

Any thoughts?

[seb567]

I broke my large file into chunks of 10000 lines each which left me with 10 files. The program runs fine for 9 files and throws this error for the last file.

Any thoughts?

Is there something unusual in the last compared to the nine others ?

[gokhulkrishnakilaru]

Is there something unusual in the last compared to the nine others ?

I am trying to figure out those things. I broke the last file into 1000 line chunks now. And all of them are giving a segmentation fault. I cross checked the end coordinate against the chrom sizes files and all of them seem to be within the limit.

I don't know whats wrong with it.

[will.jackson]

I have had the same problem with the filtering parameters of gffread (-J, -V -H etc). Can anyone suggest another program that does a similar thing to gffread? That is, filter transcripts based on CDS features and provide a multi-fasta format sequence file at the end?

[prios]

Hello all,

I had the same problem a while ago. I fragment my file into 10 and I got errors in two of them. In both cases I broke the files into smaller ones and (after more splitting) I found located the problem in two a gene with alternative splicing event. In my case the secondary transcript was bigger than the annotated gene sequence (the last exon coordinates were placed outside the gene).

I just erased the two features from my original file (I considered that it was not a great loss for my purposes) and it works fine now.

Hope it helps,

Pablo

[will.jackson]

Hi,
Thanks for your reply Pablo, I tried doing what you said and I broke down my files and kept getting the segfault even when the file was only 100 lines long(!). I think maybe I have a lot of alternate splicing in my organism (a basidiomycete (fungi)). Does anyone know of a program like gffread that can handle alternate splicing? Or another way I could get around this problem?

Thanks very much

Will

[asperjelly]

I'm having the same problem as Will. I've done the tophat2-cufflinks-cuffmerge pipeline to generate a merged GTF file. However, my organism has fairly high gene density, so cufflinks is predicting very long transcripts, which are not correct. I wanted to filter the merged GTF file using gffread to discard any transcripts that have internal stops (either the -V or -J parameter). However, I keep getting the 'segmentation_fault' error. I have tried to break up the merged GTF file into smaller sizes (such as 1000 lines), however the segmentation error persists. Does anybody know a solution to this problem?