featureCounts segmentation fault, Is there any solution yet ?

Question

[EagleEye]

http://seqanswers.com/forums/showthread.php?t=50353

featureCounts -Q 10 -F GTF -a MY_ANNOTATION.gtf -t exon -g gene_id -o mypath/out_counts.txt mypath/celline1.bam

========== _____ _ _ ____ _____ ______ _____ 
===== / ____| | | | _ \| __ \| ____| /\ | __ \ 
===== | (___ | | | | |_) | |__) | |__ / \ | | | |
==== \___ \| | | | _ <| _ /| __| / /\ \ | | | |
==== ____) | |__| | |_) | | \ \| |____ / ____ \| |__| |
========== |_____/ \____/|____/|_| \_\______/_/ \_\_____/
v1.4.5-p1

//========================== featureCounts setting ===========================\\
|| ||
|| Input files : 1 BAM file ||
Segmentation fault (core dumped)

[Istvan Albert]

They have already told you that there is a bug and you need to wait for a new release.

[EagleEye]

Thanks Istvan, I am waiting. But meanwhile looking into different solutions to this problem, if someone has.

[Devon Ryan]

In the mean time just use htseq-count. It's slower but it'll presumably work too.

Answer 1

[EagleEye]

Great there is a new patched version for Subread (subread-1.5.0-p1-source) released today. It works perfectly :-)

Thanks a lot for all your suggestions and help.... See you with some other issue.

Answer 2

[kanika.151]

It worked for me when I called it in R.

I followed this protocol:

http://bioinf.wehi.edu.au/RNAseqCaseStudy/

[EagleEye]

Hi Kanika,

Thanks for your reply, please let me know in which R version you installed Rsubread?

[kanika.151]

R version 3.2.2 on Fedora

[EagleEye]

Thanks again for the suggestion, but I get the same issue using R. I guess I should wait for the next patched version.

        ==========     _____ _    _ ____  _____  ______          _____  
        =====         / ____| |  | |  _ \|  __ \|  ____|   /\   |  __ \ 
          =====      | (___ | |  | | |_) | |__) | |__     /  \  | |  | |
            ====      \___ \| |  | |  _ <|  _  /|  __|   / /\ \ | |  | |
              ====    ____) | |__| | |_) | | \ \| |____ / ____ \| |__| |
        ==========   |_____/ \____/|____/|_|  \_\______/_/    \_\_____/
       Rsubread 1.20.2

//========================== featureCounts setting ===========================\\
||                                                                            ||
||             Input files : 1 BAM file                                       ||

 *** caught segfault ***
address 0x3020002fa, cause 'memory not mapped'

Segmentation fault (core dumped)

[kanika.151]

I think this should help you. It is because of the size of GTF file. :/

https://stat.ethz.ch/pipermail/bioconductor/2014-May/059512.html

[EagleEye]

My GTF file is very small with just 3,000 genes in it. I think it is the problem with larger BAM as mentioned by Wei Shi. You can see his comments on seqanswers link on my first post.

[kanika.151]

Have you tried using HTSeq instead? Both give you the same results.

[EagleEye]

I have not tried. I just waited for the update from featureCount. Since I wanted to maintain the consistancy with my analysis pipeline, I did not try it. But it is an interesting question.

[kanika.151]

Don't do that. All the analysis pipeline need one count module which can be done by either featureCount or HTSeq the difference is only of the platforms. I did both :)

[EagleEye]

No, I haven't done yet. I am already done with all analysis part.

How much correlation did you get between these two methods? It will be nice if you can share it.

[Devon Ryan]

The difference you get in counts is incredibly small if it even still exists. Originally, the only difference was due to an off-by-one error in htseq-count (see the featureCounts paper for a figure on this).