Tophat Stalled Error = -9

Question

[Craig Anderson]

Heya,

I've just tried out tophat on all of my QC'd RNAseq data using the the transcriptome as the index, but unfortunately its become snagged!

"Error: segment-based junction search failed with err = -9"

Can anyone offer any solutions? I'd really appreciate it! Thanks,

Craig

(Please excuse the formatting)

---

$ tophat -o /home/sbica1/tophat/tophatoutput/trinity -r 50 -m 2 -p 8 /home/sbica1/tophat/L.rubellus_trinitytranscriptomeindex/L.rubellus_trinitytranscriptomeindex trimmedop010_1.fastq,trimmedop008_1.fastq,trimmedop009_1.fastq,trimmedop008_3.fastq,trimmedop009_3.fastq trimmedop010_2.fastq,trimmedop008_2.fastq,trimmedop009_2.fastq,trimmedop008_4.fastq,trimmedop009_4.fastq

[Wed Nov  9 19:57:34 2011] Beginning TopHat run (v1.0.14)
----------------------------------------------- 
[Wed Nov  9 19:57:34 2011] Preparing output location /home/sbica1/tophat/tophatoutput/trinity/  
[Wed Nov  9 19:57:34 2011] Checking for Bowtie index files 
[Wed Nov  9 19:57:34 2011] Checking for reference FASTA file 
[Wed Nov  9 19:57:34 2011] Checking for Bowtie     Bowtie version:         0.12.5.0 
[Wed Nov  9 19:57:34 2011] Checking reads     seed length:     20bp     format:         fastq     quality scale:     phred33 (default) 
[Wed Nov  9 22:21:17 2011] Mapping reads against L.rubellus_trinitytranscriptomeindex   > with Bowtie 
[Wed Nov  9 23:51:38 2011] Joining segment hits 
[Thu Nov 10 00:37:24 2011] Mapping reads against L.rubellus_trinitytranscriptomeindex   > with Bowtie 
[Thu Nov 10 01:56:43 2011] Joining segment hits 
[Thu Nov 10 02:37:40 2011] Searching for junctions via segment mapping     
[FAILED] Error: segment-based junction search failed with err = -9

Answer 1

[Chris Penkett]

In the output directory, there is a logs/ sub-directory with a file called segment_juncs.log - you could look in there for output or error messages that may help.

When debugging failed TopHat runs, you can also try to run the individual commands from the command line yourself.

In same logs/ sub-directory there is another file called run.log. Look for a command which starts with the command segment_juncs.

e.g., $TOPHAT_DIR/src/segment_juncs --min-anchor 8 --splice-mismatches 0 --min-report-intron 50 ...

It's quite long, but you can paste it into a terminal from the same directory you started TopHat in. Also, when TopHat crashes it saves temporary files in the tmp sub-dir in the output directory, so it should work.

If it fails you may get a more helpful error message that way. It could be something like you're going over-quota and the tophat.py wrapper loses that information and just says a standard error message and exits.

Chris

[Craig Anderson]

Awesome, I'll give this a pop in the morning. Thank you!

Answer 2

[Jitendra]

Hi! this could be not-enough-memory error. Maybe you can try mapping the data set in smaller chunks and then merge the bam files at the end?