TopHat error: could not open writer pipe gzip

[JonB]

Hi,

I am running TopHat on my MacBook Air. Everything seemed to be running fine, but after maybe 30min of "Mapping left_kept_reads to genome" it stopped with this error message:

Error: could not open writer pipe gzip -cf - < ./tophat_out/tmp/left_kept_reads_seg123.fq.z

Anyone know what this means?

Thanks!

Jon

[EGrassi]

Is your disk full or almost there? (just guessing, I've never seen that error myself)

[JonB]

I have about 23 GB of disk space left. I tried to monitor the disk space as the analysis was running and couldn't see if I ran out of space.

[SDiehl]

I have the same issue:

[2013-04-25 18:00:32] Beginning TopHat run (v2.0.7)
-----------------------------------------------
[2013-04-25 18:00:32] Checking for Bowtie
Bowtie version: 2.0.5.0
[2013-04-25 18:00:32] Checking for Samtools
Samtools version: 0.1.18.0
[2013-04-25 18:00:32] Checking for Bowtie index files
[2013-04-25 18:00:32] Checking for reference FASTA file
[2013-04-25 18:00:32] Generating SAM header for ../../reference/petMar2
format: fasta
[2013-04-25 18:00:34] Preparing reads
left reads: min. length=201, max. length=16867, 266767 kept reads (0 discarded)
[2013-04-25 18:01:04] Mapping left_kept_reads to genome petMar2 with Bowtie2
Error: could not open writer pipe gzip -cf - < Sample1_petMar2/tmp/left_kept_reads_seg508.fq.z

I have enough disk space and everything. I tried with several versions of tophat2 and bowtie2 and several samples and always the same error. The file tmp/left_kept_reads_seg508.fq.z is empty, but the rest of the temp files are ok. The only thing that was the same all the time was the reference, which I build myself from the UCSC fasta from petMar2. Everything worked fine previously with a different version of the reference. I also didn't see any other error in the logs. So the reference should be fine since bowtie2 works.

[JonB]

I don't remember exactly, but I think this problem was because I ran out of memory. I haven't been running tophat locally since, but on a cluster with ~8 GB of memory I never have this problem.

[SDiehl]

Thanks for the reply JonB. I tried running tophat2 on two different powerful servers which both have much more than 8GB memory.

[Lizex]

Originally Posted by SDiehl

Thanks for the reply JonB. I tried running tophat2 on two different powerful servers which both have much more than 8GB memory.

Hi SDiehl
Did you get a solution for your problem you had regarding Tophat not opening gzip. I think I have a similar problem. I have compiled and installed Tophat2. When running tophat on the test_data, I got this error:
test_data $ tophat2 -r 20 test_ref reads_1.fq reads_2.fq
Error: cannot find compression program gzip

It says it cannot find the compression program gzip. Any idea or suggestion what I should do?

Thanks in advance

Lizex

[GenoMax]

Originally Posted by Lizex

It says it cannot find the compression program gzip. Any idea or suggestion what I should do?

Thanks in advance

Lizex

Depending on the unix distro you are using there will be different ways of getting "gzip" installed. At a minimum you can find the official download here: http://ftp.gnu.org/gnu/gzip/

It may be simple to look up (and install) gzip using your software repository (e.g. software center in Ubuntu).

[SDiehl]

Originally Posted by Lizex

Hi SDiehl
Did you get a solution for your problem you had regarding Tophat not opening gzip. I think I have a similar problem. I have compiled and installed Tophat2. When running tophat on the test_data, I got this error:
test_data $ tophat2 -r 20 test_ref reads_1.fq reads_2.fq
Error: cannot find compression program gzip

It says it cannot find the compression program gzip. Any idea or suggestion what I should do?

Thanks in advance

Lizex

In my case gzip was not the problem. For some strange reason it turned out the be the reference. After I redid the reference everything worked fine again.

As GenoMax already said, most likely you need to install gzip or add it to your path, such that tophat can find it.

[Lizex]

Thanks guys.

I downloaded the gzip and put it in my PATH. It seems to have done the trick.

Cheers

Lizex

[edm1]

In case this is useful to anyone in the future; My problem was that I was running on a Beowolf cluster and the network drive I had my reads on was only available from the login nodes, not the worker node.

[seqprone]

Originally Posted by SDiehl

In my case gzip was not the problem. For some strange reason it turned out the be the reference. After I redid the reference everything worked fine again.

As GenoMax already said, most likely you need to install gzip or add it to your path, such that tophat can find it.

Could you please explain how you 'redid the reference'? Thanks!