Cufflinks/cuffmerge text file busy error

[gormleymp]

I am getting the following error when running cuffmerge.

Traceback (most recent call last):
File "/usr/local/bin/cuffmerge", line 576, in <module>
sys.exit(main())
File "/usr/local/bin/cuffmerge", line 559, in main
compare_meta_asm_against_ref(params.ref_gtf,params.fasta,output_dir+"
/transcripts.gtf")
File "/usr/local/bin/cuffmerge" line 447, in compare_meta_asm_against_ref
os.remove(tmap)

OSError: [Errno 26] Text file busy: '.merged_asm/tmp_meta_asm.transcripts.gtf.tmap'

Has anyone come across this error and come up with a solution? It appears to be internal to cuffmerge as it is dealing with file management of temporary files.

Thanks,
Mike

[gormleymp]

FYI interestingly the --keep-tmp option does not fix this error.

[GenoMax]

What kind of disk storage system are you running these jobs on? Is this on a cluster or a standalone server?

[gormleymp]

Standalone Virtual Machine running CentOS 6.

[GenoMax]

I guess then some kind of disk I/O contention is probably not applicable. Are you consistently seeing that error with cuffmerge jobs?

[gormleymp]

New to cufflinks. This is the only RNAseq data we have in the pipeline at the moment.

[GenoMax]

Are you running anything else besides the cuffmerge job at the same time? What OS is this virtual machine running on?

[gormleymp]

Nothing else on the VM. VM is on Windows 7.

[GenoMax]

This may be an issue specific for your environment and may coming from python (based on some googling). You have not said if you are getting the error consistently with all jobs or if this was a one time issue.

You may want to email the tophat/cufflinks list to see if they have any suggestions.

[gormleymp]

Thanks for your input.

I said earlier that I am new to cufflinks and it is the only RNAseq data we have at the moment to work with. I meant to imply that I have not run cuffmerge on other jobs. Sorry if it was not clear. If it comes down to it, I may run the cufflinks protocols using the test data described in the Nature Protocols paper "Differential gene and transcript expression analysis...". This would be a good control to evaluate if it is a data issue or an environment issue.

I have emailed tophat/cufflinks with this issue. I will post the resolution to guide other users when I work through it.

[gormleymp]

Success. The issue seems to be that I was running cuffmerge from a shared directory between the VM and the host machine. Running cuffmerge from a directory on the VM alone resolved the error.

[GenoMax]

Success. The issue seems to be that I was running cuffmerge from a shared directory between the VM and the host machine. Running cuffmerge from a directory on the VM alone resolved the error.

That would certainly do it. Glad it got resolved.