problem with run_indep_wkflow_with_gnu_par.pl

[hcdenbakker]

perl /home/henk/cortex/scripts/calling/run_indep_wkflow_with_gnu_par.pl --index subclade_index.txt --ref /home/henk/outbryk/species/Listeria/refdata/LmonoJ2_064.fasta --dir_for_ref_objects ref_new_wrapper --vcftools_dir /home/henk/vcftools --outdir out_NW --stampy_bin /home/henk/stampy-1.0.23/stampy.py --kmer 31 --procs 30 --prefix Lmono_NW Use of uninitialized value $line in scalar chomp at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 99. Use of uninitialized value $line in split at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 100. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 102. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 105. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 121. ls: cannot access outNW//vcfs/_wkraw: No such file or directory

[rmcolq]

This was the thing with the memory I think. Check the build log in one of the sample directories and see if it says it ran out of memory

From: Henk den Bakker notifications@github.com<mailto:notifications@github.com> Reply-To: iqbal-lab/cortex reply@reply.github.com<mailto:reply@reply.github.com> Date: Friday, 11 September 2015 17:33 To: iqbal-lab/cortex cortex@noreply.github.com<mailto:cortex@noreply.github.com> Subject: [cortex] problem with run_indep_wkflow_with_gnu_par.pl (#7)

perl /home/henk/cortex/scripts/calling/run_indep_wkflow_with_gnu_par.pl --index subclade_index.txt --ref /home/henk/outbryk/species/Listeria/refdata/LmonoJ2_064.fasta --dir_for_ref_objects ref_new_wrapper --vcftools_dir /home/henk/vcftools --outdir out_NW --stampy_bin /home/henk/stampy-1.0.23/stampy.py --kmer 31 --procs 30 --prefix Lmono_NW Use of uninitialized value $line in scalar chomp at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 99. Use of uninitialized value $line in split at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 100. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 102. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 105. Use of uninitialized value $sample in concatenation (.) or string at /home/henk/cortex/scripts/calling/build_samples_parallel.pl line 121. ls: cannot access out_NW//vcfs/wkraw: No such file or directory

— Reply to this email directly or view it on GitHubhttps://github.com/iqbal-lab/cortex/issues/7.

[iqbal-lab]

No Rachel, this error happens before it even tries to build a sample - it is failing to read the per-sample index file

[hcdenbakker]

Yup, no memory issue, none of the per-sample index files are present, except index_0, and that one is empty

[iqbal-lab]

So that is the failure. Is itbecausee run_independent hashardcoded Parallel --gnu ?

[hcdenbakker]

OK, I think i have found the 'bug', I was able to reproduce exactly the same error while writing a bash-based wrapper; parallel --gnu does not seem to like it if you try to put a variable in the {1..".$num_samples."} part of the script ( {1..$num_samples} in bash). Maybe this is linux distro related? Solved this in the bash wrapper by replacing {1.. $num_samples} with $(seq $num_samples); I don't know if there is a perl equivalent for seq.

[iqbal-lab]

Bloody parallels is distro dependent. But if you remove --gnu, you might now find the script works

[hcdenbakker]

Nope, removing --gnu does not do the trick for Ubuntu

[hcdenbakker]

OK, may have solved it. Created a pull request.