multipop_selection_pipeline stopped at fst calculation

[binshuangli]

I am testing the multipop_selection_pipeline with the test lactase data. After running the following command, I noticed that it did not generate the Fst results and other files as described in the manual.

multipop_selection_pipeline -p CEU_ids.txt -p YRI_ids.txt -i CEU_YRI_lactase.vcf --config-file defaults.cfg -a "--imputation" -c 2 --imputation

The multi_population.log contains following info:
2017-06-28 12:01:54,452 fst_calculation_weir command = /software/selectiontools/1.0.1/bin/vcftools --fst-window-size 1000 --fst-window-step 1000 --vcf .../SelectionPipelineTestData/lactase/CEU_YRI_lactase.vcf --weir-fst-pop first_pop.tmp --weir-fst-pop second_pop.tmp
2017-06-28 12:01:54,699 fst_calculation_weir failed to run /software/selectiontools/1.0.1/bin/vcftools --fst-window-size 1000 --fst-window-step 1000 --vcf
.../SelectionPipelineTestData/lactase/CEU_YRI_lactase.vcf --weir-fst-pop first_pop.tmp --weir-fst-pop second_pop.tmp

The fst directory contains the following files:
first_pop.tmp second_pop.tmp stderr.tmp stdout.tmp

Is there any something else I should try?

[murraycadzow]

Hi,
if you run the command "/software/selectiontools/1.0.1/bin/vcftools --fst-window-size 1000 --fst-window-step 1000 --vcf .../SelectionPipelineTestData/lactase/CEU_YRI_lactase.vcf --weir-fst-pop first_pop.tmp --weir-fst-pop second_pop.tmp" inside the fst directory what do you get?

thanks
Murray

[binshuangli]

Hi Murray,

Thanks for your quick reply. I tried that command and now it can finish the fst analysis and generate the following files: out.log out.windowed.weir.fst

I am curious how to move forward from here because I am interested in applying this package to calculate different statistics like Tajima's D, Fay and Wu's H etc. for my own data.

Thank you!

[murraycadzow]

The multi_selection_pipeline is actually just a wrapper for the selection_pipeline. I normally use selection_pipeline to run on single populations which generates Tajima's D, Fay Wu's H etc so that should hopefully work for you.

Murray

[lauren-mak]

I've run into the same error when running multipop_selection_pipeline with the test lactase data. Is there a quick fix I can apply to the code on my end to address this? I'm planning to run selectionTools on a dozen populations simultaneously, so it would be great to have a pre-made script that will automatically process all of them.

I'm also getting the some weird error message after using the standalone vcftools command:

Weir and Cockerham mean Fst estimate: 0.066819
Weir and Cockerham weighted Fst estimate: 0.17305
Illegal instruction: 4

Is this due to some kind of compilation error during the installation process? It looks like I still get the 'out.windowed.weir.fst' file. Is there any output that should be there but is missing?

When I run multipop_selection_pipeline I get an error at the same step, but a different one that seems to be related to how vcftools is reading the first_pop.tmp file:

Error:Could not open Individual file: first_pop.tmp

Any idea how to address these items? Thanks!