gwas.py gwas crashes with a KeyError: 0 trying to rename the IID column

[rzetterberg]

When running gwas.py gwas it crashes with a KeyError: 0 when trying to rename the IID column:

Here's the offending line from gwas.py:

    iid_column_name = pheno_dict.loc[pheno_dict['TYPE'] == 'IID', 'FIELD'][0]

At first I was using a dict file that looked like this:

But now that you are not allowed to have multiple IID type columns, I changed the FID column to NOMINAL type,
(as I saw done in one of the use cases dict files):

When I then ran gwas.py gwas I was given the error above. Before commit 92fc902 I did not get this error.

Let me know if you need me to provide any additional information.

[espenhgn]

Hi @rzetterberg. Can you try without the FID column? This document suggests it is not required: https://github.com/comorment/containers/blob/main/gwas/pheno_geno_specification.md#phenotypes-and-covariates

[ttfiliz]

Hi @rzetterberg could you also try renaming the 'COLUMN' in the header of the dict file as 'FIELD'?

[rzetterberg]

Thanks for the suggestions, @espenhgn and @ttfiliz! 👍🏻

Renaming the COLUMN in the header of the dict file to FIELD had no effect on the outcome.
But removing the FID column resolved the problem.

In the link @espenhgn sent it says:

If FID column is included in the phenotype file it will be simply ignored.

But if you include the FID column the program will crash, like it did for me. I'm not sure how many people will encounter this crash, since they won't be adding the FID column anyway. The reason I had the FID column was that I was simply rerunning an analysis from a couple of months ago with the latest version of the containers.

[espenhgn]

Good! It would be good to add a test to check that the included columns are valid perhaps, producing a more meaningful error message than what you encountered above.

[github-actions]

This issue appears to be stale due to non-activity