Snepe2gene xqtl mapping error

[Shiko]

Hi Tanya / Kyoko / all,

I'm running into a reproducible error at the xQTL mapping step in SNP2GENE for a subset of my GWAS files, while others submitted with identical settings complete fine.

Fails reproducibly:
- Job 755990 and 755922 — ALL_HF_male (failed on two separate submissions)
- Job 755929 — I_HF_male

Succeeds, same settings:
- Job 749207 and 755624 — ALL_HF_female
- Job 755933 — NI_HF_male
- Job 755931 — NI_HF_female
- Job 749100 — MDD_female
- Job 748884 — MDD_male

What's odd is the pattern isn't a clean male/female or HF/non-HF split — NI_HF works fine for both sexes, ALL_HF works for females, but ALL_HF_male and I_HF_male both fail every time. Settings across all 8 jobs are identical: P-value cutoff 1e-5, MAF 0.01, Gene type All, MAGMA window 35/10kb, GRCh37, full xQTL dataset selection.

Input validation passes cleanly on the failed jobs too (right columns detected, expected variant count), so it really does seem specific to xQTL mapping rather than the file format itself.

Any idea what might cause this, or a way to tell which xQTL dataset is triggering it? I'd love to keep xQTL settings consistent across all 8 comparisons rather than dropping xQTL mapping just for these two, since I'm comparing results across traits.

Happy to send the input files or more logs if that helps track it down.

Regards
Fran

[Tanya Phung]

Hi Fran, 

Jobs 755990, 755922, 755929 were run on 07-21-2026. This was before the bug was fixed. The bug was fixed on 07-22-2026: 

The reason why this job 755624 was successful was because the GWAS variants were not found in the pQTLs datasets which caused the bug. On the other hand, job 755922 failed with the same settings because the GWAS variants were found in the pQTLs datasets which triggered the code with the bug. The bug has been fixed so you could retry. Hopefully it works out. 

Best,
Tanya