Hello, all. I am trying to run basic GWAS and Exome-WAS studies using LMM-BOLT and SAIGE, but I'm currently stuck early in the process with a BOLT-LMM error when constructing the GRM that I can't decipher. The command itself and the Swiss Army Knife interface appear fairly straightforward. Does anyone have access to codes I can reference that have been successfully implemented or, more broadly, general recommendations? I've attached my most recent log for reference.
[Image: BOLTLMM_log]
Thank you in advance.
Interesting. I didn't run into the same issue with SAIGE. I went back to the page with my most recent job, and I see that I included the sample file from the command and the sample files from the TopMED imputed data folder corresponding with each of the other chromosomes in my list of Inputs (using the SAK vocabulary here). Could this error be occurring because I've added too many sample files to my input?
I'll also mention that the current command includes a list of ~1100 IDs that were in the fam file but not in the sample file. So the sample file should've been unchanged, but I wrote a short script to remove the EIDs/FIDs from the plink files that were not in the sample file. Perhaps this could've caused the issue?
I have faced an issue where sample IDs are different from bgen so I had to edit a sample file and upload that for the analysis. Ur's may not be the same case, Try doing one chromosome at a time. When u kept chromosome 1..22 as an array, it looks like it is not taking all the files together for check. That could be the possible issue.
Comments
4 comments
Looks like a sample file is not matching. Have you encountered the same issue while running saige?
Interesting. I didn't run into the same issue with SAIGE. I went back to the page with my most recent job, and I see that I included the sample file from the command and the sample files from the TopMED imputed data folder corresponding with each of the other chromosomes in my list of Inputs (using the SAK vocabulary here). Could this error be occurring because I've added too many sample files to my input?
I'll also mention that the current command includes a list of ~1100 IDs that were in the fam file but not in the sample file. So the sample file should've been unchanged, but I wrote a short script to remove the EIDs/FIDs from the plink files that were not in the sample file. Perhaps this could've caused the issue?
I have faced an issue where sample IDs are different from bgen so I had to edit a sample file and upload that for the analysis. Ur's may not be the same case, Try doing one chromosome at a time. When u kept chromosome 1..22 as an array, it looks like it is not taking all the files together for check. That could be the possible issue.
Sounds reasonable. I'll try this suggestion--thanks!
Please sign in to leave a comment.