snps icon indicating copy to clipboard operation
snps copied to clipboard
verified publisher icon apriha

Force build conversion?

Open teepean opened this issue 10 months ago • 6 comments

Hello!

I have received a 23andme formatted file that is in GRCh38 but detects as GRCh37. Is there any way to force the conversion or possibly use more positions to detect the build?

Thanks!

teepean avatar Feb 27 '25 18:02 teepean

Hello and thanks for the issue! Yes, if you could provide an RSID or two from the file, the detection could add those.

apriha avatar Feb 28 '25 05:02 apriha

The root cause was that the file was created from a GRCh38 CRAM with WGSExtract and the 23andme header still has mentions of build 37.

teepean avatar Feb 28 '25 05:02 teepean

Hmm, maybe there could be an option to detect the build using only RSID positions, instead of using them only when the build isn’t in the header.

apriha avatar Feb 28 '25 05:02 apriha

I think WGSExtract needs to be fixed to create a compatible header depending on the build.

teepean avatar Feb 28 '25 06:02 teepean

In terms of snps, build detection could be made more robust to just use the data (RSID positions) first, and then fall back to the headers.

apriha avatar Feb 28 '25 14:02 apriha

To be honest that would be the best solution.

teepean avatar Feb 28 '25 17:02 teepean