LiftOver for genetic variants results in misplacement

[quattro]

Hi all,

Thanks for developing a fantastic tool for the community. I'm running into a liftOver issue within another package, and thought I would raise the issue here.

Namely, when I try to liftover the GRCh37 coordinates to GRCh38 for the SNP rs190850470, I see the following output:

rs190850470 chrX:50858748 -> rs190850470 chr2:33174205, which is at odds with the dbSNP entry.

Is this a bug within the R liftOver implementation, or is this an artifact liftOver algorithm itself? I found the UCSC browser gives the same mapping to chr2. Alternatively, I've found that NCBI remap correctly maps the variant to chrX:51115913. So I'm not sure what the path forward is here.

Any help or insights would be greatly appreciated.

[BEFH]

I think this is due to the use of the UCSC chain files and is working as intended. I generally discard cross-chromosome mappings.