search

statgen / raremetal

A flexible tool for meta-analysis
8 stars 9 forks source link

Multiallelic variant causes rest of VCF to be skipped #7

Open welchr opened 6 years ago

welchr commented 6 years ago

If a VCF contains a multiallelic variant/record, the remaining rows are skipped without an error message.

I think it's tripped up in this bit of code:

https://github.com/statgen/savvy/blob/c2cec45b5af8a306f65133ba5d060af90ef3e9cd/include/savvy/vcf_reader.hpp#L779-L787

The reader stops producing records once it hits that point.

Probably would be good to show the user an error message that it hit a multiallelic variant, and offer an option to skip them (maybe --skip-multiallelic?)

suraj86nong commented 4 years ago

Hi, my vcf file has multi-allelic variants and I would like to include multi-allelic variants in the meta-analysis by RAREMETAL. There was no problem with single variant analysis in RAREMETALWORKER. But when I run the RAREMETAL for single variant as well as group-based meta-analysis, the multi-allelic variants were skipped from the analysis. How can we solve this issue? Are there any alternatives for multi-allelic variants analysis in RAREMETAL?

welchr commented 4 years ago

One of our collaborators has apparently patched in support for multi-allelic variants. Once we are able to see the code, we can assess how difficult it is to merge in, and whether it would make it into the next release.