search

getian107 / PRScsx

Cross-population polygenic prediction
MIT License
65 stars 20 forks source link

The UKB Reference is GRCh37? #43

Closed koujiaodahan closed 7 months ago

koujiaodahan commented 8 months ago

The UKB Reference is GRCh37?

koujiaodahan commented 8 months ago

And I noticed that the 1000genome LD panel in the directory is also GRCh37?

getian107 commented 8 months ago

Yes they are in GRCh37. But note that PRS-CS/x uses rs ID, A1 and A2 to match SNPs; the base pair position is for reference only and is not used in any calculation. As a result genome build usually has a minimal impact on the result.

koujiaodahan commented 8 months ago

Yes they are in GRCh37. But note that PRS-CS/x uses rs ID, A1 and A2 to match SNPs; the base pair position is for reference only and is not used in any calculation. As a result genome build usually has a minimal impact on the result.

Thanks for your reply, and I will not liftover the position, and just guarantee the rsID, A1, A2.

koujiaodahan commented 7 months ago

Yes they are in GRCh37. But note that PRS-CS/x uses rs ID, A1 and A2 to match SNPs; the base pair position is for reference only and is not used in any calculation. As a result genome build usually has a minimal impact on the result.

Hi, would you recommend the dbsnp version for annotating rsID? I noticed that a specific position is named by different rsID in different dbsnp version

image
getian107 commented 7 months ago

Hi - I think you can either use the version that is consistent with the reference panel or the latest version. A small number of SNPs may have inconsistent mapping between versions but in the vast majority of cases this should be minimal impact of the predictive power of the final score.