Question: hg19 and hg38 format

getian107 / PRScsx

Cross-population polygenic prediction

MIT License

65 stars 20 forks source link

Question: hg19 and hg38 format #5

Closed jasmine9764 closed 3 years ago

jasmine9764 commented 3 years ago

Dear Ge Tian, We are trying to test your excellent tool in east asian population, however, as our biobank imputation data is released in hg38 version while the PRS-CSx LD reference panels was constructed as hg19 from 1000 Genomes. To be specific, we might need .bed instead of .hdf5. We know there is tool online (eg ucsc annotation) but it might be time-consuming and we are not sure if the coverage rate will affect our resul.

I look forward to your opinion. Thank you so much!

Jasmine

getian107 commented 3 years ago

Hi Jasmine, PRS-CS and PRS-CSx use rs ID, A1 and A2 to match SNPs between summary statistics, reference panels and the target dataset. Base pair positions are not used in any computation and are included for reference only. I don't think there will be any meaningful differences in results using hg19 vs. hg38.