Post by javiermr on Feb 4, 2018 10:09:50 GMT
I am trying to conduct an LD-score regression for a GWAs conducted in Latin Americans.
As recommended by the authors (Bulik-Sullivan et al. 2015 Nat Genet), I decided to use the 1KG data (AMR individuals) to do this.
I am using this command to get LD-scores in GCTA:
However, checking the paper, they said they used only SNPs with MAF>1% for the LD-score regression. I am bit confused about that statement. Does it mean that they computed LD-SCORES with all SNPs, but when they did the regression they only used SNPs with MAF>1%? Or they computed LD-scores with only SNPs with MAF>1%?
What are the standard filters I should use for this analysis?
Thank you.
As recommended by the authors (Bulik-Sullivan et al. 2015 Nat Genet), I decided to use the 1KG data (AMR individuals) to do this.
I am using this command to get LD-scores in GCTA:
gcta64 --bfile amr1kg --ld-score --ld-wind 1000 --ld-score-adj --out amr1kg
However, checking the paper, they said they used only SNPs with MAF>1% for the LD-score regression. I am bit confused about that statement. Does it mean that they computed LD-SCORES with all SNPs, but when they did the regression they only used SNPs with MAF>1%? Or they computed LD-scores with only SNPs with MAF>1%?
What are the standard filters I should use for this analysis?
Thank you.