calculate f statistic for SNPs from GWAS

Question

[mk19726]

Hi,

I have two different types of list of SNPs for MR analysis. One list is just the significant SNPs from a published GWAS (not the whole summary data). Other lists of SNPs have the whole summary data

In R I would like to calculate the overall F statistic to assess instrument strength.

I saw a paper where they wrote this, "we estimated the r2 of each genetic instrument and summed them up to compute the overall r2 and F statistics using the sample size "

I would love to do this with my data too!

Can anyone explain how to do this in R?

I have rsid, beta, se, pval, N

Incase this effects things, one of my SNP lists for which I have summary data is binary, all others continuous

Thank you for your help!!!

[panmy2015]

Yes,I have the same question as you, have you got any answers?

[SGMS]

I also need to do the same. Did you find a way of doing this?

Thanks a lot

Answer 1

[daniel.coral]

In the supplementary material of this paper the F-statistic for a SNP is approximated by βₓ² / σₓ² for exposure X.