Power Analysis for LD score Regression Genetic Correlation

[Ella Wigmore]

Hi,

I'm trying to calculate how much power I have in a genetic correlation analysis in LD score Regression. My sample sizes are 11,800 and 17,200 with about 1 million common SNPs between the samples. I was just wondering if there was a method to calculate the power for running a genetic correlation in LD score? Or whether the GCTA-GREML power calculator would be acceptable to use?

I would really appreciate you help.

Many thanks,

Ella

[Raymond Walters]

Hi Ella,

We don’t have a direct power calculation, but the GCTA-GREML one should be reasonably close. Power for LD score regression can be expected to be a bit lower than GCTA-GREML (the cost of using summary statistics rather than genotypes, and for being robust to population stratification). E.g. see Table 1 of this note for a brief comparison of standard deviations in simple simulations.

Cheers,

Raymond

--
You received this message because you are subscribed to the Google Groups "ldsc_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ldsc_users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ldsc_users/d9bd69ac-301f-4e25-8686-3e4c2c8a3449%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.