--robust = 1 bring inflated result for rare variants?

[Shicheng-Guo]

Dear GEM team,

I am wondering whether --robust = 1 bring inflated result for rare variants compared with --robust = 0. Meanwhile, --robust = 1 always have more significant results compared with --robust = 0, correct?

Thanks

Shicheng

[hanchenphd]

Hi Shicheng,

Thank you for your interest in GEM. I would not recommend using the robust standard errors for rare variants, because asymptotics depends on the minor allele count and may not work well for rare variants.

Best, Han

[Shicheng-Guo]

Thank you Han!! It is consistent with what I noticed!! Thank you for your confirmation and Thank you for this great tool for GxE analysis!!