search

saigegit / SAIGE

Development for SAIGE and SAIGE-GENE(+)
GNU General Public License v3.0
64 stars 27 forks source link

About Pvalue #50

Closed SheaCheng2000 closed 1 year ago

SheaCheng2000 commented 2 years ago

Hi,

Did you use P-value adjustment method (e.g., false discovery rate) when you calculate the P-value of Burden or SKAT in SAIGE-GENE?

Thanks, Shea

weizhou0 commented 2 years ago

Hi Shea,

We don't have any P-value adjustment method when calculate the P-value of Burden or SKAT in SAIGE-GENE.

Thanks, Wei

SheaCheng2000 commented 2 years ago

Got it :) Thanks!

SheaCheng2000 commented 1 year ago

Hi,

May I ask another question about P-value :) When you calculated P-value, you used one-tailed test, right? I am not so good at statistics, just extrapolated from your code pchisq(...,lower.tail=False). From the R documentation, "lower.tail=False" means probability P[X>x] In my understanding, the essence of p-value calculation in burden test of one certain gene/region should be a 2*2 chi-square test (case vs control and observed vs expected). From the plot and formula of chi-square(df=1), it should be one-tailed. Did I understand it correctly?

Thanks again, Shea

saigegit commented 1 year ago

Hi Shea,

Yes your understanding is correct expect that in the mode we use, covariates are included, while accounting for sample relatedness and unbalanced case control ratios.

Thanks, Wei