Difference of p.value and p.value.NA in output files about saige HOT 3 CLOSED

Hi Tenghao,

You may check this page
https://github.com/weizhouUMICH/SAIGE/wiki/SAIGE-Hands-On-Practical
for the information

Thanks,
Wei

CHR: chromosome
POS: genome position
SNPID: variant ID
Allele1: Ref allele
Allele2: Alt allele
AC_Allele2: allele count of Alt allele
AF_Allele2: allele frequency of Alt allele
N: sample size
BETA: effect size
SE: standard error of BETA
Tstat: score statistic
p.value: p value with SPA applied
p.value.NA: p value when SPA is not applied
Is.SPA.converge: whether SPA is converged or not
varT: estimated variance of score statistic with sample related incorporated
varTstar: variance of score statistic without sample related incorporated

from saige.

@weizhouUMICH

Hi - I have a follow up on this. The column description is helpful. It would be good to know which pvalue to be used for downstream analyses (manhattan plot, gene based):

p.value: p value with SPA applied
p.value.NA: p value when SPA is not applied

Would you say p.value.NA is FDR adjusted?

Thank you.

from saige.

Hi Sariya,
It shouldn't be FDR adjusted, you can do FDR adjustment using either column p.valye.NA or p.value.
The advantage of using column p.value is when the ratio of cases to controls is a concern and you decide to continue the analysis with the values when SPA is applied.

from saige.