We have performed a meta-analysis to test the association of various Single Nucleotide Gene polymorphisms (SNPs) with the risk of ischemic stroke.
We have stratified the studies on the basis of Asian and Caucasian and have performed the meta-analysis for each SNP.
We wish to adjust the reported p-values by performing the Benjamini-Hochberg procedure to decrease the False Discovery Rate (FDR) in our meta-analysis.
Kindly tell me how to proceed with the same using STATA software and which commands to use?
If I understand correctly, you have one p-value (resulting from a meta-analysis) per SNP and you want to perform a Benjamini Hochberg correction. If that's the case, I'd take all the p-values to excel and do the correction there, which is really easy to do - I think this link explains it really well: http://www.biostathandbook.com/multiplecomparisons.html.
Shubham,
In Stata you could try: (two variables: id (subgroup/strata ID) and p for p-values)
gsort -p
gen order=_n
count
gen criticalBH=((r(N) - order+1)*.05)/(2*r(N) )
gen value="."
replace value="*" if p< criticalBH
replace value="-" if value=="."
list id p criticalBH value, sep(0) noo
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis