Entering edit mode
9.1 years ago
snowy
•
0
Hello, I used qtscore()
in GenABEL to do GWAS. As my sample size is small, someone suggested me to do bootstraps without replacement and gave me a formula (A)
A: qtscore(sample(phdata(b.dat)$disease, replace=F), b.dat, trait ="binomial")
but I think this formula is wrong, because it only sampled phenptypes but without relevant genotype information.
I think the right formula is (B):
B: qtsocre(disease, b.dat, trait="binomial", idsubset=sample(gtdata(b.dat)@idnames, replace=F), times=1000)
Am I right? Or any other good suggestions?
I also have a question;
What's the meaning of "times" in qtscore()
, if times=1000, does it mean that in my formula the sample process run 1000 times?
If times= 1000, what's the meaning of P1df
in results? How can I calculate FDR(false discovery rate)?