|
Post by vrisner on Jul 1, 2019 14:25:05 GMT
Hello,
I'm having issues using the --cojo-file flag because the summary statistic data I am working with does not have a sample size. When I downloaded the summary statistic data, the only information I received was CHR, SNP, BP, A1, A2, FRQ_A, FRQ_U, INFO, OR, SE, and P. According to the GCTA documentation, files should have SNP, A1, A2, freq, b, se, p, N. When I try to run my analyses without including N since I don't have that information, I get the error Error: invalid sample size in line: "rs9442371 T C 0.5675 0.0112 0.1072 0.9169" (which is the first line of my data). Is there a way to run my analysis without having a N value (like putting 0 in the N column)?
gcta64 --bfile LD_reference_UKBB --cojo-file CPD_in_ukbb.txt --cojo-sblup 2.619e7 --cojo-wind 1000 --out sblup_cpd
I'm trying to compare traditional Polygenic Risk Score (PRS) methods (which has worked in the past without the N that SBLUP requires) with SBLUP methods, so my end goal is to adjust the weight of the effect size using SBLUP and re-run PRS analysis.
Thanks!
|
|
ax
New Member
Posts: 27
|
Post by ax on Dec 6, 2019 1:50:31 GMT
Hi vrisner,
Usually, you can find the sample size in the original paper. You can put that number in the N column.
|
|