how can i transfer the format into the input like your example?

mglev1n / ldscr

LD Score Regression in R

https://mglev1n.github.io/ldscr/

Other

31 stars 3 forks source link

how can i transfer the format into the input like your example? #6

Closed pubTools closed 1 year ago

pubTools commented 1 year ago

thanks for your excellent job!

for example: from 248afb2be678117588ac7fd452b2dfc to eb834c42f8ef8f52e3a84ca6bee7970

mglev1n commented 1 year ago

There are several tools to transform your summary statistics into the "munged" format required by the ldsc tools. Some examples include:

munge_sumstats.py (original python script provided by the ldsc authors): https://github.com/bulik/ldsc
MungeSumstats (package for R): https://neurogenomics.github.io/MungeSumstats/articles/MungeSumstats.html
munge (function from the GenomicSEM R package): https://github.com/GenomicSEM/GenomicSEM

fangzhoufan commented 3 weeks ago

If I have a GWAS dataframe which includes A1 A2 SNP N FRE and so on, could I use the math (Z=beta/sebeta) to calulate the Z, instead of using the MungeSumstats packages ? That correct ?

mglev1n commented 3 weeks ago

Mathematically that calculation of Z-score is correct. However, because the Z-scores are signed, you'll want to ensure your alleles correspond to those used in the reference file (and the sign of the Z-score is adjusted accordingly. The munge functions referenced above perform several checks of the data (see: https://github.com/GenomicSEM/GenomicSEM/blob/master/R/munge_main.R), and using them is strongly encouraged.