Hi everyone!
Does it matter if you use N effective or N total, when running LDSC for cross trait rG? I’m aware it will impact heritability estimates but not sure if only worried about rG.
You need h2_1 and h2_2 to move from covG to rG and you also need N (or Neff) to calculate covG. The original 2015 paper from Bulik-Sullivan et al ( An atlas of genetic correlations across human diseases and traits | Nature Genetics ) doesn’t distinguish between N and Neff. I assume Neff should be used, but I don’t know this for sure.
Do the sumstats of the traits you are interested in provide both N and Neff? Would love to hear from the faculty myself regarding your question.