search

stephenslab / susieR

R package for "sum of single effects" regression.
https://stephenslab.github.io/susieR
Other
169 stars 42 forks source link

whys lambda is so large #212

Open koujiaodahan opened 6 months ago

koujiaodahan commented 6 months ago

I try to run fitted_rss1 <- susie_rss(bhat = sumread$beta_alt, shat = sumread$V10, n = N, R = ldmatrix, L = 10, estimate_residual_variance = FALSE,verbose=TRUE) based on my gwas summary statistics and the LD matrix of same samples. After that,I run kriging_rss for diagnose. but the lambda is always above 0.4 and the plot is that. why?

image
koujiaodahan commented 6 months ago

Besides that , running the susie_rss function at a locus(chr6), it breaks with "Error in susie_suff_stat(XtX = XtX, Xty = Xty, n = n, yty = (n - 1) * : The estimated prior variance is unreasonably large. This is usually caused by mismatch between the summary statistics and the LD matrix. Please check the input. Calls: susie_rss -> susie_suff_stat Execution halted"

stephens999 commented 6 months ago

The results suggest something wrong with the way you are computing your ld matrix or z scores