Unstable MiXeR Results

[nasser1el]

Hi there,

I've been using MiXeR to compare and analyze five different GWAS files. For the most part, MiXeR has been running properly, but MiXeR either produces no figures or produces figures with unreasonably high values for any analysis involving one of the GWAS files (let's call this GWAS file "trait 1").

When looking over the results of the univariate analysis, trait 1 was found to have a high polygenicity (pi_mean = 0.0395), low heritability (h2 mean = 0.0627), and a Mixture_vs_Inft_BIC value of -11.3. I understand that these factors are likely why MiXeR is producing unstable results whenever trying to do bivariate analysis with the trait 1 file. However, I was able to get MiXeR to successfully run when using the munged version of the trait 1 file with LDSR (due to the lower polygenicity and higher heritability that resulted). I understand that this is not recommended, so I was wondering what alternatives you would suggest to be able to properly run MiXeR using the original trait 1 file without having to use munge_sumstats.py?

Any help is appreciated. Thank you!