Minimum Requirements and Assessing Reliability of Estimates

[alterecho11]

Hello,

Can MiXeR analysis be performed on any phenotype, or are there minimum N, Neff or SNP-h2 requirements?

After reading some papers, I understood it is suggested to have over 50,000 sample size (N > 50,000) and SNP-h2 above 0.05, but from my results, I see that a larger sample size may make up for lower SNP-h2 (and vice-versa), and I get both positive AIC and BIC values.

How should I determine the reliability of MiXeR analysis results? Should I rely solely on AIC and BIC values, or should other factors (mentioned above) be considered before reporting the results?