mixed linear model

Question

[toyan.j.p]

Hi,

Would like to know when to use mixed linear model based association analysis as implemented in GCTA-MLMA compared to Plink analysis for quantitative traits.I have tried both but the results are very different in the sense the most prominent SNP after Plink analysis didn't come associated at all in GCTA.

Thanks in advance for help

Answer 1

[Mo]

Mixed linear model is when you want to cope with the fixed effects and random error, on the other hand association analysis is done to handle problem of a data which has rare variants or common variants or the combination of the two. So they are used for different purposes. There are two things you must keep in mind.

1. What you want to do and
2. What is the problem of your data (structure, dimension , type etc)

All these influence on the results you will achieve.