Categorize genotype data into 0s,1s and 2s to run GWAS

Question

[doodle]

Hello,

I have run a GWAS on X chromosome for a few phenotypes. But I was told that an additive model would not be suitable for it. I was then asked to categorize the genotypes into 0s, 1s and 2s for females and 0s and 1s for males and run the gwas. Can you please help me with how to run such a GWAS? Is it a type of non-additive model?

[Kevin Blighe]

The person may mean the 'genotypic' model? - 0, 1, and 2 are encoded as categorical variables. Irrespective, although females have 2 X chromosomes, one of these undergoes inactivation, meaning that genes will only be expressed from one chromosome or the other (although certain genes can 'escape' X inactivation, too).

[doodle]

I am not sure if it's a model he was referring to. So what happened is this : I ran an association test using the additive model for 28 SNPs and I'm getting very significant p-values - to the order of 10e-16! So my supervisor was wondering if it is an artifact and wanted to study it a little deeper and asked me to categorize the genotypes into 0, 1 and 2 and run the analysis. I'm aware of the inactivation process.