The answer to this question is "it depends on the biology being studied". Is there a reason why that could matter for your phenotype? What is your phenotype?? Generally, you should read about recessive, additive, and dominant models of inheritance.

Most sophisticated GWAS association testing can do any of these tests. For instance, with SNP Test, you could write:

./snptest  -summary_stats_only -data ./example/cohort1.gen ./example/cohort1.sample \
**-frequentist 2**   -o ./example/ex.out

see the section called "Priors for Binary Trait models".

Plink does this too, of course; as per the documentation these commands tend to look like this:

./plink --bfile mydata --model-dom --fisher