Some tips for quality control (QC):
Despite controlling for HWE, MAF and Call Rate, some funny SNPs can scape. I recently analysed some SNPs which pass all QC criteria but had only 1 individual with genotype 22 and an extreme value. This individual screwed up my analysis. Solutions:
1. Eliminate phenotypic outliers.
2. Add another QC criterion: minimum genotype frequency (MGF) 5%.
Moreover, if you know the structure of your population, e.g. breeds, HWE should be tested having such structure into account, e.g. within breeds.

Everyday we know a bit more than yesterday and a bit less than the morrow
Jules