Dear Sumedha

There are many assumptions of HWE can not be found in the real human population, like the absence of selection, and we know the health care in the human community give a chance for weak genotypes to survive ( treated diseases and care in hospitals ) and reproduction ( treated sterility and improve fertility ) as an exception of HWE, another factor is the active migration among the human populations.

The final factor is the population size must be large ( hundreds ) and randomly mating.

I think HWE is theoretically expression to detect the occurrence of evolution in the population or not, and it is very difficult to practice in the real world.

Good Luck

It's necessary to test this samples to be in HE. The assumptions and conditions it's not in your data.

Hi Sumedha,

Yes, you can and you should for comparability across different SNPs (if you are testing a couple of SNPs).

However, given that your derived allele is rare, it is almost impossible for you to deviate from HWE when you have observed no rare homozygotes. Imagine the two possible scenarios below (N=100):

1) Genotype counts of 99, 1 and 0 gives you a HWE exact test p-value of 1 (no evidence of deviation).

2) Genotype counts of 65, 35 and 0 gives you a p-value of 0.037 (significant deviation).

In the second scenario, you expect to see at least a few 'rare homozygotes' given that 35% of the population are heterozygotes.

Hope it helps.

P.S. Use the hwe.exact function in R (due to presence of genotype counts below 5) and for more info see our paper (see link).

Article The Essentiality of Reporting Hardy-Weinberg Equilibrium Cal...

Hi all! Thank you for your answers. Thank you Naser for the article. It was very illustrative, clarified some of my doubts and given rise to further doubts. 

I have genotyped thalassaemia patients and further, have genotyped SNPs that are unrelated to the causative mutation. I am studying the spectrum of different SNPs in beta thalassaemia patients and correlating it to phenotype. In this scenario, if I am not wrong, all the genotyped SNPs need to be in HWE before I attempt to correlate it to phenotype. 

So, since the alleles stated in the question are not related to the causation of beta thalassaemia (sampling has been done for beta thalassaemia patients), I will not have rare homozygotes of the SNP in question in my study sample (since sampling was not done for this SNP). So will this not be reflected in the HWE?

Secondly, HWE uses, as mentioned in the article, Chi-square test. I need to use Fisher's Exact test. But when calculating the expected values for Fisher's, will it not be calculated assuming the alleles are in HWE? So is it right to use Fisher's Exact test for calculation and consequently test for alleles to be in HWE?

Thank you in advance

Yes HWE is necessary to test. If the alleles you are testing are in HWE, it represents the real population and your results are justified and correct.