COMPARISON OF STATISTICAL METHODS IN TESTING OF HARDY-WEINBERG EQUILIBRIUM IN CASE-CONTROL GENETIC ASSOCIATION STUDIES
Objective(s).Assessment of the assumption of Hardy-Weinberg Equilibrium (HWE) has been commonly used in genetic association studies because it can detect the genotyping errors and describe the genotype distribution in a population. Different statistics have been suggested for testing HWE in case-control studies. In this paper, we compared the results of Chi-square, Wald, Chen and Chatterjee and Likelihood ratio tests to find out which one offers a better performance. Methods.Four statistics are applied to four different simulated datasets and real data from a genetic association study entitled ‘Association of MTHFR C677T (Methylenetetrahydrofolate reductase) Polymorphism and Vascular Thrombosis’. Simulated datasets were constructed considering two different scenarios with as the violation of HWE excess heterozygosity. The rejection probability of each test was compared with others. Results.The results of Chi-square, Wald and Chen and Chatterjee Wald tests were almost similar. The likelihood ratio test showed the same results in some cases but yielded different results from other tests on real data. Conclusion(s).Among the four statistics considered, it seems that likelihood ratio test has a better performance than the others and genetic researchers can obtain more details if they use this test in case-control genetic association studies.
Chen and Chatterjee, likelihood ratio, MTHFR C677T c2 Hardy-Weinberg.