A powerful MAF-neutral allele-based test for case-control association studies

In a case-control study aimed at locating autosomal disease variants for a disease of interest, association between markers and the disease status is often tested by comparing the marker minor allele frequencies (MAFs) between cases and controls. For most common allele-based tests the statistical power is highly dependent on the actual values of these MAFs, where associated markers with low MAFs have less power to be detected compared to associated markers with high MAFs. Therefore, the popular strategy of selecting markers for follow-up studies based primarily on their p-values is likely to preferentially select markers with high MAFs. We propose a new test which does not favor markers with high MAFs and improves the power for markers with low to moderate MAFs without sacrificing performance for markers with high MAFs and is therefore superior to most existing tests in this regard. An explicit formula for the asymptotic power function of the proposed test is derived theoretically, which allows for fast and easy computation of the corresponding p-values. The performance of the proposed test is compared with several existing tests both in the asymptotic and the finite sample size settings.