A powerful allele based test for case-control association studies

Abstract

In a case-control study aimed at localizing disease variants, association between a marker and the disease status is often tested by comparing the marker allele frequencies among cases and controls. These marker allele frequencies are expected to be different if the marker is associated with the disease. The power of the commonly used allele based test is based on the marker allele frequency; markers with a low minor allele frequency have less power to be detected (if they are associated with the disease), than markers with high minor allele frequency. Therefore the strategy of selecting markers for follow-up study based on their p-values, favors markers with a high minor allele frequency. We propose an allele based test that does not have this (unwanted) property and is therefore more powerful for markers with a low minor allele frequency. This test may, therefore, be more effective when searching for rare causal variants. The asymptotic power function of the test is derived and simulation studies are performed for finite sample properties of the test. Next, the existing and the proposed tests are applied to data; this is not included yet. In the light of the current interest in detecting association between complex phenotypes and causal variants with a low minor allele frequencies, this test is expected to be of relevance.