Estimating genetic inheritance in case-control studies

Case-control genetic association study is an efficient tool to search for the deleterious genetic variants predispose to human complex diseases, where the additive mode of inheritance is commonly assumed. However, how the genetic variants influence the occurrence of a certain disease is impossible to know beforehand. We show numerically that the existing procedures using the Hardy-Weinberg equilibrium test to choose the genetic model might be inconsistent. We propose a new method to choose the genetic model by adopting co-dominant code for risk allele and logistic regression model. Extensive computer simulation results demonstrate superiorities of the new method. Applications to six single nucleotide polymorphisms(SNPs) for breast cancer and eight SNPs for Type 2 Diabetes further show good performances of proposed method. In order to specify a genetic model for an allele, our method has some merits and is consistent as sample size is large. We propose to apply our method in related fields.