For many common heritable traits such as schizophrenia, multiple underlying genes seem to contribute to the risk of disease. As an alternative to genotype patterns (diplotypes) conferring risk, investigated in previous years under this grant, we developed an approach to case-control association studies based on a polygenic model of disease inheritance. The model assumes a number m of loci, each with two alleles, 1 and 2, where the 1 allele (risk allele) confers risk to disease. Thus, each individual can have a number of risk alleles ranging from m1 = 0 through m1 = 2m. We assume a common allele frequency, q = P(1), at each locus, and a threshold, k, such that an individual with m1 ≥ k is affected with the trait and otherwise is unaffected. We investigated properties of this model in terms of predicted segregation ratio and recurrence risk and compared resulting values with those for simple recessive and dominant models and with empirical values observed for schizophrenia. In addition, we implemented case-control analysis under such a model in a computer program and applied it to several published datasets.
修改评论