Computational Methods to determine Epistatic Effects

点击翻译按钮获取中文摘要

Hanlon0073785 The investigator designs effective algorithms to identifyepistatic effects that contribute to a complex quantitativephenotype. The algorithms apply to datasets that, for eachindividual in a sample, identifies alleles at a fixed set ofmarkers along with the value of the quanitative trait. Thealgorithms seek an optimum fit for the phenotype as a sum ofepistatic effects. The approach is to perform repeated searchesfor an optimum fit in the space of all possible combinations ofeffects. The search utilizes random walk hill-climbingtechniques. These hill-climbs are adaptive in the sense thatcertain parameters, which control the choice of steps in therandom walks, evolve based on the outcomes of previous walks. It is widely accepted that the genetic components of mostinteresting quantitative traits (eg. blood pressure, height)involve interactions between many genes. So, computationalmethods to infer the genetic basis of quantitative traits mustinclude an analysis of epistatic effects, i.e., instances wherethe choice of allele at two or more genetic markers influencesthe value of the quantitative trait in ways that are differentfrom the sum of the effects of the individual alleles. In thisproject the investigator develops computational methods toidentify these epistatic effects by a process of doing repeatedsearches through the space of possible effects looking for onesthat explain a significant amount of the variation of the trait.These searches are self-educating in the sense that latersearches learn from earlier searches. As the number of searchesincreases, the adaptive mechanism focuses the search onparticularly strong epistatic effects.