Estimation and testing for the effect of a genetic pathway on a disease outcome using logistic kernel machine regression via logistic mixed models

摘要: Background: Growing interest on biological pathways has called for new statistical methods modeling and testing a genetic pathway effect health outcome. The fact that genes within tend to interact with each other relate the outcome in complicated way makes nonparametric more desirable. kernel machine method provides convenient, powerful unified multi-dimensional parametric of effect. Results: In this paper we propose logistic regression model binary outcomes. This relates disease risk covariates parametrically, parametrically or nonparametrically using machines. allows possible interactions among same relationship We show estimation components can be formulated mixed model. Estimation hence proceed framework standard software. A score test based Gaussian process approximation is developed are illustrated prostate cancer data set evaluated simulations. An extension continuous discrete outcomes generalized models its connection linear discussed. Conclusion: Logistic provide novel flexible tool effects Their close attractive performance make them have promising wide applications bioinformatics biomedical areas.