Abstract
A complex disease is usually characterized by a few relevant disease phenotypes which are dictated by complex genetical factors through different biological pathways. These pathways are very likely to overlap and interact with one another leading to a more intricate network. Identification of genes that are associated with these phenotypes will help understand the mechanism of the disease development in a comprehensive manner. However, no analytical model has been reported to deal with multiple phenotypes simultaneously in gene-phenotype association study. Typically, a phenotype is inquired at one time. The conclusion is then made simply by fusing the results from individual analysis based on single phenotype. We believe that the certain information among phenotypes may be lost by not analyzing the phenotypes jointly. In current study, we proposed to investigate the associations between expressed genes and multiple phenotypes with a single statistics model. The relationship between gene expression level and phenotypes is described by a multiple linear regression equation. Each regression coefficient, representing gene-phenotype(s) association strength, is assumed to be sampled from a mixture of two normal distributions. The two normal components are used to model the behaviors of phenotype(s)-relevant genes and phenotype(s)-irrelevant genes, respectively. The conclusive classification of coefficients determines the association status between genes and phenotypes. The new method is demonstrated by simulated study as well as a real data analysis.
Paper version not known (
Free)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call