Modeling SNP and quantitative trait association from GWAS catalog using CLG Bayesian network
Document Type
Article - Abstract Only
Publication Date
2017
Keywords
Bayes methods, Genetics, Mathematical model, Bioinformatics, Diseases, Linear regression, Analysis of variance, belief networks, genomics, quantitative trait association, GWAS catalog, CLG Bayesian network, genome-wide association studies, genetic methods, Conditional Linear Gaussian Bayesian network, CLG distribution
Abstract
Genome-wide association studies (GWAS) are a type of genetic methods that have recently received intensive attention. In this paper, we study the construction of the Bayesian network from the GWAS catalog for modeling SNP and quantitative trait associations. Existing methods in the literature can only deal with categorical traits. We address this limitation by leveraging the Conditional Linear Gaussian (CLG) Bayesian network, which can handle a mixture of discrete and continuous variables. A two-layered CLG Bayesian network is built where the SNPs are represented as discrete variables in one layer and quantitative traits are represented as continuous variables in another layer. We propose the method for specifying the CLG Bayesian network, focusing on the specification of the CLG distribution for quantitative traits. We empirically evaluate the construction method, and results demonstrate the effectiveness of our method.
Citation
L. Zhang, Q. Pan and X. Wu, "Modeling SNP and quantitative trait association from GWAS catalog using CLG Bayesian network," 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Kansas City, MO, 2017, pp. 266-269. doi: 10.1109/BIBM.2017.8217661
Comments
Principal Investigator: Xintao Wu
Acknowledgements: This work is supported in part by U.S. National Science Foundation (DGE-1523115 and IIS-1502273). We also would like to thank Dr. Xinghua Shi from UNC Charlotte for helpful discussions and feedback.