Bayesian Exploration of Multilocus Interactions on the Genome-Wide Scale
Ivan Kozyryev and Jing Zhang
DOI : 10.3844/ajbsp.2012.70.78
Current Research in Bioinformatics
Volume 1, Issue 1
Problem statement: Recent technological and scientific advances propelled the field of Genome-Wide Association Study (GWAS), which promises to be instrumental in linking many common complex diseases to their genetic origin. While so far such large-scale surveys have been moderately successful in identifying disease related genetic variants, much of disease heritability is still not accounted for by the discovered loci. There is an urgent need for advanced statistical methods for efficient automatic detection of complicated multilocus interactions on significant scales. Approach: Novel statistical methods based on Bayesian data analysis ideas, specifically Bayesian modeling, Bayesian variable partitioning, graphical and network models are promising to aid in search for missing disease heritability and shed light on complex biological processes involved in disease development. First crucial difference setting these methods apart from all the mainstream previous approaches (hypothesis testing methods) is their joint disease mapping capability via the simultaneous fitting of a statistical model for the whole case-control data set. Additionally, such Bayesian methods allow for the construction of complicated data models and quantitative incorporation of diverse prior information into the final statistical model. Results: The use of Bayesian techniques has already yielded new insights into the details of epistatic interactions across the genome associated with various important diseases. Conclusion/Recommendations: Bayesian approaches provide a way to detect and understand complicated multilocus interactions that already started to elucidate important disease pathways. As the field of GWAS matures, Bayesian strategies can surely aid in converting such multiple surveys into useful biomedical information.
© 2012 Ivan Kozyryev and Jing Zhang. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.