Bayesian detection of expression quantitative trait loci hot spots.

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• High-throughput genomics enables genome-wide gene expression quantification.
• Combining expression data with sequence variation identifies expression quantitative trait loci (eQTL).
• Detecting clusters of eQTL, or regulatory hotspots, is crucial but challenging.

Purpose of the Study:

• To develop a novel statistical modeling strategy to identify regulatory hotspots.
• To estimate the propensity of genetic markers to influence multiple expression traits simultaneously.
• To pinpoint individual genetic effects (eQTL) and uncover regulatory hotspots.

Main Methods:

• A hierarchical regression model implemented within a Bayesian framework.
• A stochastic search algorithm, HESS, for efficient probing of sparse genetic marker subsets.
• Simulation studies and application to real mouse and human genetic datasets.

Main Results:

• HESS outperforms existing methods, especially with large numbers of transcripts.
• The method successfully identified regulatory hotspots not detected by conventional approaches.
• Demonstrated applicability and new insights in diverse real-case datasets.

Conclusions:

• The HESS method offers significant advantages for identifying functional eQTL hotspots.
• This approach effectively reveals key regulators underlying biological pathways.
• Provides a powerful tool for dissecting complex gene regulatory networks.