Network-guided search for genetic heterogeneity between gene pairs

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Standard genetic analyses often overlook genetic heterogeneity and locus interactions, limiting understanding of complex diseases.
• Genetic heterogeneity can increase statistical power by aggregating variants with low individual signals.
• Analyzing gene interactions presents significant computational and statistical challenges.

Purpose of the Study:

• To introduce SiNIMin, a novel computational method designed to identify pairs of interacting genes associated with phenotypes.
• To address genetic heterogeneity and locus interactions in complex disease genetics.
• To leverage biological prior knowledge for guiding gene interaction searches.

Main Methods:

• SiNIMin identifies interacting gene pairs associated with phenotypes under a genetic heterogeneity model.
• The method utilizes protein-protein interaction networks to guide the search for gene interactions.
• Type I error is controlled, and statistical power is enhanced compared to existing methods.

Main Results:

• SiNIMin successfully identifies novel genetic associations for Arabidopsis thaliana phenotypes.
• An adapted version of SiNIMin revealed associations with rare variants in migraine patients.
• The method demonstrates superior statistical power over state-of-the-art approaches.

Conclusions:

• SiNIMin offers a powerful new approach for dissecting the genetic architecture of complex diseases.
• The method effectively integrates genetic heterogeneity and gene-gene interactions.
• SiNIMin has broad applicability in plant and human genetics research.