SiNIMin

SiNIMin detects candidate genomic intervals and interacting gene pairs associated with phenotypic variation by aggregating low-signal variants and mining protein-protein interaction networks to detect genetic heterogeneity.

Key Features:

• Genetic heterogeneity detection: Aggregates low-signal variants across different loci to increase statistical power for associations under genetic heterogeneity.
• Gene-gene interaction discovery: Identifies pairs of interacting genes that jointly show significant association with a phenotype under a heterogeneity model.
• Biological prior integration: Uses protein-protein interaction networks to guide the search and prioritize candidate interactions and intervals.
• Statistical control: Implements procedures that control type I error while assessing interval- and interaction-level associations.
• Performance benchmarking: Demonstrates superior statistical power compared to existing state-of-the-art methods.

Scientific Applications:

• Complex disease genetics: Applied to detect multi-locus contributions and interactions relevant to complex phenotypes.
• Arabidopsis thaliana association mapping: Used to discover novel genetic associations for multiple phenotypes in Arabidopsis thaliana.
• Rare-variant studies of migraine: An adapted variant of SiNIMin has been used to study rare variants associated with migraine in patients.

Methodology:

Network-guided mining of candidate intervals using protein-protein interaction networks; aggregation of low-signal variants into intervals; identification of interacting gene pairs under a genetic heterogeneity model; statistical testing with type I error control.