SCANG

SCANG implements a dynamic scan-statistic procedure to detect rare-variant association regions in whole-genome sequencing (WGS) studies for mapping associations between rare genetic variants and complex traits or diseases.

Key Features:

• R package implementation: Provided as an R package for analysis of WGS data.
• Dynamic scan-statistic methodology: Uses a scan-statistic approach that employs the p-value from variant set-based tests as the statistic for each moving window.
• No fixed window sizes: Dynamically detects both the size and location of association regions without requiring pre-specification of window sizes.
• Control of genome-wise type I error: Controls the genome-wise type I error rate for robust statistical inference.
• Accounting for linkage disequilibrium: Explicitly accounts for linkage disequilibrium among genetic variants during scanning.
• Tailored to WGS and non-coding regions: Designed to handle the large number of intergenic and intronic non-coding variants present in WGS datasets.
• Computational efficiency: Employs a computationally efficient scanning procedure suitable for large-scale WGS analyses.

Scientific Applications:

• Rare-variant association discovery: Detects regions harboring rare variants associated with complex traits or diseases in WGS studies.
• Genome-wide scanning across phenotypes: Applies to different phenotypes by adapting window size and location to varying genetic-association region dimensions.
• Method validation and real-data analysis: Demonstrated by extensive simulation studies and applied to WGS lipids data from the Atherosclerosis Risk in Communities (ARIC) study.

Methodology:

SCANG applies a dynamic scan-statistic procedure that uses p-values from variant set-based tests as the scan statistic across moving windows, dynamically determines window size and location, accounts for linkage disequilibrium, and controls the genome-wise type I error rate; implemented as an R package.