RegionScan: a comprehensive R package for region-level genome-wide association testing with integration and visualization of multiple-variant and single-variant hypothesis testing

TL;DR

RegionScan is an R package that enables genome-wide association testing at the region level, integrating and visualizing multiple and single-variant hypothesis tests.

Contribution

RegionScan introduces a comprehensive framework for region-level genome-wide association testing with multiple-variant and single-variant methods.

Findings

RegionScan implements three state-of-the-art region-level tests for association detection.

The package supports analysis of multi-allelic variants and unbalanced binary phenotypes.

RegionScan provides detailed outputs and utilities for result comparison and visualization.

Abstract

RegionScan is designed for scalable genome-wide association testing of both multiple-variant and single-variant region-level statistics, with visualization of the results. For detection of association under various regional architectures, it implements three classes of state-of-the-art region-level tests, including multiple-variant linear/logistic regression (with and without dimension reduction), a variance-component score test, and region-level minP tests. RegionScan also supports the analysis of multi-allelic variants and unbalanced binary phenotypes and is compatible with widely used variant call format (VCF) files for both genotyped and imputed variants. Association testing leverages linkage disequilibrium (LD) structure in pre-defined regions, for example, LD-adaptive regions obtained by genomic partitioning, and accommodates parallel processing to improve computational and memory efficiency. Detailed outputs (with allele frequencies, variant-LD bin assignment, single/joint variant effect estimates and region-level results) and utility functions are provided to assist comparison, visualization, and interpretation of results. Thus, RegionScan analysis offers valuable insights into region-level genetic architecture, which supports a wide range of potential applications.

RegionScan is freely available for download on GitHub (https://github.com/brossardMyriam/RegionScan).

Figures