BPrimm

BPrimm performs genomewide multiple-loci mapping by applying Bayesian Adaptive Lasso and Iterative Adaptive Lasso approaches to select genetic markers associated with traits in high-dimensional data exhibiting linkage or linkage disequilibrium.

Key Features:

• Bayesian Adaptive Lasso: Uses Bayesian principles to adaptively adjust penalty parameters for refined selection of significant loci.
• Iterative Adaptive Lasso: Iteratively refines weights assigned to markers to improve variable selection and computational efficiency.
• Adaptive weights and iterative updating: Assigns adaptive weights to genetic markers that are iteratively updated to account for linkage and linkage disequilibrium among markers.
• High-dimensional mapping: Targets settings where the number of genetic markers exceeds sample size, enabling analysis of large-scale genomic data.
• Comparative performance: Demonstrated superior efficacy relative to eight existing techniques in simulation studies and real data analyses.
• Computational speed: Iterative Adaptive Lasso achieves substantially faster runtimes than marginal regression and stepwise regression for large datasets.
• Generalizability: Provides a framework applicable to other variable selection problems beyond multiple-loci mapping.

Scientific Applications:

• Genomewide multiple-loci mapping: Identify loci associated with complex traits using multivariate variable selection across the genome.
• Identification of trait-associated markers: Select genetic markers significantly associated with phenotypic traits in high-dimensional datasets.
• Analysis under linkage/linkage disequilibrium: Resolve marker selection in regions with strong linkage or linkage disequilibrium.
• General variable selection: Apply adaptive-lasso-based selection methods to other high-dimensional biological and statistical problems.

Methodology:

Implements Bayesian Adaptive Lasso and Iterative Adaptive Lasso that assign adaptive weights to markers which are iteratively updated; the Bayesian method adapts penalty parameters, the iterative method refines weights for computational efficiency, and performance was evaluated via simulation studies and real data analyses.