iMAP

iMAP maps pleiotropic SNP associations across multiple phenotypic traits using GWAS summary statistics to identify shared genetic architecture and inform joint trait association analysis.

Key Features:

• Joint Modeling of Multiple Traits: Uses GWAS summary statistics to model genome-wide SNP associations across multiple phenotypic traits via a multivariate Gaussian distribution that accounts for phenotypic correlations.
• Mixture Modeling Approach: Employs mixture modeling to infer genome-wide SNP association patterns and capture complex genetic architectures across traits.
• Integration of Functional Annotations: Incorporates large numbers of SNP functional annotations, including tissue-specific annotations, to increase power for association mapping.
• Annotation Selection with Sparsity-Inducing Penalty: Applies a sparsity-inducing penalty to select informative annotations from many candidates.
• Scalable Inference via Efficient Algorithms: Implements an expectation-maximization algorithm based on approximate penalized regression for scalable inference on large GWAS datasets.

Scientific Applications:

• Pleiotropic mapping in GWAS: Enables identification and interpretation of loci that influence multiple complex traits to elucidate genetic relationships and disease etiology.
• Large-scale multi-trait analysis: Has been applied to joint analyses of 48 traits from 31 GWAS consortia using 40 tissue-specific SNP annotations from the Roadmap Project.

Methodology:

Uses GWAS summary statistics with multivariate Gaussian modeling of phenotypic correlations, mixture modeling for SNP association patterns, integration of functional annotations with a sparsity-inducing penalty, and inference via an expectation-maximization algorithm based on approximate penalized regression.