GEMMA is a software toolkit for fast application of linear mixed models (LMMs) and related models to genome-wide association studies (GWAS) and large-scale datasets. It addresses the need for efficient statistical analysis of genetic data, offering solutions for population structure correction, heritability estimation, and multi-marker modeling for researchers and bioinformaticians.

How It Works

GEMMA implements univariate and multivariate linear mixed models (LMMs) for association tests, efficiently correcting for population structure and sample non-exchangeability. It also offers a Bayesian sparse linear mixed model (BSLMM) for phenotype prediction and multi-marker modeling. Variance components can be estimated using raw data (HE regression, REML AI) or summary statistics (MQS algorithm), providing flexibility in analysis.

Quick Start & Requirements

• Installation: Precompiled binaries, Docker, Conda, Homebrew, or Guix.
• Prerequisites: None explicitly stated for precompiled binaries. Compilation from source may benefit from specialized C++ compilers and numerical libraries.
• Resources: Docker images available for MacOS, Windows, and Linux (amd64 & arm64).
• Links: INSTALL.md, GEMMA manual, Tutorial

Highlighted Details

• Fast association tests using univariate LMMs.
• Multivariate LMMs for jointly analyzing multiple phenotypes.
• BSLMM for polygenic modeling and phenotype prediction.
• Variance component estimation partitioned by SNP functional categories.

Maintenance & Community

• Development has moved to PanGEMMA (as of Dec 2024).
• Community support via GEMMA Google Group.
• Bug reporting via GitHub issues.
• Contributions encouraged via pull requests.

Licensing & Compatibility

• License: GNU General Public License (GPL).
• Compatibility: GPL is a copyleft license, potentially restricting use in closed-source commercial products without careful consideration of derivative works. Included libraries have LGPL and Boost Software License.

Limitations & Caveats

• Main software development has transitioned to PanGEMMA, suggesting potential for reduced updates or support for this repository.
• The GPL license may impose significant restrictions on commercial use or integration into proprietary software.