gsuite
Modular computational libraries for statistical genetics
gsuite is an ecosystem for statistical genetics, bringing together standalone R packages, reusable computational libraries and analysis documentation. The public packages cover genomic workflows; mixed-model examples use the current development integration.
What can gsuite do?
- Fit pedigree mixed models: specify effects and covariance components, estimate variance parameters and obtain predictions.
- Work with genotype data: prepare resources, filter markers and construct sparse linkage-disequilibrium matrices.
- Analyse GWAS summaries: estimate heritability and genetic correlation, fit covariance models, and perform Mendelian randomisation.
- Fit Bayesian models: analyse traits or independent populations jointly, learn parameters and annotation effects, and summarize genetic variation.
- Map and predict: run association scans, fine-map regions, assess colocalisation and reuse estimated effects for prediction.
Start with the capability overview for current R access and limits. The mixed-model catalog records specific LMM combinations and their evidence.
Explore gsuite
Tutorials
Package-focused and connected analysis paths, from genotype preparation to prediction and pathways.
Guides
Method details, inputs and data preparation.
Software validation
Focused checks of gsuite implementations using simulated truth and numerical references.
Architecture and libraries
Understand the components, their responsibilities and how they connect.
Capabilities and R workflows
Find an analysis, its R interface and the relevant scope notes.
Benchmarks
Explore numerical comparisons and runtime evidence, including DMU.
Software availability
The seven standalone genomic R packages are available as public source packages. A root gsuite R package release is not currently planned. R packaging for greml and gsolve is deferred; separate or combined packaging is undecided. See software availability and current status for installation requirements and qualification limits.
The companion packages gsim for simulation and gact for genomic data integration are public and maintain their own documentation.
For broader thoughts on working with AI, see the presentations.