Home

SUMMIT user guide

SUMMIT estimates heritability and genetic correlation, computes reference LD scores, and fits gene–environment interaction models and polygenic scores.

Start with Installation, try the synthetic example below, then choose an analysis guide. Input files describes the formats needed for your own data. Commands and options lists the commands and shared controls.

python example/prepare_example_inputs.py
bash example/estimate_partitioned_gwldscore.sh
bash example/h2_ldscore.sh

Run these commands from an installed SUMMIT checkout. Generated inputs and results go to example/out/.

Choose an analysis

Analysis	Guide
Genome-wide or windowed LD scores	LD scores
Total or partitioned h² and rg	Heritability and genetic correlation
Many traits or annotation models	Batch analyses
One environment, including heterogeneous residual variance	G×E models
Joint continuous and categorical environments	Multiple environments
Shared genetic responses across traits	Cross-trait response models
Binary traits with case–control ascertainment	Binary traits and PCGC
Context-dependent genetic prediction	Polygenic scores
Python model construction and mathematical details	PGS API, Methods
Research extensions	Contextual Python API

Results and background

• Real-data results: genetic-variance estimates from UK Biobank, with aggregate figures and downloadable plot data.
• Benchmarks: runtime, memory, and reproduction commands.
• Methods: model definitions and estimating equations.
• Troubleshooting: input, installation, and fitting errors.

The main summit command supports LD scores, h²/rg, one-environment G×E, and binary-trait PCGC preparation and inference. Joint generalized G×E estimation is available through Python; summit reference provides planning and inspection. PGS uses summit pgs. Research functions in summit.context and binary cross-trait functions in summit.pcgc.cross are identified in their guides.