-
Notifications
You must be signed in to change notification settings - Fork 0
Installation
Use Linux for the widest feature support. SUMMIT requires Python 3.10 or newer,
a C++17 compiler, OpenMP, and BLAS/LAPACK. Conda installs the numerical libraries
and Python packages listed in environment.yml; install a compiler separately
if one is unavailable on your system.
On macOS, a compiler with OpenMP support is needed. The Linux direct G×E backend is unavailable there.
git clone https://github.com/sriramlab/SUMMIT.git
cd SUMMIT
conda env create -f environment.yml
conda activate summit
python -m pip install .Verify the commands:
summit --help
summit pgs --help
summit reference --helpThe distribution is currently named gwldcore; its Python package and main
command are named summit. A normal install compiles the native modules.
The generalized reference executor currently requires a Linux build linked to pthread BLIS. The standard OpenBLAS installation supports the other analysis paths, but does not satisfy this executor's build check.
For a site-specific BLIS build, configure GXELDCORE_USE_PRIVATE_BLIS and the
archive/include locations in CMakeLists.txt. The accompanying source metadata
is supplied by the person building that library. This setup is needed only for
the direct generalized reference executor; it is not an extra input to an analysis.
From an activated environment:
python -m pip install --no-build-isolation -e .Reinstall after changing native code or console entry points, using the same
native build configuration. python -m summit also runs the command-line
interface.
See Troubleshooting for compiler and library errors.
Start here
Analyses
- LD scores
- h² and rg
- Batch analyses
- G×E models
- Multiple environments
- Cross-trait response models
- Binary traits and PCGC
- Polygenic scores
Results and reference