GenKI 0.1.0

GenKI

GenKI (Gene Knock-out Inference)

A VGAE (Variational Graph Auto-Encoder) based model to learn perturbation using scRNA-seq data.
New! Data has been added.
Paper

Prerequisites

Before installing GenKI, install PyTorch and PyTorch Geometric matching your CUDA version.

Install GenKI with pip (PyPI):

pip install GenKI

Or install from source:

pip install git+https://github.com/yjgeno/GenKI.git

Or clone and install manually:

git clone https://github.com/yjgeno/GenKI.git
cd GenKI
pip install .

Alternatively, use conda to set up the full environment:

conda env create -f environment.yml
conda activate ogenki

Tutorial

Virtual KO experiment:
https://github.com/yjgeno/GenKI/blob/master/notebook/Example.ipynb