dnbaker / minicore-experiments

Experiments for minicore: fast scRNA-seq clustering with various distances

This repository contains experiments for minicore: Fast scRNA-seq clustering with various distances.

Dependencies

1. sklearn
2. numpy
3. scipy
4. minicore v0.3

Installing minicore

To install minicore's Python bindings, clone minicore recursively, change into the Python directory and run python3 setup.py install. Alternatively, a single-command install would be python3 -m pip install git+git://github.com/dnbaker/minicore@main. Please create an issue if you encounter any difficulties.

Because compilation can be somewhat slow, the build is parallelized by default with $OMP_NUM_THREADS threads. Run your command as OMP_NUM_THREADS=number python3 setup.py install to use number threads.

Setting up data

Data can be downloaded in compressed CSR-format here. For simplicity, simply run this download script, which will download and decompress the tarballs necessary. It will prompt you to append MINICORE_DATA to your .bashrc; if you choose not to, you simply need to export export MINICORE_DATA=${PATH} to the path where you downloaded minicore_data. Then you should be able to perform these experiments.