Ligand Force Matching (LFM)

This repository contains all the code for the Ligand Force Matching (LFM) workflow. See the paper for more details.

Setting up environment

Once you have conda (or mamba, preferred) installed, simply run:

source env.sh

This will create the lfm conda environment.

Reproducing paper metrics

The outputs from running the benchmarking virtual screening campaigns can be found here. Once you've downloaded it, simply run:

python -m analysis.results /path/to/downloaded/folder

Running the LFM workflow on a new target

The data generation + training code is pretty specific to our infrasctuture and we're in the process of cleaning it up + documenting it. It currently involves orchestrating hundreds of instances on Vast.ai to do the data generation and running the training on our SLURM cluster so there are a lot of moving parts.

In the meantime, please reach out if you're interesting in using this for your work!

License

All the code is available under the MIT license.

This repo includes modified code from torch-cubic-spline-grids (BSD license).