https://github.com/paccmann/paccmann_gp
PyTorch/skopt based implementation of Bayesian optimization with Gaussian processes - build to optimize latent spaces of VAEs to generate molecules with desired properties
https://github.com/paccmann/paccmann_gp
bayesian-optimization chemoinformatics gaussian-processes generative-model machine-learning
Last synced: 11 months ago
JSON representation
PyTorch/skopt based implementation of Bayesian optimization with Gaussian processes - build to optimize latent spaces of VAEs to generate molecules with desired properties
- Host: GitHub
- URL: https://github.com/paccmann/paccmann_gp
- Owner: PaccMann
- License: mit
- Created: 2021-03-25T18:32:43.000Z (almost 5 years ago)
- Default Branch: main
- Last Pushed: 2024-03-23T20:56:53.000Z (almost 2 years ago)
- Last Synced: 2025-03-24T16:53:10.700Z (11 months ago)
- Topics: bayesian-optimization, chemoinformatics, gaussian-processes, generative-model, machine-learning
- Language: Python
- Homepage:
- Size: 48.7 MB
- Stars: 11
- Watchers: 3
- Forks: 5
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://github.com/PaccMann/paccmann_gp/actions/workflows/python-package.yml)
[](https://opensource.org/licenses/MIT)
[](https://doi.org/10.1021/acs.jcim.1c00889)
# paccmann_gp
Bayesian Optimisation with Gaussian Processes for molecular generative models.
## Installation
Create a conda environment:
```sh
conda env create -f conda.yml
```
Activate the environment:
```sh
conda activate paccmann_gp
```
## Example usage
In the `examples` directory is an example script [example.py](./examples/example.py) that makes use of `paccmann_gp` for a combined optimisation for QED, SAscore and affinity to the transcription factor ERG.
```console
(paccmann_gp) $ python examples/example.py -h
usage: example.py [-h]
svae_path affinity_path
optimisation_name
positional arguments:
svae_path Path to downloaded SVAE model.
affinity_path Path to the downloaded affinity prediction model.
optimisation_name Name for the optimisation.
```
The trained SVAE and affinity models can be downloaded from the SELFIESVAE and affinity folders located [here](https://ibm.ent.box.com/v/paccmann-sarscov2/folder/122603752964).
## Citation
If you use this repo in your projects, please temporarily cite the following:
```bib
@article{born2022active,
author = {Born, Jannis and Huynh, Tien and Stroobants, Astrid and Cornell, Wendy D. and Manica, Matteo},
title = {Active Site Sequence Representations of Human Kinases Outperform Full Sequence Representations for Affinity Prediction and Inhibitor Generation: 3D Effects in a 1D Model},
journal = {Journal of Chemical Information and Modeling},
volume = {62},
number = {2},
pages = {240-257},
year = {2022},
doi = {10.1021/acs.jcim.1c00889},
note ={PMID: 34905358},
URL = {https://doi.org/10.1021/acs.jcim.1c00889}
}
```