https://github.com/whitead/synspace
Synthesis generative model
https://github.com/whitead/synspace
Last synced: 3 months ago
JSON representation
Synthesis generative model
- Host: GitHub
- URL: https://github.com/whitead/synspace
- Owner: whitead
- License: mit
- Created: 2022-12-28T00:59:14.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2025-04-24T22:25:19.000Z (5 months ago)
- Last Synced: 2025-07-01T05:47:22.858Z (3 months ago)
- Language: Python
- Size: 19.2 MB
- Stars: 45
- Watchers: 3
- Forks: 4
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- best-of-atomistic-machine-learning - GitHub - 50% open · ⏱️ 24.04.2025): (Generative Models)
README
# synspace
**This is early-stage code that is in progress. It is in flux**
This package generates a local chemical space around a given molecule using retro and forward synthesis rules. The reactions used are the 50 robust medchem reactions proposed by [Hartenfeller et al.](https://pubs.acs.org/doi/10.1021/ci200379p). The retrosynthesis is done by reversing the 50 robust reactions. The purchasable building blocks come from the [Purchasable Mcule supplier building block catalogs](https://mcule.com/database/). All of these things can be customized though.
## Installation
```sh
pip install synspace
```
## Usage
Generate local chemical space given a SMILES string
```py
mols, props = synspace.chemical_space('CCC=O')
```
`props` contains information like the synthesis route for the molecules. Note that all synthesis routes are relative to the given molecule (it is assumed to be synthetically feasible).
## Citation
Cite the ChemCrow paper for now. Haven't really written up much on this package.
```bibtex
@article{bran2023chemcrow,
title={Chemcrow: Augmenting large-language models with chemistry tools},
author={Bran, Andres M and Cox, Sam and Schilter, Oliver and Baldassari, Carlo and White, Andrew D and Schwaller, Philippe},
journal={arXiv preprint arXiv:2304.05376},
year={2023}
}
```
Also, this idea is similar to [Dolfus et al.](https://pubs.acs.org/doi/10.1021/acs.jcim.2c00246)
## NOTICE
This product includes software developed by Pat Walters (MIT Licensed)
https://github.com/PatWalters/useful_rdkit_utils
Copyright (c) 2022 Pat Walters