README.md

EASIE - Exploration of chemical Analog Space, Implicitly and Explicitly

This repository contains the code described in Computer-aided evaluation and exploration of chemical spaces constrained by reaction pathways

Usage:

1. Set up Conda environment

a. Initialize the Conda environment: conda env create --file environment.yml b. Activate the Conda environment conda activate easie

2. Download the ASKCOS building block database:

a. cd easie/building_blocks b. wget https://github.com/ASKCOS/askcos-data/blob/main/buyables/buyables.json.gz

3a. Run implicit/explicit enumeration from a manually defined route

To count the number of analogs or perform an explicit enumeration for a manually defined route, use count_analogs.py or enumerate_analogs.py, respectively. Manually defined routes should be written into a file as a list of reaction SMILES with double quotation marks, separated by commas. See data/mitapivat_experimental_route.txt as an example input file. The order of the reactions does not matter.

For example calls and usage examples, see exmample_from_manual_route.ipynb.

3b. Run implicit/explicit enumeration from routes returned by ASKCOS

To count the number of analogs or perform an explicit enumeration for routes generated by ASKCOS use count_analogs_askcos_route.py or enumerate_analogs_askcos_route.py, respectively. To generate routes from ASKCOS in the appropariate format, use the script easie/run_askcos/query_askcos.py. Example input files for querying ASKCOS are provided in the easie/run_askcos directory. This will require a separate deployment of ASKCOS. An example output file is provided at data/valsartan_routes.json.

For example calls and usage examples, see exmample_from_ASKCOS_route.ipynb.

How to cite:

Levin, I, Fortunato, ME, Tan, KL, Coley, CW. Computer-aided evaluation and exploration of chemical spaces constrained by reaction pathways. AIChE J. 2023;e18234. doi:10.1002/aic.18234

Code contributors: Itai Levin and Michael Fortunato