FragGT: Fragment-based Evolutionary Molecule Generation using Gene Types

This directory contains data requires to run frag-gt (Meyers and Brown, 2023), a fragment-based evolutionary algorithm for generating optimal molecules released as part of the guacamol_baselines GitHub repository – https://github.com/BenevolentAI/guacamol_baselines.
Scripts for generating the data are available from github. The compressed data directory contains (A) Processed and filtered SMILES derived from ChEMBL v.33 produced by `download_chembl_smiles` and (B) fragment stores generated by `generate_fragstore` and `filter_fragstore` for both the above file in (A) and the original GuacaMol dataset.
smiles_files and fragstores were derived from molecule data downloaded from ChEMBL (https://www.ebi.ac.uk/chembl).
Liability: We do not represent and/or warrant that no third party rights exist which might prevent the use of the database or that no third party rights would be infringed by said use.
(data updated for frag-gt version 0.0.2)