Enhanced Thompson Sampling for Virtual Screening

This repo accompanies the preprint "Enhanced Thompson Sampling by Roulette Wheel Selection for Screening Ultra-Large Combinatorial Libraries".

Setting up the environment for running Enhanced Thompson Sampling

Create a new conda environment: conda create -n ts_test python=3.11

Activate your environment and install the rest of the requirements: conda activate ts_test pip install -r requirements.txt

Reproduce Figure 3 in the preprint

python ./src/ts_paper.py quinazoline_fp_sim.json

Code in src is outdated and will not be maintained

Use multiprocessing from src_multiprocess

python ./src_multiprocess/ts_main.py input.json

Note that there is a multiprocessing overhead. If time_for_scoring_single_compound * min_cpds_per_core > 0.1 s, there could be a gain from multiprocessing.

Parameters

Required params:

• evaluator_arg: The argument to be passed to the instantiation of the evaluator (e.g. smiles string of the query molecule, filepath to the mol file, etc.) See the relevant Evaluator for more info.
• evaluator_class_name: Required. The scoring function to use. Must be "FPEvaluator" for 2D similarity. To use your own scoring function, implement a subclass of the Evaluator baseclass in evaluators.py.
• reaction_smarts: Required. The SMARTS string for the reaction.
• reagent_file_list: Required. List of filepaths - one for each component of the reaction. Each file should contain the smiles strings of valid reagents for that component.

Optional params:

• percent_of_libray: Optional. Default 0.1%. Percent of library to be screened. If 0.1% to be screened, set as 0.001 instead of 0.1 (a bit confusing).
• num_warmup_trials: Optional. Default 3.
• min_cpds_per_core: Optional. Default 50. The minimum number of compounds collected for scoring per core per iteration.
• scaling: Optional. Default 1. +1 if higher score is preferred; -1 otherwise.
• stop: Optional. Default 6000. Stop searching when without sampling a new compound for a specified number of consecutive attempts.
• results_filename: Optional. Default "./results.csv". Name of the file to output results to.
• log_filename: Optional. Log filename to save logs to. If not set, logging will be printed to stdout.
• hide_progress: Optional. Defaut true. false otherwise.
• nprocesses: Optional. Default to use all CPU cores.