lbvs-environment

Ligand-based virtual screening (LBVS) environment (LBVS-environment) is designed to support researchers in task of development and benchmarking of chemical representations. The process of LBVS is divided into several phases:

• data preparation
• screening
• evaluation.

This environment consists of a set of scripts enabling easy implementation of the above mentioned steps.

Requirements and dependencies

• RDKit
• Python 3.5

The download_data.py can be executed also with Python 2.6.

Usage examples

Ligand-based virtual screening

Benchmark performance of tt_tanimoto on the 01 collection.

cd scripts
# Download datasets - Choose download all, when done exit the script.
python download_data.py
# Perform screening.
python screen_localhost.py -c 01 -m tt/tt_tanimoto
# Evaluate data.
python evaluate_screening.py
python export_results.py

Molecule export

Export molecules (test, train, validation set) used in collection 01.

cd scripts
# Download datasets - Choose download all, when done exit the script.
python download_data.py
python export_molecules.py -c 01 -o ../molecule-export --flat --info