Consensus queries in ligand-based virtual screening experiments

Abstract

BackgroundIn ligand-based virtual screening experiments, a known active ligand is used in similarity searches to find putative active compounds for the same protein target. When there are several known active molecules, screening using all of them is more powerful than screening using a single ligand. A consensus query can be created by either screening serially with different ligands before merging the obtained similarity scores, or by combining the molecular descriptors (i.e. chemical fingerprints) of those ligands.ResultsWe report on the discriminative power and speed of several consensus methods, on two datasets only made of experimentally verified molecules. The two datasets contain a total of 19 protein targets, 3776 known active and ~ 2 × 106 inactive molecules. Three chemical fingerprints are investigated: MACCS 166 bits, ECFP4 2048 bits and an unfolded version of MOLPRINT2D. Four different consensus policies and five consensus sizes were benchmarked.ConclusionsThe best consensus method is to rank candidate molecules using the maximum score obtained by each candidate molecule versus all known actives. When the number of actives used is small, the same screening performance can be approached by a consensus fingerprint. However, if the computational exploration of the chemical space is limited by speed (i.e. throughput), a consensus fingerprint allows to outperform this consensus of scores.

Chemoinformatics and structural bioinformatics in OCaml

“…Consent [14, 85] (opam package lbvs_consent [86]) performs ligand-based virtual screening using consensus queries. When several active molecules are known, screening with all of them is recommended (instead of using just one).…”

“…The compilation of OCaml programs is fast. For example, the OCaml lines (without comments) of the consent software [14] and its four executables compile from scratch and link in s (resp ) using dune (version 1.6.2) and a single core (resp. up to all cores) of our desktop computer (16 cores, Intel Xeon 2.1GHz, 64GB RAM, Linux Ubuntu 18.04.1 LTS).…”

“…Most of its academic users work in computer science, on compilers and formal methods. But, OCaml is also used in bioinformatics [21–23], structural bioinformatics [24–27], chemoinformatics [14, 28], systems biology [29–32] and ecotoxicology [33].…”

ETCM: an encyclopaedia of traditional Chinese medicine

“…To facilitate mechanistic studies on TCM herbs and formulas, we predicted target genes of TCM ingredients according to the structural and chemical similarity of ingredients with known drugs. Although the prediction method we used has been evaluated as one of the best performed methods in similarity based drug discovery (20,21), there may still be many false positives in the prediction results. Thus, the target prediction results could only serve as a mechanism indication of TCM ingredients, herbs and formulas, and await to be investigated in the future.…”

Fingerprinting CANDO: Increased Accuracy with Structure and Ligand Based Shotgun Drug Repurposing

“…The CANDO platform for shotgun drug repurposing is not dependent on any particular method for determining compound similarity, such as the protein-centric one used in v1. Here, we consider the utility of ligand-based pipelines by constructing two dimensional molecular fingerprints of the 3733 compounds in the CANDO putative drug library using the open-source cheminformatics software RDKit Python API 30 and performing an all-against-all comparison using the Tanimoto coefficient. Once the features of a molecule have been quantized into a vector, the Tanimoto coefficient is a score of how many bits two vectors have in common divided by the number of bits by which they differ, i.e., | A ∩ B | / | A ∪ B |, where A and B represent compounds in binary vector form, and | A | is the length of the vector.…”