| Literature DB >> 25250508 |
Nathalie Lagarde1, Jean-François Zagury, Matthieu Montes.
Abstract
The evaluation of virtual ligand screening methods is of major importance to ensure their reliability. Taking into account the agonist/antagonist pharmacological profile should improve the quality of the benchmarking data sets since ligand binding can induce conformational changes in the nuclear receptor structure and such changes may vary according to the agonist/antagonist ligand profile. We indeed found that splitting the agonist and antagonist ligands into two separate data sets for a given nuclear receptor target significantly enhances the quality of the evaluation. The pharmacological profile of the ligand bound in the binding site of the target structure was also found to be an additional critical parameter. We also illustrate that active compound data sets for a given pharmacological activity can be used as a set of experimentally validated decoy ligands for another pharmacological activity to ensure a reliable and challenging evaluation of virtual screening methods.Mesh:
Substances:
Year: 2014 PMID: 25250508 DOI: 10.1021/ci500305c
Source DB: PubMed Journal: J Chem Inf Model ISSN: 1549-9596 Impact factor: 4.956