Here, we report on the development of a novel methodology to aid in design of Hsp90 inhibitors, using molecular docking combined with artificial neural network (ANN) modelling. Inhibitors are first docked into the ATPase site of the Human Hsp90α crystal structures and the thermodynamic properties of the complexes together with various physical-chemical properties of the ligands are used as input to train a simple feed-forward, back propagation ANN, to predict the inhibitors' pIC 50s. For an objective test set of 60 known Hsp90 inhibitors for which there are no crystallographic data available, the trained ANN is shown to give pIC 50s accurate to within ±1 log unit, and the predictions are sufficiently good as to allow the majority of the inhibitors to be ranked correctly according to their potency.
Prediction of inhibitory activities of Hsp90 inhibitors / P. Swuec, D.J. Barlow. - In: BIOORGANIC & MEDICINAL CHEMISTRY. - ISSN 0968-0896. - 20:1(2012 Jan 01), pp. 408-414.
|Titolo:||Prediction of inhibitory activities of Hsp90 inhibitors|
SWUEC, PAOLO (Primo)
|Parole Chiave:||Docking; Enzyme inhibition; Heat shock proteins; Neural networks; QSAR; Binding Sites; Computer Simulation; HSP90 Heat-Shock Proteins; Humans; Ligands; Protein Structure, Tertiary; Quantitative Structure-Activity Relationship; Software; Thermodynamics; Biochemistry; Molecular Medicine; Molecular Biology; 3003; Drug Discovery3003 Pharmaceutical Science; Clinical Biochemistry; Organic Chemistry|
|Settore Scientifico Disciplinare:||Settore BIO/10 - Biochimica|
|Data di pubblicazione:||1-gen-2012|
|Digital Object Identifier (DOI):||http://dx.doi.org/10.1016/j.bmc.2011.10.069|
|Appare nelle tipologie:||01 - Articolo su periodico|