The use of docking-based comparative intermolecular contacts analysis to identify optimal docking conditions within glucokinase and to discover of new GK activators

Abstract

Glucokinase (GK) is involved in normal glucose homeostasis and therefore it is a valid target for drug design and discovery efforts. GK activators (GKAs) have excellent potential as treatments of hyperglycemia and diabetes. The combined recent interest in GKAs, together with docking limitations and shortages of docking validation methods prompted us to use our new 3D-QSAR analysis, namely, docking-based comparative intermolecular contacts analysis (dbCICA), to validate docking configurations performed on a group of GKAs within GK binding site. dbCICA assesses the consistency of docking by assessing the correlation between ligands' affinities and their contacts with binding site spots. Optimal dbCICA models were validated by receiver operating characteristic curve analysis and comparative molecular field analysis. dbCICA models were also converted into valid pharmacophores that were used as search queries to mine 3D structural databases for new GKAs. The search yielded several potent bioactivators that experimentally increased GK bioactivity up to 7.5-folds at 10 μM.

Unsupervised pharmacophore modeling combined with QSAR analyses revealed novel low micromolar SIRT2 inhibitors

“…Multiple linear regression analysis assumes the existence of linear correlation between molecular descriptors and corresponding bioactivities . The kNN, on the other hand, is a non‐linear nonparametric method that predicts a ligand's bioactivity as distance weighted average of the bioactivities of its k‐nearest neighbors.…”

“…However, 2 separate methodologies were implemented to evaluate Multiple linear regression analysis assumes the existence of linear correlation between molecular descriptors and corresponding bioactivities. [12][13][14][15][16][17] The kNN, on the other hand, is a non-linear nonparametric method that predicts a ligand's bioactivity as distance weighted average of the bioactivities of its k-nearest neighbors. The neighborhood is defined based on certain selected descriptor(s).…”

“…where log(IC 50 )/HA is ligand efficiency (HA is the number of ligands' heavy atoms), r 2 77 is the correlation coefficient against 77 training compounds, r 2 LOO is the leave-one-out cross-validation correlation coefficient, and r 2 PRESS is the predictive r 2 determined for 19 randomly selected test compounds (1,5,11,14,17,23,32,42,45,50,51,64,67,74,78,128,138,139, and 141, as inTable S1 under supporting information).r 2 m and Δr 2 m are the average and delta r 2 m values calculated for training compounds, respectively. These are recently developed metrics that extensively test the internal and external predictive capacities of a QSAR model through establishing the proximity between predicted and observed response data among training compounds.…”

“…The pharmacophore model(s) can be used as 3D search query(ies) to mine 3D libraries for new SIRT2 inhibitors, while the associated QSAR model(s) can be used to predict the bioactivities of captured hits to prioritize them for in vitro evaluation. Similar approach was used towards the discovery of new mTOR inhibitors, PI3Kγ inhibitors, glucokinase activators, urokinase plasminogen inhibitors, acetylcholinesterase inhibitors, and PPAR‐γ ligands …”

Beskrywing, modellering en dok-studies van Plasmodium falciparum kinase PfCDPK4

“…Die bindingposisie is afgelei van die reseptor (PfCDPK4) se holruimtes deur middel van 'n uitvee-algoritme (Kang et al, 2017;Taha et al, 2014;Venkatachalam et al, 2003).…”