Improved approach for proteochemometrics modeling: application to organic compound—amine G protein-coupled receptor interactions

Abstract

The model exploited data for the binding of 22 compounds to 31 amine GPCRs, correlating chemical descriptions and cross-descriptions of compounds and receptors to binding affinity using a novel strategy. A highly valid model (q2 = 0.76) was obtained which was further validated by external predictions using data for 10 other entirely independent compounds, yielding the high q2ext = 0.67. Interpretation of the model reveals molecular interactions that govern psychoactive organic amines overall affinity for amine GPCRs, as well as their selectivity for particular amine GPCRs. The new modeling procedure allows us to obtain fully interpretable proteochemometrics models using essentially unlimited number of ligand and protein descriptors.

Negative electrospray ionization mode in mass spectrometry: A new perspective via modeling

“…These plots do not present outliers for neither models (flow and column injection analysis) since there scatter around the straight line is small and explains about 83% of the variation of Y with predictive ability (Q 2 ) > 70%.An additional internal validation was performed applying 20 permutation tests and then comparing them with the original PLS model in terms of goodness of fit (R 2 ) and predictive ability (Q 2 )[29,30]. In order to verify full statistical significance of the original estimates, the limits of intercepts were R 2 intercept <0.3 and Q 2 intercept <0.05[31,32].The effect of each variable (xk) on Y in the model is labelled as VIP (Variable Importance in the Projection). VIP is the sum over all model dimensions of the contributions VIN (variable influence).…”

Kinome-wide interaction modelling using alignment-based and alignment-independent approaches for kinase description and linear and non-linear data analysis techniques

“…Cross-terms were scaled to Pareto variance; the block weight for cross-terms was initially set to 0 and thereafter increased by a regular step size until an optimal PLS model (according to inner loop cross-validation) was obtained. We have earlier shown that this approach exerts no negative influence on the final modelling results [42]. …”

Exploring the origin of phosphodiesterase inhibition via proteochemometric modeling

“…29 Due to the crucial physiological impact resulted upon selective inhibition of different PDE4 subclasses, as well as owing to the fact that there have been no PCM studies regarding the PDE family as yet, therefore, here for the rst time we applied the PCM approach to study the interaction space of PDE4 subfamilies and their inhibitors. In addition, thus far PCM have been successfully applied to investigate protein families such as G protein-coupled receptors, 30,31 proteases, 32,33 kinases, [34][35][36] antibodies, 37 cytochrome P450 38,39 and carbonic anhydrase. 40,41 While the majority of these researches have used sequence based descriptors for describing proteins, whereby the latter has shown a positive impact on molecular interaction eld (MIF) based descriptors called GRid-INdependent Descriptors (GRIND) on modeling and on the signicance of using z scales-GRINDs combinatorial descriptors.…”