Activity Cliffs As Protein-Related Phenomenon: Investigation Using Machine Learning Against Numerous Protein Kinases

Activity cliffs (ACs) are analogous compounds of significant affinity discrepancies against certain biotarget. We propose that the ACs phenomenon is protein-related and that the propensity of certain target to have ACs can be predicted by some intrinsic protein properties. We pursued this assumption by collecting the crystallographic structures of 84 protein kinases, each of which has numerous reported inhibitors (hundreds). Following data augmentation using synthetic minority oversampling technique (SMOTE), we attempted to correlate the presence/absence of ACs within the ligand pools of collected protein kinases with their corresponding protein properties using genetic algorithm (GA) coupled with variety of machine learners (MLs). Very good GA-ML models were achieved with accuracies of around 75% against external testing set. The models were further validated by Y-scrambling. Shapely additive explanations highlighted the significance of protein rotatable bonds, hydrophobic and acidic residues in relation to the presence of ACs. These results support the hypothesis that ACs are protein-related.

Benchmarking molecular feature attribution methods with activity cliffs

Feature attribution techniques are popular choices within the explainable artificial intelligence toolbox, as they can help elucidate which parts of the provided inputs used by an underlying supervised-learning method are considered relevant for a specific prediction. In the context of molecular design, these approaches typically involve the coloring of molecular graphs, whose presentation to medicinal chemists can be useful for making a decision of which compounds to synthesize or prioritize. The consistency of the highlighted moieties alongside expert background knowledge is expected to contribute to the understanding of machine-learning models in drug design. Quantitative evaluation of such coloring approaches, however, has so far been limited to substructure identification tasks. We here present an approach that is based on maximum common substructure algorithms applied to experimentally-determined activity cliffs. Using the proposed benchmark, we found that molecule coloring approaches in conjunction with classical machine-learning models tend to outperform more modern, deep-learning-based alternatives. However, none of the tested feature attribution methods sufficiently and consistently generalized when confronted with unseen examples.

2-Phenoxy-3-trichloromethylquinoxalines Are Antiplasmodial Derivatives with Activity against the Apicoplast of Plasmodium falciparum

The malaria parasite harbors a relict plastid called the apicoplast. Although not photosynthetic, the apicoplast retains unusual, non-mammalian metabolic pathways that are essential to the parasite, opening up a new perspective for the development of novel antimalarials which display a new mechanism of action. Based on the previous antiplasmodial hit-molecules identified in the 2-trichloromethylquinoxaline series, we report herein a structure–activity relationship (SAR) study at position two of the quinoxaline ring by synthesizing 20 new compounds. The biological evaluation highlighted a hit compound (3i) with a potent PfK1 EC50 value of 0.2 µM and a HepG2 CC50 value of 32 µM (Selectivity index = 160). Nitro-containing (3i) was not genotoxic, both in the Ames test and in vitro comet assay. Activity cliffs were observed when the 2-CCl3 group was replaced, showing that it played a key role in the antiplasmodial activity. Investigation of the mechanism of action showed that 3i presents a drug response by targeting the apicoplast and a quick-killing mechanism acting on another target site.

Structure-based Assessment of Homologous Analogues of Natural products: A computational approach to predict the therapeutic effects of natural products

The present study describes a novel strategy to screen natural products (NPs) for their therapeutic effects with the predicted mechanism of action. The method entitled 'Structure-based Assessment of Homologous Analogues of Natural products-SAHANA' follows the comparison of NPs against prescribed synthetic chemical drugs to deduce activity cliffs and core fragments, based on the molecular properties and 2-dimensional structural similarities. The method was applied to predict the biological effect of the identified NPs as antidiabetic molecules. Selected NPs were assessed for their pharmacokinetic and pharmacodynamics properties. The biological interactions and structural stability of the bound structures were evaluated using molecular docking and molecular dynamics simulations. The study yielded NPs with significant structural similarities to prescribed drugs. Further, their binding interactions stabilized the macromolecular structure. The results envisage a strong indication that the natural products can produce therapeutic effects efficiently if administered individually. The results also encourage using the current screening strategy to identify competent natural product drugs against any disease condition ad libitum.

Prediction of activity cliffs on the basis of images using convolutional neural networks

An activity cliff (AC) is formed by a pair of structurally similar compounds with a large difference in potency. Accordingly, ACs reveal structure–activity relationship (SAR) discontinuity and provide SAR information for compound optimization. Herein, we have investigated the question if ACs could be predicted from image data. Therefore, pairs of structural analogs were extracted from different compound activity classes that formed or did not form ACs. From these compound pairs, consistently formatted images were generated. Image sets were used to train and test convolutional neural network (CNN) models to systematically distinguish between ACs and non-ACs. The CNN models were found to predict ACs with overall high accuracy, as assessed using alternative performance measures, hence establishing proof-of-principle. Moreover, gradient weights from convolutional layers were mapped to test compounds and identified characteristic structural features that contributed to successful predictions. Weight-based feature visualization revealed the ability of CNN models to learn chemistry from images at a high level of resolution and aided in the interpretation of model decisions with intrinsic black box character.