QPHAR: quantitative pharmacophore activity relationship: method and validation

Unveiling structural determinants for FXR antagonism in 1,3,4-trisubstituted-Pyrazol amide derivatives: A multi-scale in silico modelling approach

Non-alcoholic fatty liver disease (NAFLD) is a growing global health concern due to its potential to progress into severe liver diseases. Targeting the bile acid receptor FXR has emerged as a promising strategy for managing NAFLD. Building upon our previous research on FXR partial agonism, the present study investigates a series of 1,3,4-trisubstituted-pyrazol amide derivatives as FXR antagonists, aiming to delineate the structural features for antagonism. By means of 2D-QSAR (quantitative structure-activity relationships) modelling techniques, we elucidated the key structural elements responsible for the antagonistic properties of these derivatives. We then employed QPhAR, an open-access software, to identify key molecular features within the compounds that enhance their antagonistic activity. Additionally, 3D-QSAR modelling allowed us to analyse the steric and electrostatic fields of aligned 3D structures, further refining our understanding of structure-activity relationships. Subsequent molecular dynamics simulations provided insights into the binding mode interactions between the compounds and FXR, with varying potencies, confirming and complementing the findings from 2D-QSAR, pharmacophore, and 3D-QSAR modelling. Particularly, our study highlighted the significance of hydrophobic interactions in conferring potent antagonism by the 1,3,4-trisubstituted-pyrazol amide derivatives against FXR. Overall, this work underscores the potential of 1,3,4-trisubstituted-pyrazol amides as FXR antagonists for NAFLD treatment. Notably, our reliance on open-access software fosters reproducibility and broadens the accessibility of our findings.

Ligand-Based Compound Activity Prediction via Few-Shot Learning

Predicting the activities of new compounds against biophysical or phenotypic assays based on the known activities of one or a few existing compounds is a common goal in early stage drug discovery. This problem can be cast as a "few-shot learning" challenge, and prior studies have developed few-shot learning methods to classify compounds as active versus inactive. However, the ability to go beyond classification and rank compounds by expected affinity is more valuable. We describe Few-Shot Compound Activity Prediction (FS-CAP), a novel neural architecture trained on a large bioactivity data set to predict compound activities against an assay outside the training set, based on only the activities of a few known compounds against the same assay. Our model aggregates encodings generated from the known compounds and their activities to capture assay information and uses a separate encoder for the new compound whose activity is to be predicted. The new method provides encouraging results relative to traditional chemical-similarity-based techniques as well as other state-of-the-art few-shot learning methods in tests on a variety of ligand-based drug discovery settings and data sets. The code for FS-CAP is available at https://github.com/Rose-STL-Lab/FS-CAP.

bbSelect - An Open-Source Tool for Performing a 3D Pharmacophore-Driven Diverse Selection of R-Groups

The design of compounds during hit-to-lead often seeks to explore a vector from a core scaffold to form additional interactions with the target protein. A rational approach to this is to probe the region of a protein accessed by a vector with a systematic placement of pharmacophore features in 3D, particularly when bound structures are not available. Herein, we present bbSelect, an open-source tool built to map the placements of pharmacophore features in 3D Euclidean space from a library of R-groups, employing partitioning to drive a diverse and systematic selection to a user-defined size. An evaluation of bbSelect against established methods exemplified the superiority of bbSelect in its ability to perform diverse selections, achieving high levels of pharmacophore feature placement coverage with selection sizes of a fraction of the total set and without the introduction of excess complexity. bbSelect also reports visualizations and rationale to enable users to understand and interrogate results. This provides a tool for the drug discovery community to guide their hit-to-lead activities.

Shaping the Future of Obesity Treatment: In Silico Multi-Modeling of IP6K1 Inhibitors for Obesity and Metabolic Dysfunction

Recent research has uncovered a promising approach to addressing the growing global health concern of obesity and related disorders. The inhibition of inositol hexakisphosphate kinase 1 (IP6K1) has emerged as a potential therapeutic strategy. This study employs multiple ligand-based in silico modeling techniques to investigate the structural requirements for benzisoxazole derivatives as IP6K1 inhibitors. Firstly, we developed linear 2D Quantitative Structure-Activity Relationship (2D-QSAR) models to ensure both their mechanistic interpretability and predictive accuracy. Then, ligand-based pharmacophore modeling was performed to identify the essential features responsible for the compounds' high activity. To gain insights into the 3D requirements for enhanced potency against the IP6K1 enzyme, we employed multiple alignment techniques to set up 3D-QSAR models. Given the absence of an available X-ray crystal structure for IP6K1, a reliable homology model for the enzyme was developed and structurally validated in order to perform structure-based analyses on the selected dataset compounds. Finally, molecular dynamic simulations, using the docked poses of these compounds, provided further insights. Our findings consistently supported the mechanistic interpretations derived from both ligand-based and structure-based analyses. This study offers valuable guidance on the design of novel IP6K1 inhibitors. Importantly, our work exclusively relies on non-commercial software packages, ensuring accessibility for reproducing the reported models.

Synthesis, molecular docking, and in vivo antidiabetic evaluation of new benzylidene-2,4-thiazolidinediones as partial PPAR-γ agonists

Peroxisome proliferator-activated receptor-γ (PPAR-γ) partial agonists or antagonists, also termed as selective PPAR-γ modulators, are more beneficial than full agonists because they can avoid the adverse effects associated with PPAR-γ full agonists, such as weight gain and congestive heart disorders, while retaining the antidiabetic efficiency. In this study, we designed and synthesized new benzylidene-thiazolidine-2,4-diones while keeping the acidic thiazolidinedione (TZD) ring at the center, which is in contrast with the typical pharmacophore of PPAR-γ agonists. Five compounds (5a-e) were designed and synthesized in moderate to good yields and were characterized using spectral techniques. The in vivo antidiabetic efficacy of the synthesized compounds was assessed on streptozotocin-induced diabetic mice using standard protocols, and their effect on weight gain was also studied. Molecular docking and molecular dynamics (MD) simulation studies were performed to investigate the binding interactions of the title compounds with the PPAR-γ receptor and to establish their binding mechanism. Antidiabetic activity results revealed that compounds 5d and 5e possess promising antidiabetic activity comparable with the standard drug rosiglitazone. No compound showed considerable effect on the body weight of animals after 21 days of administration, and the findings showed statistical difference (p < 0.05 to p < 0.0001) among the diabetic control and standard drug rosiglitazone groups. In molecular docking study, compounds 5c and 5d exhibited higher binding energies (- 10.1 and - 10.0 kcal/mol, respectively) than the native ligand, non-thiazolidinedione PPAR-γ partial agonist (nTZDpa) (- 9.8 kcal/mol). MD simulation further authenticated the stability of compound 5c-PPAR-γ complex over the 150 ns duration. The RMSD, RMSF, rGyr, SASA, and binding interactions of compound 5c-PPAR-γ complex were comparable to those of native ligand nTZDpa-PPAR-γ complex, suggesting that the title compounds have the potential to be developed as partial PPAR-γ agonists.

Discovery of novel S6K1 inhibitors by an ensemble-based virtual screening method and molecular dynamics simulation

Ribosomal protein S6 kinase beta-1 (S6K1) is considered a potential target for the treatment of various diseases, such as obesity, type II diabetes, and cancer. Development of novel S6K1 inhibitors is an urgent and important task for the medicinal chemists. In this research, an effective ensemble-based virtual screening method, including common feature pharmacophore model, 3D-QSAR pharmacophore model, naïve Bayes classifier model, and molecular docking, was applied to discover potential S6K1 inhibitors from BioDiversity database with 29,158 compounds. Finally, 7 hits displayed considerable properties and considered as potential inhibitors against S6K1. Further, carefully analyzing the interactions between these 7 hits and key residues in the S6K1 active site, and comparing them with the reference compound PF-4708671, it was found that 2 hits exhibited better binding patterns. In order to further investigate the mechanism of the interactions between 2 hits and S6K1 at simulated physiological conditions, the molecular dynamics simulation was performed. The ΔG_bind energies for S6K1-Hit1 and S6K1-Hit2 were - 111.47 ± 1.29 and - 54.29 ± 1.19 kJ mol^-1, respectively. Furthermore, deep analysis of these results revealed that Hit1 was the most stable complex, which can stably bind to S6K1 active site, interact with all of the key residues, and induce H1, H2, and M-loop regions changes. Therefore, the identified Hit1 may be a promising lead compound for developing new S6K1 inhibitor for various metabolic diseases treatment.

Multi-model in silico characterization of 3-benzamidobenzoic acid derivatives as partial agonists of Farnesoid X receptor in the management of NAFLD

Non-alcoholic fatty liver disease (NAFLD) is a pathological condition which is strongly correlated with fat accumulation in the liver that has become a major health hazard globally. So far, limited treatment options are available for the management of NAFLD and partial agonism of Farnesoid X receptor (FXR) has proven to be one of the most promising strategies for treatment of NAFLD. In present work, a range of validated predictive cheminformatics and molecular modeling studies were performed with a series of 3-benzamidobenzoic acid derivatives in order to recognize their structural requirements for possessing higher potency towards FXR. 2D-QSAR models were able to extract the most significant structural attributes determining the higher activity towards the receptor. Ligand-based pharmacophore model was created with a novel and less-explored open access tool named QPhAR to acquire information regarding important 3D-pharmacophoric features that lead to higher agonistic potential towards the FXR. The alignment of the dataset compounds based on pharmacophore mapping led to 3D-QSAR models that pointed out the most crucial steric and electrostatic influence. Molecular dynamics (MD) simulation performed with the most potent and the least potent derivatives of the current dataset helped us to understand how to link the structural interpretations obtained from 2D-QSAR, 3D-QSAR and pharmacophore models with the involvement of specific amino acid residues in the FXR protein. The current study revealed that hydrogen bond interactions with carboxylate group of the ligands play an important role in the ligand receptor binding but higher stabilization of different helices close to the binding site of FXR (e.g., H5, H6 and H8) through aromatic scaffolds of the ligands should lead to higher activity for these ligands. The present work affords important guidelines towards designing novel FXR partial agonists for new therapeutic options in the management of NAFLD. Moreover, we relied mainly on open-access tools to develop the in-silico models in order to ensure their reproducibility as well as utilization.

Applications of the Novel Quantitative Pharmacophore Activity Relationship Method QPhAR in Virtual Screening and Lead-Optimisation

Pharmacophores are an established concept for the modelling of ligand–receptor interactions based on the abstract representations of stereoelectronic molecular features. They became widely popular as filters for the fast virtual screening of large compound libraries. A lot of effort has been put into the development of sophisticated algorithms and strategies to increase the computational efficiency of the screening process. However, hardly any focus has been put on the development of automated procedures that optimise pharmacophores towards higher discriminatory power, which still has to be done manually by a human expert. In the age of machine learning, the researcher has become the decision-maker at the top level, outsourcing analysis tasks and recurrent work to advanced algorithms and automation workflows. Here, we propose an algorithm for the automated selection of features driving pharmacophore model quality using SAR information extracted from validated QPhAR models. By integrating the developed method into an end-to-end workflow, we present a fully automated method that is able to derive best-quality pharmacophores from a given input dataset. Finally, we show how the QPhAR-generated models can be used to guide the researcher with insights regarding (un-)favourable interactions for compounds of interest.