An Integrated Machine Learning Model To Spot Peptide Binding Pockets in 3D Protein Screening

The prediction of peptide-protein binding sites is of utmost importance to tackle the onset of severe neurodegenerative diseases and cancer. In this work, we detail a novel machine learning model based on Linear Discriminant Analysis (LDA) demonstrating to be highly predictive in detecting the putative protein binding regions of small peptides. Starting from 439 high-quality pockets derived from peptide-protein crystallographic complexes, three sets of well-established peptide-binding regions were first selected through a Partitioning Around Medoids (PAM) clustering algorithm based on morphological and energetic 3D GRID-MIF molecular descriptors. Next, the best combination between all the putative interacting peptide pockets and related GRID-MIF scores was automatically explored by using the LDA-based protocol implemented in BioGPS. This approach proved successful to recognize the actual interacting peptide regions (that is, AUC = 0.86 and partial ROC enrichment at 5% of 0.48) from all the other pockets of the protein. Validated on two external collections sets, including 445 and 347 crystallographic peptide-protein complexes, our LDA-based model could be effective to further run peptide-protein virtual screening campaigns.

Identification of Novel Influenza Polymerase PB2 Inhibitors Using a Cascade Docking Virtual Screening Approach

To discover novel inhibitors that target the influenza polymerase basic protein 2 (PB2) cap-binding domain (CBD), commercial ChemBridge compound libraries containing 384,796 compounds were screened using a cascade docking of LibDock-LigandFit-GOLD, and 60 compounds were selected for testing with cytopathic effect (CPE) inhibition assays and surface plasmon resonance (SPR) assay. Ten compounds were identified to rescue cells from H1N1 virus-mediated death at non-cytotoxic concentrations with EC₅₀ values ranging from 0.30 to 67.65 μM and could bind to the PB2 CBD of H1N1 with Kd values ranging from 0.21 to 6.77 μM. Among these, four compounds (11D4, 12C5, 21A5, and 21B1) showed inhibition of a broad spectrum of influenza virus strains, including oseltamivir-resistant ones, the PR/8-R292K mutant (H1N1, recombinant oseltamivir-resistant strain), the PR/8-I38T mutant (H1N1, recombinant baloxavir-resistant strain), and the influenza B/Lee/40 virus strain. These compounds have novel chemical scaffolds and relatively small molecular weights and are suitable for optimization as lead compounds. Based on sequence and structure comparisons of PB2 CBDs of various influenza virus subtypes, we propose that the Phe323/Gln325, Asn429/Ser431, and Arg355/Gly357 mutations, particularly the Arg355/Gly357 mutation, have a marked impact on the selectivities of PB2 CBD-targeted inhibitors of influenza A and influenza B.

Classification of Staphylococcus Aureus FabI Inhibitors by Machine Learning Techniques

Enoyl-acyl carrier protein reductase (FabI) is a key enzyme in the fatty acid metabolism of gram-positive bacteria and is considered a potential target for new antibacterial drugs development. Indeed, triclosan is a widely employed antibacterial and AFN-1252 is currently under phase-II clinical trials, both are known as FabI inhibitors. Nowadays, there is an urgent need for new drug discovery due to increasing antibacterial resistance. In the present study, classification models using machine learning techniques were generated to distinguish SaFabI inhibitors from non-inhibitors successfully (e.g., Mathews correlation coefficient values equal to 0.837 and 0.789 calculated with internal and external validations). The interpretation of a selected model indicates that larger compounds, number of N atoms and the distance between central amide and naphthyridinone ring are important to biological activity, corroborating previous studies. Therefore, these obtained information and generated models can be useful for design/discovery of novel bioactive ligands as potential antibacterial agents.

3D QSAR studies, molecular docking and ADMET evaluation, using thiazolidine derivatives as template to obtain new inhibitors of PIM1 kinase

Proviral Integration site for Moloney murine leukemia virus-1 (PIM1) belongs to the serine/threonine kinase family of Ca²⁺-calmodulin-dependent protein kinase (CAMK) group, which is involved in cell survival and proliferation as well as a number of other signal transduction pathways. Thus, PIM1 is regarded as a promising target for treatment of cancers. In the present paper, a three-dimensional quantitative structure activity relationship (3D-QSAR) and molecular docking were performed to investigate the binding between PIM1 and thiazolidine inhibitors in order to design potent inhibitors. The comparative molecular similarity indices analysis (CoMSIA) was developed using twenty-six molecules having pIC₅₀ ranging from 8.854 to 6.011 (IC₅₀ in nM). The best CoMSIA model gave significant statistical quality. The determination coefficient (R²) and Leave-One-Out cross-validation coefficient (Q²) are 0.85 and 0.58, respectively. Furthermore, the predictive ability of this model was evaluated by external validation((n = 11, R²_test = 0.72, and MAE = 0.170 log units). The graphical contour maps could provide structural features to improve inhibitory activity. Furthermore, a good consistency between contour maps and molecular docking strongly demonstrates that the molecular modeling is reliable. Based on these satisfactory results, we designed several new potent PIM1 inhibitors and their inhibitory activities were predicted by the molecular models. Additionally, those newly designed inhibitors, showed promising results in the preliminary in silico ADMET evaluations, compared to the best inhibitor from the studied dataset. The results expand our understanding of thiazolidines as inhibitors of PIM1 and could be of great help in lead optimization for early drug discovery of highly potent inhibitors.

Discrete Fourier Transform-Based Multivariate Image Analysis: Application to Modeling of Aromatase Inhibitory Activity

We recently generalized the formerly alignment-dependent multivariate image analysis applied to quantitative structure-activity relationships (MIA-QSAR) method through the application of the discrete Fourier transform (DFT), allowing for its application to noncongruent and structurally diverse chemical compound data sets. Here we report the first practical application of this method in the screening of molecular entities of therapeutic interest, with human aromatase inhibitory activity as the case study. We developed an ensemble classification model based on the two-dimensional (2D) DFT MIA-QSAR descriptors, with which we screened the NCI Diversity Set V (1593 compounds) and obtained 34 chemical compounds with possible aromatase inhibitory activity. These compounds were docked into the aromatase active site, and the 10 most promising compounds were selected for in vitro experimental validation. Of these compounds, 7419 (nonsteroidal) and 89 201 (steroidal) demonstrated satisfactory antiproliferative and aromatase inhibitory activities. The obtained results suggest that the 2D-DFT MIA-QSAR method may be useful in ligand-based virtual screening of new molecular entities of therapeutic utility.