SAR and QSAR in Environmental Research最新文献

Unc-51 like kinase 1 (ULK1), a key regulator of autophagy initiation, is a novel target for anticancer drug design. In this work, we collected 846 ULK1 inhibitors with IC₅₀ values from 30 references. Based on ECFP_4, MACCS fingerprints, and Mordred descriptors, we established a list of classification models by using Support Vector Machine (SVM), Random Forest (RF), extreme Gradient Boosting (XGBoost) and Deep Neural Networks (DNN). Additionally, several Fingerprint and Graph Neural Network (FP-GNN) models were also constructed using mixed molecular fingerprints and molecular graph. A total of 39 classification models were developed. Model_1D_1, an ECFP4-based DNN model, performed the best, achieving accuracies over 95% and Matthews correlation coefficient (MCC) over 0.9 on both validation and test sets. The applicability domain calculated by weighted Euclidean distance indicated that Model_1D_1 could reliably predict the activity for over 84% compounds in both training and test sets. We conducted structure-activity relationship (SAR) analysis through K-means and SHAP. The dataset's molecular structures were classified into 7 subsets by K-means clustering. We identified three high-activity subsets sharing a common scaffold, 2-amino-4-(2-thienyl)-5-(trifluoromethyl)pyrimidine. SHAP analysis highlighted critical molecular fragments influencing activity, enhancing our understanding of model predictions and providing a theoretical basis for optimizing ULK1 inhibitors.

Plasmodium falciparum dihydroorotate dehydrogenase (PfDHODH) is a well-established target for developing novel antimalarial agents. Novel imidazo[1,2-a]pyrimidines were designed as PfDHODH inhibitors using a fragment-based drug design (FADD) approach. A library of active molecules targeting PfDHODH was analysed to generate fragments using the RDKit BRICS module. These fragments were screened by docking them into the active site of the PfDHODH enzyme. Among them, the lead fragment, fragment-11, demonstrated a significant binding affinity of -6.895 kcal/mol. This fragment was optimized using a fragment-growing approach via the FragGrow webserver. From the 471 generated molecules, two showed binding scores of -7.9 and -7.0 kcal/mol. These molecules were further optimized, resulting in a lead molecule with a binding score of -11.3 kcal/mol. Based on the results from the FragGrow webserver, 216 novel imidazo[1,2-a]pyrimidines were designed using the scaffold-hopping approach. The ADMET properties of these compounds revealing that all the designed compounds exhibited drug-like properties. Docking studies indicated that compounds 28d, 46d, and 49d had strong binding affinities, with 28d showing the highest score of -10.41 kcal/mol. Furthermore, molecular dynamics (MD) simulations of 28d demonstrated good stability in the enzyme-ligand complex. This comprehensive in silico study suggests that imidazo[1,2-a]pyrimidines can serve as potent PfDHODH inhibitors.

Leishmaniasis, caused by Leishmania spp. remains a major global health concern due to drug resistance, toxicity, non-specificity, and prolonged treatments. Addressing the need for new therapeutics, we investigated a range of bioactive compounds, including chalcones, pyrimidines, quinolines, azoles, sulphonamides, flavonoids, and quinazoline derivatives, targeting Leishmania donovani amastigotes. Key molecular descriptors influencing anti-leishmanial activity were identified using LASSO and multiple linear regression (MLR), yielding robust QSAR models (r² > 0.84) validated through rigorous statistical analysis. Virtual screening and scaffold-hopping strategies led to the design of 12 novel compounds, among which six; mainly benzothiazole and benzoxazole derivatives exhibited clear predicted pIC₅₀ values and promising ADMET profiles. Quinoline-based compounds showed moderate activity, consistent with prior experimental data. Structural analysis revealed the significance of quinoline rings linked to thiazole or benzoxazole moieties, with modifications like alkyl halides and methyl groups enhancing bioactivity. Further molecular docking against Leishmania donovani N-myristoyltransferase (Ld-NMT) and sterol 14-α demethylase CYP51 demonstrated strong binding affinities with compounds N8, N9, and N11. Structure-based similarity searches using ChEMBL confirmed selective bioactivity and low predicted cytotoxicity, supporting minimal off-target interactions. These findings present a computationally guided framework for developing effective, targeted anti-leishmanial agents.

Thymidine phosphorylase (TP) is a key enzyme involved in angiogenesis, tumour growth and closely linked to cancer progression and metastasis. This study represents the first comprehensive 3D-QSAR pharmacophore-based approach to identifying potential 1,3,4-oxadiazole derivatives as targeted TPIs for anticancer therapy. A dataset of 76 analogues with an experimental IC₅₀ values was used to develop pharmacophore models. The BEST conformation method identified an optimal model (Hypo 2), featuring HBA, HBD and RA as key activity determinants with strong statistical validation (r² = 0.69, ΔCost = 77.41, Q² = 0.68 and MAE = 0.23). A virtual screening of 12,353 drug-like 1,3,4-oxadiazole compounds from PubChem and ChEMBL yielded 329 potential TPIs (IC₅₀ < 10 μM). MD Docking using CDOCKER (PDB ID: 1UOU) identified the top hits interacting with critical TP residues (Thr151, Gly152, Lys221, Ser217, Thr118). ADMET analysis confirmed their drug-likeness with no significant toxicity. ChEMBL2058305 exhibited the highest binding stability (-85.508 kcal/mol), the lowest HOMO-LUMO gap (0.066 ha), and superior TP affinity, highlighting its potential as a promising TP inhibitor for anticancer therapy. This first report with integration of pharmacophore modelling, virtual screening, MD Docking, ADMET, MMGBSA and DFT will be beneficial for the discovery of novel TPIs.

This study integrates a robust cheminformatics approach (including chemical space exploration, Bayesian model-based fingerprint analysis, and cluster-driven molecular profiling) to reveal the key structural features influencing peroxisome proliferator activated receptor-gamma (PPARγ) modulatory activity. The Bayesian classification model effectively differentiates between the beneficial and adverse structural characteristics of PPARγ modulators. Structural motifs such as substituted benzylamine, phenoxyphenyl groups, 5-phenyl-1,3-thiazolidine scaffolds, and indole rings have been identified as positive contributors, enhancing PPARγ activity. Conversely, features like substituted tertiary amines and sulphonamide groups were found to have detrimental effects, suggesting that these should be deprioritized in the design of future PPARγ modulators. Additionally, molecular clustering provided a means to categorize structurally similar compounds, facilitating scaffold analysis, diversity calculation, and lead optimization for PPARγ modulators. To extend these findings to the broader scientific community, we have developed an open-access online tool, 'Fasda_v1.0', (https://fasdav1web.streamlit.app/) designed for cluster-driven molecular profiling of any dataset, enabling further exploration and application of these methods. This study offers valuable guidance for designing and developing novel therapeutics targeting PPARγ, thereby contributing to advancements in treating type 2 diabetes mellitus and related diseases.

Milk is one of the primary sources of food. Pesticides and veterinary drugs are reaching directly or indirectly (pesticides containing grass or other cattle foods) into the milk of the cattle, which are serious health concerns to the animals, infants, babies, and humans. So, in-silico approaches like QSPR, read-across, etc., are used as an alternative (reduce time, cost, complex analytical process) for calculating retention time (RT). The present work involves the development of the first multiple PLS-based QSAR models for the estimation of RT of pesticides, veterinary drugs, and related chemical hazards in milk by strictly obeying the OECD principles. Based on the results, the quality of the models is good enough. In the current work, it was observed that lipophilicity, binding property, rotatable bonds, and reactivity are responsible for high RT while hydrophilicity, the presence of primary amines, aqueous solubility, and branching reduce the RT of the compounds. The established models were utilized to screen the PPDB database to justify its real-world application. The present study will be vital in the food-informatics area for the RT data-gap filling and identification of hazardous chemicals in milk. Thus, it will be helpful to maintain a healthier, safer, and eco-friendly ecosystem.

The trade-off between applicability domain size and prediction accuracy is a well-known phenomenon in QSAR. We have developed a modelling approach where multiple models with different applicability domain sizes and with different prediction accuracy are selected instead of a single best model. This approach is implemented in DanishQSAR, a new software for binary classification QSAR modelling, integrating descriptor calculation, descriptor selection, model development, validation and application. The various methods and options available in the software are automatically tested and efficiently combined during model development using a version of cross-validation-based grid search and post-hoc ensemble modelling. The resulting large and diverse pool of model candidates is then analysed to generate three hierarchies of models, optimized for sensitivity, specificity or balanced accuracy, respectively, for minimum to maximum coverage levels. When predicting a query compound, the system provides predictions from all models in the three hierarchies, at all coverage levels with user-defined steps, together with the individual model predictivity performances, producing a prediction profile rather than one prediction from a single model. Twenty data sets from the Danish (Q)SAR Database (https://qsar.food.dtu.dk) are used to demonstrate the performance. The developed binary classification models are highly accurate by cross- and external validation.

Triple Negative Breast Cancer (TNBC) is the most aggressive type of breast cancer unveiling negative expression on oestrogen receptors, progesterone receptors, and HER2. The anomalous activation of signalling pathways and specific types of mutations characterize the progression of TNBC. Protein-protein interaction in the tumour microenvironment plays a crucial role in tumour aggressiveness. Disrupting the signalling pathways that promote cell progression, migration, and survival opens up a promising avenue for targeting the aggressive form of TNBC. The present study emphasizes the molecular interaction mechanism driving the aggressive and recalcitrant TNBC between BBOX1-IP3R3. The BBOX1-IP3R3 complex destabilization was accomplished using compounds obtained from various databases through virtual screening, molecular, and essential dynamics. The interaction study revealed that the four hits bound at the interface and facilitated better binding affinity with the highest docking score and optimal binding free energy. In addition, the molecular dynamics simulation, PCA/FEL, and MM/PBSA analysis conclusively evaluate the binding potential of the compounds and unequivocally stabilize specific conformations or deception of the complexes in high-energy states. Thus, the identified compounds lead to the disruption of BBOX1-IP3R3 interaction, which aids in the therapeutic option of TNBC.

Protoporphyrinogen IX oxidase (PPO) inhibition is a critical strategy for weed control in crop production. This study employed a computational approach integrating QSAR modelling, docking studies, and molecular dynamics to investigate the inhibitory activities of benzothiazole- and phthalimide-derived compounds against PPO. The MIA-QSAR method modelled pKi values for 52 compounds, complemented by docking and molecular dynamics to analyse ligand-enzyme interactions and identify potential agrochemical candidates. QSAR analysis yielded predictive models with r² = 0.77, q² = 0.55, and r² = 0.74. MIA plots guided the design of 12 derivatives, 5 of which showed promising pKi values (7.31-8.69). Docking and molecular dynamics revealed strong binding affinity and stability for these candidates. The presence of fluorine substituents and C=O and C=S bonds in tetrahydroisoindole moieties enhanced biological activity, leading to the proposition of effective PPO inhibitors. Synthetic routes for the top candidates were outlined for future development, aiming to improve agrochemical efficacy and address resistance issues in crop protection.

Mycobacterium tuberculosis remains one of the leading causes of death from a single infectious agent, posing a major global health challenge. The rise of drug-resistant strains has intensified the need for novel therapeutic agents. Pretomanid and delamanid, two recently developed antitubercular drugs, are bicyclic nitroimidazoles that act as prodrugs, requiring activation by specific mycobacterial enzymes. However, the precise molecular targets of their active metabolites are not fully explained. Recent studies have identified DprE2, an essential enzyme in the biosynthesis of decaprenylphosphoryl-β-D-arabinofuranose (DPA) and arabinogalactan, as a potential target of delamanid. In this study, we applied structure-based pharmacophore modelling to identify potential inhibitors targeting DprE2. High-throughput virtual screening, followed by molecular docking, was used to evaluate binding affinities. ADMET predictions were incorporated to assess drug likeness and pharmacokinetic profiles. Nine promising hits were shortlisted, and their binding stability was further evaluated using 250 ns molecular dynamics simulations. Binding free energy calculations using the MM-GBSA method were then applied to refine the selection, identifying five potent lead molecules. These candidates show strong potential for further development as DprE2 inhibitors, offering a new path in the fight against drug-resistant tuberculosis.