Search Results (478)

Bioactive peptides have health benefits, but the intense bitterness associated with their hydrolysis severely restricts their industrial applications. This paper systematically constructs a collaborative theoretical framework that integrates intelligent prediction, targeted debittering, and multi-dimensional evaluation. Firstly, it reviews the core applications of deep learning (such as quantitative structure–activity relationship (QSAR) and graph convolutional network (GCN)) combined with molecular docking technology in the high-throughput identification of bitter peptides and the analysis of target receptor interaction mechanisms. Secondly, it discusses how artificial intelligence and computational simulation can improve the efficiency of traditional debittering processes, emphasizing the advantages of multifunctional composite wall materials in the targeted encapsulation and delivery of bitter peptides, as well as the metabolic regulatory mechanisms behind controlling microbial fermentation for the debittering of specific peptide substrates. Finally, to provide a high-fidelity data closed loop for artificial intelligence (AI) models, a three-dimensional cross-validation system integrating standardized quantitative sensory evaluation and biomimetic electronic tongues was established. Future research should focus on developing large models for flavor generation to drive the green and targeted creation of low-bitterness and highly active peptides. Full article

Quantitative structure–activity relationship and quantitative structure–property relationship (QSAR/QSPR)-based molecular toxicity prediction provides an in silico strategy for prioritizing environmental contaminants when longer-duration bioassay data are sparse. However, many Simplified Molecular-Input Line-Entry System (SMILES)-based machine learning models treat exposure duration as an unconstrained numerical covariate and provide limited information on whether predictions are supported by the observed temporal domain. Here, we evaluated an applicability-domain-aware chemoinformatics framework that combines transformer-derived molecular representations with toxicokinetic-informed temporal encoding and evidential uncertainty estimation. The approach replaces conventional ${log}_{10}$-transformed time encoding with a bounded first-order toxicokinetic saturation feature and combines this representation with Deep Evidential Regression to support a joint chemical–temporal view of the QSAR/QSPR applicability domain. Using experimentally derived U.S. EPA Ecotoxicology Knowledgebase (ECOTOX) fish EC50 mortality records, models were trained on 48,728 acute-duration observations and evaluated retrospectively on 2090 temporally separated longer-duration observations. The combined toxicokinetic and evidential model reduced temporal extrapolation error relative to conventional time encoding while maintaining comparable within-domain validation performance. The learned population-level timescale converged to 221 ± 3 h, consistent with accumulation timescales extending beyond standard acute fish test durations. Epistemic uncertainty was positively associated with absolute prediction error across all 10 folds, suggesting that the uncertainty estimates retained sample-level information relevant to applicability-domain-aware molecular toxicity screening. Cross-species analyses further showed that model behavior depended on training time coverage, with greater convergence when available assays covered a larger fraction of the learned timescale. These results suggest that toxicokinetic-informed temporal encoding can improve uncertainty-aware QSAR/QSPR modeling of environmental contaminant toxicity and support prioritization of compounds for further testing, while complementing rather than replacing chronic bioassays. Full article

Characterizing the human health risk posed by constituents in drinking water is often challenging due to a lack of published toxicity values. The PIANO (Paraffin, Isoparaffin, Aromatic, Naphthene, and Olefin) analytical method measures nearly 300 compounds in JP-5 jet fuel, 43 of which have published oral reference doses (RfDs). The remaining compounds are typically assigned surrogate toxicity values. We predict RfDs for 290 PIANO compounds using Quantitative Structure–Activity Relationship (QSAR) models based on stepwise linear regression of 2-dimensional molecular descriptors (MDs) and published toxicity values. Five training groups, created by randomly selecting 80% of the non-PIANO compounds and 50% of the 43 PIANO compounds that have RfDs within a master dataset of 1113 compounds, were used to develop five QSAR models. We used the geometric means of four QSAR model results of sufficient quality to predict RfDs for compounds lacking toxicological information. For compounds with known RfDs, 884 (79%) were within 8-fold of published RfDs, well within the acknowledged uncertainty inherent in published RfDs. Our approach has applicability beyond PIANO compounds and represents a new alternative methodology (NAM) that may be used to reduce uncertainty in human health risk assessment and guide regulatory decisions. Full article

Background/Objectives: Coagulation factor XIa (FXIa) has emerged as a promising therapeutic target for the development of safer anticoagulant therapies with reduced bleeding risk. This study aimed to develop an interpretable machine learning-driven quantitative structure–activity relationship (QSAR) framework for predicting the inhibitory activity of FXIa inhibitors and supporting virtual screening applications. Methods: A total of 3026 curated compounds retrieved from the ChEMBL database were used for regression modeling, whereas 2119 compounds were retained for classification modeling after excluding intermediate-activity molecules. Molecular descriptors were generated using RDKit, Mordred, and Morgan fingerprint representations. Following preprocessing and feature selection, multiple machine learning algorithms were systematically benchmarked. Model robustness and reliability were further evaluated using 5-fold cross-validation, scaffold-aware validation, applicability domain analysis, and Y-randomization testing. Results: Nonlinear ensemble learning approaches consistently outperformed conventional linear algorithms. The optimized HistGradientBoostingRegressor achieved the best regression performance, with an independent test-set R² value of 0.711 and an RMSE value of 0.759, whereas the optimized classification model achieved accuracies approaching 95%. SHAP analysis identified lipophilicity-related descriptors, aromatic scaffold organization, electrostatic surface properties, and molecular topology as major contributors to FXIa inhibitory activity prediction. In addition, a proof-of-concept virtual screening workflow successfully identified several candidate compounds exhibiting high predicted pKi values and elevated active-class probabilities. Conclusions: The proposed framework provides a robust, interpretable, and reproducible machine learning-driven QSAR strategy for FXIa inhibitor discovery and may facilitate future virtual screening campaigns and medicinal chemistry optimization studies targeting FXIa-associated anticoagulant drug discovery. Full article

This study addresses the growing need for efficient and reliable application of New Approach Methodologies (NAMs) to assess aquatic toxicity of chemicals in response to increasing environmental contamination and regulatory demands. Particular emphasis is placed on in silico methods, especially quantitative structure–activity relationship (QSAR) modeling. Curated and structurally diverse datasets were compiled for representative aquatic organisms from different trophic levels, including the microalga Raphidocelis subcapitata, the crustacean Daphnia magna, and fish species (zebrafish embryo and fathead minnow). The models demonstrated consistently strong predictive performance across the evaluated assays. They were based on interpretable molecular descriptors associated with lipophilicity, polarity, and molecular reactivity. Furthermore, interspecies quantitative structure–activity–activity relationship (QSAAR) models were developed, demonstrating that toxicity data from lower trophic levels, combined with structural descriptors, can effectively predict fish toxicity. These models support cross-species extrapolation and contribute to environmental hazard assessment and regulatory decision-making. Full article

Quantitative structure–activity relationship (QSAR) modeling has conventionally relied on expert-designed molecular descriptors to encode chemical structures. DeepSnap is a descriptor-free QSAR approach that converts prepared three-dimensional molecular conformers into image representations and feeds them directly into convolutional neural networks for activity prediction. This focused narrative review traces DeepSnap from its introduction in 2018 to its current state and places it within the broader landscape of descriptor-based QSAR, topology-based and 3D-aware graph neural networks, and related image-based or semi-image-based molecular representation approaches. Previous studies applied DeepSnap to Tox21 nuclear receptor and molecular initiating event endpoints, rat hepatic clearance, blood–brain barrier penetration, acute oral toxicity, and cosmetics–pharmaceutical compound classification. Across the DeepSnap series, image-based and descriptor-based predictions have provided complementary information, particularly in ensemble or consensus models. However, high or near-ceiling ROC–AUC values reported for selected endpoints should not be interpreted as indicating deterministic or universally generalizable predictions; rather, they should be considered in the context of endpoint-specific model development, image-rendering parameter optimization, possible class imbalance, split dependence, limited matched external replication, and incomplete benchmarking against modern molecular representation models. Limitations include a dependence on nonphysical rendering parameters, single- or representative-conformer input, incomplete matched benchmarking against 2D and 3D molecular representation models, and an interpretability gap addressed in part by CAM-family visualization in the AI-based Substance Hazard Integrated Prediction System (AI-SHIPS) and S-COPHY (a model developed by Shiseido for cosmetics–pharmaceutical compound classification). Future directions include standardized image-generation protocols, conformer-ensemble extensions, systematic interpretability analysis, matched benchmarking, and potential integration with graph-based and 3D-aware molecular learning approaches. Full article

A hybrid method combining biomimetic liquid chromatography with immobilized artificial membrane (IAM) and quantitative structure–activity relationships (QSARs) was used to derive helpful models for predicting selected properties related to distribution (binding to human serum albumin (log P_w/HSA)) and absorption (skin permeation (log K_w/sp), plant cuticle permeation (log P_w/pc), and human intestinal permeability (Caco-2)), and therefore influencing the effectiveness or unwanted effects of 199 synthetic compounds that are regarded as potential drugs or plant protection products. The molecules under investigation—derivatives of 5H-6,7-dihydroimidazo [2,1-c][1,2,4]triazole, 7,8-dihydroimidazo[2,1-c][1,2,4]triazin-4(6H)-one, 2,6,7,8-tetrahydroimidazo[2,1-c][1,2,4]triazine-3,4-dione, 1H-1,2,4-triazole, carbamic and phenoxyacetic acid—differ in their properties but all meet the requirements for xenobiotics to be considered as medicinal products. Reliable high-concept models were developed, indicating lipophilicity, molecular size, electronic properties, and the number of rotatable bonds as descriptors that determine the biological properties of these compounds. These models have been optimized and cross-validated, confirming their reliability and high predictivity. Full article

This study investigated the application of six machine learning regression algorithms such as Random Forest, CatBoost, K-Nearest Neighbours, XGBoost, LightGBM, and Gradient Boosting, paired with Molecular ACCess System (MACCS) key fingerprints for the quantitative prediction of aromatase (CYP19A1) inhibitory potency, expressed as pIC₅₀. A dataset of 187 compounds was assembled from the ChEMBL database (version 33, Target ID: CHEMBL1978) following by systematic data curation workflow encompassing duplicate removal, pIC₅₀ transformation, and activity-based filtering. Model performance was rigorously evaluated using an 80/20 stratified train/test split, 5-fold cross-validation, and Y-randomisation testing to ensure unbiased assessment of predictive generalisation. Feature selection via CatBoost permutation importance on the held-out test set identified the top 20 predictive MACCS keys from an initial 166-bit space, substantially reducing dimensionality and improving generalisation across all models. Among the algorithms evaluated, CatBoost trained on the top 20 features achieved the strongest test-set performance (R² = 0.693, RMSE = 0.794, MAE = 0.659) with the most stable cross-validation R² (0.062 ± 0.304), outperforming all other algorithms. Y-randomisation testing returned an empirical p-value of <0.01, confirming that model performance reflects genuine structure–activity relationships rather than statistical chance. Permutation importance and SHAP analysis identified nitrogen-containing heterocyclic fragments (MACCS_41, MACCS_145) and halide-bearing substructures (MACCS_109) as the primary structural determinants of aromatase inhibitory potency, consistent with established CYP19A1 pharmacophoric requirements. Application of the model to ten representative plasticizers demonstrated that the refined applicability domain (h* = 0.423) accommodated eight of the ten compounds, enabling reliable potency predictions across phthalate esters and bisphenol analogues. These findings establish a transparent and reproducible QSAR framework for first-tier endocrine disruption risk screening of plasticizers and highlight the importance of permutation-based feature selection and applicability domain assessment in QSAR model development. Full article

As key therapeutic targets for symptomatic treatment of Alzheimer’s disease (AD) according to the cholinergic hypothesis, acetylcholinesterase (AChE; EC 3.1.1.7) and butyrylcholinesterase (BChE; EC 3.1.1.8) have been the subject of numerous studies over decades, leading to large collections of different ligands with corresponding AChE and BChE activity. This vast amount of data provides an ideal basis for the implementation of different machine learning (ML) and deep learning (DL) tools in different steps of the drug discovery process. Mainly applied to identify potential strong inhibitors of AChE and to a lesser extent BChE, many quantitative structure–activity relationship (QSAR) models and other predictive tools have been constructed utilizing different ML algorithms and DL techniques with various success depending on the input data and specific context. Here, we provide an extensive overview of such cases reported in the literature in recent years. Full article

The use of bio-based and biodegradable plastic products (BBpPs) ensures the mitigation of environmental effects of fossil-based plastics, especially in humanitarian crises where waste management is challenging. Polyhydroxyalkanoates (PHAs) are promising biodegradable biopolymers that are biocompatible and do not cause microplastic pollution. However, experimental assessment of PHA biodegradation is challenged by its time- and resource-intensiveness. In this study, a comprehensive computational Quantitative Structure–Activity Relationship (QSAR)-based approach was developed to predict biodegradability of short chain length (scl)-PHA-based formulations consisting of various additives and building blocks. A novel curated dataset for the (scl)-PHA poly(-3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV), with literature-reported environmental and biodegradation parameters from lab-scale experiments in soil, marine, freshwater and compost systems, was constructed and used to develop and validate the introduced approach. Random forest (RF) and Extreme Gradient Boosting (XGBoost) machine learning (ML) models were optimized and validated with cross-validation and test set predictions. The optimal models reported high accuracy values of the coefficient of determination R², indicating excellent relationships between structure and biodegradation metrics. Further analysis of descriptor variable importance confirmed that biopolymer biodegradability was favorably affected by biodegradation time, while mechanisms, environmental conditions, and additives contributed secondary yet physically consistent effects. The proposed QSAR framework demonstrated a robust and interpretable web-based tool for predicting the environmental fate of PHBV in natural environments and supported the sustainable safe-by-design (SSbD) approach of next-generation biodegradable polymers. Full article

Background: Cysteinyl leukotrienes are components of slow-reacting substances of anaphylactic shock (SRS-A) and play a key role in asthma and inflammatory responses. Although chromone-2-carboxylic acids and substituted (aryloxy)alkanoic acids have the potential to be SRS-A antagonists, their comprehensive structure–activity relationships and pharmacokinetic characteristics remain understudied. Objective: This study integrated computational and experimental approaches, including QSAR modeling, molecular docking, ADMET analysis, molecular dynamics (MD) simulations, and antioxidant evaluation to identify and prioritize bifunctional compounds with anti-inflammatory and free radical-scavenging properties. Methods: A set of 68 compounds was analyzed using 2D and 3D quantitative structure–activity relationships (QSAR) (MLR, MNLR, SVR, ANN, and atom-based partial least squares). Molecular docking and 100 ns MD simulations were performed against the CysLT₁ receptor (PDB ID: 6RZ5). ADMET and drug-like properties of the compounds were predicted using ADMETlab 2.0 and SwissADME, and the in vitro antioxidant activity of the top-ranked compounds was evaluated using the DPPH method. Results: The atom-based 3D-QSAR model showed strong predictive power (R² = 0.9524, Q² = 0.5382). Compounds 25, 41, and 47 stood out with the most significant binding energies: −9.5 kcal/mol for 25, −10.0 kcal/mol for 41, and −9.4 kcal/mol for 47. MD simulations confirmed the structural stability and consistent interactions of the protein-compound 47 complex. ADMET analysis showed that compounds 25 and 41 had good pharmacokinetic properties, and in vitro antioxidant assays verified their free radical-scavenging efficacy. Conclusion: Our results highlight the utility of an integrated computational–experimental strategy for the discovery of dual-acting SRS-A antagonists. Compound 25 is highlighted as a promising lead compound for further preclinical development, which effectively combines leukotriene receptor antagonism and antioxidant activity. This framework provides an effective strategy for prioritizing lead compounds in anti-inflammatory drug development. Full article

Polyhydroxyalkanoates (PHAs) emerge as promising biodegradable polymers for sustainable applications, yet predicting their biodegradation behavior under different environmental conditions remains challenging. In this study, we propose a novel data-driven computational framework for predicting biodegradation-induced weight/mass loss in PHA-based materials. A comprehensive database of poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV)-based formulations was manually curated by systematically collecting and harmonizing material descriptors, environmental parameters, and experimental biodegradation outcomes from laboratory- and large-scale studies conducted in soil, marine, freshwater, and compost environments. Multiple regression-based quantitative structure–activity relationship (QSAR) models were developed and rigorously validated, demonstrating high predictive performance and strong correlations between polymer structure, environmental conditions and degradation behavior. “Exposure time”, “degradation environment” and “hydroxybutyrate (HB) ratio” were identified as the most important features for weight loss. Finally, the predictive model was integrated into the Jaqpot computational platform, enabling open access and facilitating data-driven assessment and design of biodegradable polymer systems. Full article

Understanding the adsorption of organic pollutants onto microplastics in aqueous environments is crucial for assessing their environmental behavior and ecological risks. Herein, we used density functional theory (DFT) computations to simulate the aqueous adsorption of 54 organic compounds onto three representative microplastics, namely polyethylene (PE), polyoxymethylene (POM), and polyvinyl alcohol (PVA). Afterwards, based on theoretical molecular structural descriptors, we developed six quantitative structure activity relationship (QSAR) models based on datasets of 43 and 54 organic compounds, respectively. The results demonstrated that the oxygen-containing POM and PVA microplastics exhibited weaker adsorption in the aqueous phase compared to that in the gas phase. Furthermore, it revealed that the electron-rich atoms, van der Waals volumes and molecular polarizability exert substantial effects on the adsorption process on microplastics in water. These robust QSAR models can enable the prediction of adsorption energies for various organic pollutants on microplastics, which can offer a rapid approach for generating adsorption data. Moreover, the insights into adsorption mechanisms can provide a theoretical basis for designing modified or alternative plastics with lower environmental risks. Full article

Quantitative structure–activity/property relationship (QSAR/QSPR) is a well-established methodology widely used to model molecular properties based on structure and is applied in fields such as drug design and environmental protection. The knowledge and procedures developed and used in QSPR modelling will be applied to the validation of protein folding rate models. Understanding the protein folding process is considered one of the most important scientific topics, and identifying the fundamental factors responsible for protein folding has been the subject of intensive research over the past 30 years. Among the structural descriptors determining the protein folding rate, the length of the protein sequence, the content of regular secondary structures, and the average contact row distance between amino acids in the 3D structure are the most important. Comparative studies of different methods for predicting protein folding rates are occasionally published, and we conducted one such study. We found that the experimental data in literature databases and the data available online are inconsistent and scattered. This is partly due to differences in experimental data and protein sequence lengths, but more so due to the questionable quality of the models themselves. We observed very large deviations in the predictions of ln(k_f) by some of the analysed models implemented as web servers. The root mean square errors (RMSEs) of some of the analysed models in predicting ln(k_f) for a new external set of proteins are much larger than the RMSEs obtained for the same models on the training sets. External validation demonstrates that protein folding rate models available on web servers have accuracy for external protein sets comparable to that of a simple model based solely on the logarithm of protein chain length. This finding, which highlights the importance of external model validation as recommended by the OECD guidelines for QSAR validation, is fundamental and offers a new perspective for improving protein folding rate models by applying the knowledge and procedures used in the QSPR methodology. Full article

Mineral scale deposition remains a major flow-assurance constraint in oil and gas operations, especially in water-flooding and produced-water reinjection, where mixing between incompatible brines promotes super-saturation and precipitation of poorly soluble salts. This work introduces a novel extension of traditional methods used for modeling chemical inhibition and the predictive evaluation of oilfield scale-inhibitor molecules. A systematically optimized Two-Dimensional Quantitative Structure–Activity Relationship Model based on the k-Nearest Neighbors algorithm 2D-QSAR-KNN model was developed to quantitatively link molecular constitution of phosphonate inhibitors, brine chemistry, and operating factors with inhibition efficiency IE %. The optimized model achieved strong accuracy and generalization R²_train = 0.9182, R²_test = 0.9306, and R²_global = 0.9208 with low prediction errors RMSE_train = 4.7888%, RMSE_test = 4.5485%, and RMSE_global = 4.7421%. Median absolute errors remained minimal for the train set = 0.80%, and test set = 1.63%, and model stability was confirmed by high correlation with experimental IE % r = 0.94 and R²_train/R²_test ≈ 0.99, showing no sign of overfitting. Additionally, an inverse-2D-QSAR framework was applied to identify the optimal molecular descriptor profile expected to maximize inhibitory performance within normalized bounds, providing rational rules for next-generation inhibitor design. The findings highlight the practical value of QSAR-inspired AI modeling to accelerate molecule screening and dosage exploration prior to laboratory validation, supporting more cost-effective, interpretable, and environmentally aware sulfate-scale inhibition strategies under high-salinity reservoir conditions. Full article