QSAR Classification Modeling Using Machine Learning with a Consensus-Based Approach for Multivariate Chemical Hazard End Points

ACS Omega, Год журнала: 2024, Номер 9(51), С. 50796 - 50808

Опубликована: Дек. 12, 2024

This study introduces an innovative computational approach using hybrid machine learning models to predict toxicity across eight critical end points: cardiac toxicity, inhalation dermal oral skin irritation, sensitization, eye and respiratory irritation. Leveraging advanced cheminformatics tools, we extracted relevant features from curated data sets, incorporating a range of descriptors such as Morgan circular fingerprints, MACCS keys, Mordred calculation descriptors, physicochemical properties. The consensus model was developed by selecting the best-performing classifier-Random Forest (RF), eXtreme Gradient Boosting (XGBoost), or Support Vector Machines (SVM)-for each descriptor, optimizing predictive accuracy robustness points. obtained strong performance, with area under curve (AUC) scores ranging 0.78 0.90. framework offers reliable, ethical, effective in silico chemical safety assessment, underscoring potential methods support both regulatory research applications prediction.

Язык: Английский

---

Molecular docking and pharmacokinetic evaluations of curcumin-based scaffolds as MDM2-p53 inhibitors

Discover Chemistry., Год журнала: 2025, Номер 2(1)

Опубликована: Март 24, 2025

Язык: Английский

---

The Goldilocks paradigm: comparing classical machine learning, large language models, and few-shot learning for drug discovery applications

Communications Chemistry, Год журнала: 2024, Номер 7(1)

Опубликована: Июнь 12, 2024

Abstract Recent advances in machine learning (ML) have led to newer model architectures including transformers (large language models, LLMs) showing state of the art results text generation and image analysis as well few-shot (FSLC) models which offer predictive power with extremely small datasets. These new may promise, yet ‘no-free lunch’ theorem suggests that no single algorithm can outperform at all possible tasks. Here, we explore capabilities classical (SVR), FSLC, transformer (MolBART) over a range dataset tasks show ‘goldilocks zone’ for each type, size feature distribution (i.e. “diversity”) determines optimal strategy. When datasets are ( < 50 molecules), FSLC tend both ML transformers. small-to-medium sized (50-240 molecules) diverse, learning. Finally, when larger sufficient size, then perform best, suggesting choose likely depends on available, its diversity. findings help answer perennial question is be used faced dataset.

Язык: Английский

---

Hydrolysis of plasma-polymerized poly(ethylene glycol)/ZnO nanocomposites in food simulants: Identification of components and potential toxicity

Food Chemistry, Год журнала: 2024, Номер 464, С. 141571 - 141571

Опубликована: Окт. 10, 2024

Язык: Английский

---

New Theobromine Apoptotic Analogue with Anticancer Potential Targeting the EGFR Protein: Computational and In Vitro Studies

ACS Omega, Год журнала: 2024, Номер 9(14), С. 15861 - 15881

Опубликована: Март 27, 2024

Aim: The aim of this study was to design and examine a novel epidermal growth factor receptor (EGFR) inhibitor with apoptotic properties by utilizing the essential structural characteristics existing EGFR inhibitors as foundation. Method: began natural alkaloid theobromine developed new semisynthetic derivative (T-1-PMPA). Computational ADMET assessments were conducted first evaluate its anticipated safety general drug-likeness. Deep density functional theory (DFT) computations initially performed validate three-dimensional (3D) structure reactivity T-1-PMPA. Molecular docking against proteins investigate T-1-PMPA's binding affinity inhibitory potential. Additional molecular dynamics (MD) simulations over 200 ns along MM-GPSA, PLIP, principal component analysis trajectories (PCAT) experiments employed verify Afterward, T-1-PMPA semisynthesized proposed in silico findings through several vitro examinations. Results: DFT studies indicated using electrostatic potential, global reactive indices, total states. docking, MD simulations, ED suggested protein. predicted In demonstrated that effectively inhibited EGFRWT EGFR790m, IC50 values 86 561 nM, respectively, compared Erlotinib (31 456 nM). also showed significant suppression proliferation HepG2 MCF7 malignant cell lines, 3.51 4.13 μM, respectively. selectivity indices two cancer lines overall Flow cytometry confirmed effects increasing percentage apoptosis 42% 31, 3% Erlotinib-treated control cells, qRT-PCR further supported revealing increases levels Casp3 Casp9. Additionally, controlled TNFα IL2 74 50%, comparing Erlotinib's (84 74%), Conclusion: conclusion, our study's suggest potential promising anticancer lead compound targeting EGFR.

Язык: Английский

---

Anti-breast cancer potential of a new xanthine derivative: In silico, antiproliferative, selectivity, VEGFR-2 inhibition, apoptosis induction and migration inhibition studies

Pathology - Research and Practice, Год журнала: 2023, Номер 251, С. 154894 - 154894

Опубликована: Окт. 14, 2023

Язык: Английский

---

Predicting the Hallucinogenic Potential of Molecules Using Artificial Intelligence

ACS Chemical Neuroscience, Год журнала: 2024, Номер 15(16), С. 3078 - 3089

Опубликована: Авг. 2, 2024

The development of new drugs addressing serious mental health and other disorders should avoid the psychedelic experience. Analogs can have clinical utility are termed "psychoplastogens". These represent promising candidates for treating opioid use disorder to reduce drug dependence, with rarely reported adverse effects. This abuse cessation is linked induction neuritogenesis increased neuroplasticity, a hallmark molecules, such as lysergic acid diethylamine. Some, but not all psychoplastogens may act through G-protein coupled receptor (GPCR) 5HT

Язык: Английский

---

ApisTox: a new benchmark dataset for the classification of small molecules toxicity on honey bees

Scientific Data, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 2, 2025

Abstract The global decline in bee populations poses significant risks to agriculture, biodiversity, and environmental stability. To bridge the gap existing data, we introduce ApisTox, a comprehensive dataset focusing on toxicity of pesticides honey bees (Apis mellifera). This combines leverages data from sources such as ECOTOX PPDB, providing an extensive, consistent, curated collection that surpasses previous datasets. ApisTox incorporates wide array including levels for chemicals, details time their publication literature, identifiers linking them external chemical databases. may serve important tool agricultural research, but also can support development policies practices aimed at minimizing harm populations. Finally, offers unique resource benchmarking molecular property prediction methods agrochemical compounds, facilitating advancements both science chemoinformatics. makes it valuable academic research practical applications conservation.

Язык: Английский

---

First report on q-RASTR modelling of hazardous dose (HD ₅ ) for acute toxicity of pesticides: an efficient and reliable approach towards safeguarding the sensitive avian species

SAR and QSAR in environmental research, Год журнала: 2025, Номер unknown, С. 1 - 17

Опубликована: Фев. 11, 2025

Pesticides are crucial in modern agriculture, significantly enhancing crop productivity by managing pests. It is important to evaluate their toxicity minimize health risks bird species and preserve ecosystem balance. Traditional parameters including lethal concentration (LC50) or median dose (LD50) often underestimate hazards due limited data uncertainty about the most sensitive tested. This limitation can be addressed using extrapolation factors like HD5 accounting for 50% mortality of 5% species. In this research, a QSTR model was developed utilizing diverse set 480 pesticides partial least squares (PLS) regression with 2D descriptors. Additionally, PLS-based quantitative read-across structure-toxicity relationship (q-RASTR) classification based models were constructed. The q-RASTR outperformed traditional approaches, achieving robust statistical performance internal validation metrics r2 = 0.623, Q2 0.569 external Q2F1 0.541, Q2F2 0.540. Key influencing avian identified. used screen Pesticide Properties Database (PPDB) recognize toxic species, aligning well real-world data. work provides more economical ethical alternative conventional vivo testing methods, aiding regulatory bodies industries developing safer, environmentally friendly pesticides.

Язык: Английский

---

Computational Approaches for Predicting Drug Interactions with Human Organic Anion Transporter 4 (OAT4)

Molecular Pharmaceutics, Год журнала: 2025, Номер unknown

Опубликована: Март 20, 2025

Human Organic Anion Transporter 4 (OAT4) is predominantly expressed in the kidneys, particularly apical membrane of proximal tubule cells. This transporter involved renal handling endogenous and exogenous organic anions (OAs), making it an important for drug–drug interactions (DDIs). To better understand OAT4-compound interactions, we generated single concentration (25 μM) vitro inhibition data over 1400 small molecules against uptake fluorescent OA 6-carboxyfluorescein (6-CF) Chinese hamster ovary (CHO) Several drugs exhibiting higher than 50% this initial screen were selected to determine IC50 values three structurally distinct OAT4 substrates: estrone sulfate (ES), ochratoxin A (OTA), 6-CF. These then compared drug plasma as per 2020 FDA interaction (DDI) guidance. screened compounds, including some not previously reported, emerged novel inhibitors OAT4. also used build machine learning classification models predict activity potential inhibitors. We multiple algorithms cleaning techniques model these screening investigated utility conformal predictors a leave-out set. experimental computational approaches allowed us diverse unbalanced enable predictions DDIs mediated by transporter.

Язык: Английский

---