Machine Learning Accelerated Discovery of Antimicrobial Inorganic Nanomaterials

The Journal of Physical Chemistry Letters, Год журнала: 2025, Номер unknown, С. 5627 - 5635

Опубликована: Май 29, 2025

The growing prevalence of infectious diseases and the increasing threat bacterial resistance have drawn widespread attention to antimicrobial inorganic nanomaterials. However, diversity, abundance, complex mechanisms these materials present significant challenges in identifying new agents that are both efficient cost-effective with broad-spectrum activity. In response, this study applied machine learning for first time discover Information on over 2,000 nanomaterials was extracted from more than 8,000 papers. An unsupervised analysis conducted assess data distribution explore relationships between material features activity high-dimensional space. A series models were trained. Through evaluation six performance metrics, five key identified 27 dimensions. To further quantify structure-activity relationships, a genetic programming-symbolic classification model employed generate precise mathematical formula prediction accuracy 0.83. Using formula, 43 predicted. Of these, four synthesized their antibacterial properties experimentally validated. This work not only provides next generation approach designing but also opens avenues applying science.

Язык: Английский

---

Interpretable machine learning models for predicting the antitumor effects of metal and metal oxide nanomaterials

RSC Advances, Год журнала: 2025, Номер 15(21), С. 17036 - 17048

Опубликована: Янв. 1, 2025

Understanding the toxic behavior of metal and oxide nanoparticles (M/MOx NPs) is essential for effective tumor diagnosis treatment, yet generalizing findings remains challenging due to limited data, sampling variability, unreported complexities, low model accuracy, a lack interpretability. To address these issues minimize extensive experimentation, we combined quantum chemistry calculations with published toxicity data develop machine learning achieving over 90% accuracy in cross-validation. Utilizing 39 descriptors extracted from 152 articles, our dataset comprises 2765 instances covering various nanoparticle types, detection methods, cell types. We enhanced representation Jaccard similarity coefficient employed Feature Importance Shapley Additive Explanations (SHAP) identify key factors influencing cytotoxicity, such as concentration, exposure time, zeta potential, diameter, COSMO area (CA), coating, testing electronegativity, HOMO energy, molecular weight. Additionally, analyzed interactions among features their influence on predictions, synthesized novel nanoparticles, assessed physicochemical properties anti-tumor toxicity. Cytotoxicity experiments newly further validated model's generalizability, revealing hidden relationships enabling predictions previously unseen samples. This approach supports preliminary computer-aided screenings, significantly reducing need labor-intensive experimentation.

Язык: Английский

---

Application of Computing as a High-Practicability and -Efficiency Auxiliary Tool in Nanodrugs Discovery

Pharmaceutics, Год журнала: 2023, Номер 15(4), С. 1064 - 1064

Опубликована: Март 25, 2023

Research and development (R&D) of nanodrugs is a long, complex uncertain process. Since the 1960s, computing has been used as an auxiliary tool in field drug discovery. Many cases have proven practicability efficiency Over past decade, computing, especially model prediction molecular simulation, gradually applied to nanodrug R&D, providing substantive solutions many problems. Computing made important contributions promoting data-driven decision-making reducing failure rates time costs discovery nanodrugs. However, there are still few articles examine, it necessary summarize research direction. In review, we application various stages including physicochemical properties biological activities prediction, pharmacokinetics analysis, toxicological assessment other related applications. Moreover, current challenges future perspectives methods also discussed, with view help become high-practicability -efficiency development.

Язык: Английский

---

Layered nanomaterials for renewable energy generation and storage

Materials Advances, Год журнала: 2023, Номер 5(2), С. 394 - 408

Опубликована: Дек. 11, 2023

This study focuses on potential applications of two-dimensional (2D) materials in renewable energy research.

Язык: Английский

---

In chemico methodology for engineered nanomaterial categorization according to number, nature and oxidative potential of reactive surface sites

Environmental Science Nano, Год журнала: 2024, Номер 11(9), С. 3744 - 3760

Опубликована: Янв. 1, 2024

Methanol probe chemisorption quantifies the number of reactive sites at surface engineered nanomaterials, enabling normalization per site in reactivity and toxicity tests, rather than mass or physical area. Subsequent temperature-programmed reaction (TPSR) chemisorbed methanol identifies nature (acidic, basic, redox combination thereof) their reactivity. Complementary to assay, a dithiothreitol (DTT) oxidation is used evaluate capacity. These acellular approaches quantify number, nature, constitute new approach methodology (NAM) for site-specific classification nanomaterials. As proof concept, CuO, CeO2, ZnO, Fe3O4, CuFe2O4, Co3O4 two TiO2 nanomaterials were probed. A harmonized descriptor ENMs was obtained: DTT rate site, oxidative turnover frequency (OxTOF). CuO CuFe2O4 exhibit largest density possess highest oxidizing ability series, as estimated by reaction, followed CeO2 NM-211 then titania (DT-51 NM-101) Fe3O4. depletion ZnO NM-110 associated with dissolved zinc ions particles; however, basic characteristics particles evidenced TPSR. assays allow ranking eight into three categories statistically different potentials: are most reactive; ceria exhibits moderate reactivity; iron oxide titanias low potential.

Язык: Английский

---