SpookyNet: Learning force fields with electronic degrees of freedom and nonlocal effects

Nature Communications, Год журнала: 2021, Номер 12(1)

Опубликована: Дек. 14, 2021

Machine-learned force fields (ML-FFs) combine the accuracy of ab initio methods with efficiency conventional fields. However, current ML-FFs typically ignore electronic degrees freedom, such as total charge or spin state, and assume chemical locality, which is problematic when molecules have inconsistent states, nonlocal effects play a significant role. This work introduces SpookyNet, deep neural network for constructing explicit treatment freedom quantum nonlocality. Chemically meaningful inductive biases analytical corrections built into architecture allow it to properly model physical limits. SpookyNet improves upon state-of-the-art (or achieves similar performance) on popular chemistry data sets. Notably, able generalize across conformational space can leverage learned insights, e.g. by predicting unknown thus helping close further important remaining gap today's machine learning models in chemistry.

Язык: Английский

---

Machine Learning Methods for Small Data Challenges in Molecular Science

Chemical Reviews, Год журнала: 2023, Номер 123(13), С. 8736 - 8780

Опубликована: Июнь 29, 2023

Small data are often used in scientific and engineering research due to the presence of various constraints, such as time, cost, ethics, privacy, security, technical limitations acquisition. However, big have been focus for past decade, small their challenges received little attention, even though they technically more severe machine learning (ML) deep (DL) studies. Overall, challenge is compounded by issues, diversity, imputation, noise, imbalance, high-dimensionality. Fortunately, current era characterized technological breakthroughs ML, DL, artificial intelligence (AI), which enable data-driven discovery, many advanced ML DL technologies developed inadvertently provided solutions problems. As a result, significant progress has made decade. In this review, we summarize analyze several emerging potential molecular science, including chemical biological sciences. We review both basic algorithms, linear regression, logistic regression (LR),

Язык: Английский

---

SELFIES and the future of molecular string representations

Patterns, Год журнала: 2022, Номер 3(10), С. 100588 - 100588

Опубликована: Окт. 1, 2022

Artificial intelligence (AI) and machine learning (ML) are expanding in popularity for broad applications to challenging tasks chemistry materials science. Examples include the prediction of properties, discovery new reaction pathways, or design molecules. The needs read write fluently a chemical language each these tasks. Strings common tool represent molecular graphs, most popular string representation, Smiles, has powered cheminformatics since late 1980s. However, context AI ML chemistry, Smiles several shortcomings—most pertinently, combinations symbols lead invalid results with no valid interpretation. To overcome this issue, molecules was introduced 2020 that guarantees 100% robustness: SELF-referencing embedded (Selfies). Selfies simplified enabled numerous chemistry. In perspective, we look future discuss representations, along their respective opportunities challenges. We propose 16 concrete projects robust representations. These involve extension toward domains, exciting questions at interface languages, interpretability both humans machines. hope proposals will inspire follow-up works exploiting full potential representations

Язык: Английский

---

Inverse design of 3d molecular structures with conditional generative neural networks

Nature Communications, Год журнала: 2022, Номер 13(1)

Опубликована: Фев. 21, 2022

The rational design of molecules with desired properties is a long-standing challenge in chemistry. Generative neural networks have emerged as powerful approach to sample novel from learned distribution. Here, we propose conditional generative network for 3d molecular structures specified chemical and structural properties. This agnostic bonding enables targeted sampling distributions, even domains where reference calculations are sparse. We demonstrate the utility our method inverse by generating motifs or composition, discovering particularly stable molecules, jointly targeting multiple electronic beyond training regime.

Язык: Английский

---

Design, Synthesis, Docking, DFT, MD Simulation Studies of a New Nicotinamide-Based Derivative: In Vitro Anticancer and VEGFR-2 Inhibitory Effects

Molecules, Год журнала: 2022, Номер 27(14), С. 4606 - 4606

Опубликована: Июль 19, 2022

A nicotinamide-based derivative was designed as an antiproliferative VEGFR-2 inhibitor with the key pharmacophoric features needed to interact catalytic pocket. The ability of congener ((E)-N-(4-(1-(2-(4-benzamidobenzoyl)hydrazono)ethyl)phenyl)nicotinamide), compound 10, bind enzyme demonstrated by molecular docking studies. Furthermore, six various MD simulations studies established excellent binding 10 over 100 ns, exhibiting optimum dynamics. MM-GBSA confirmed proper a total exact energy -38.36 Kcal/Mol. also revealed crucial amino acids in through free decomposition and declared interactions variation inside via Protein-Ligand Interaction Profiler (PLIP). Being new, its structure optimized DFT. DFT mode VEGFR-2. ADMET (in silico) profiling indicated examined compound's acceptable range drug-likeness. synthesized condensation N-(4-(hydrazinecarbonyl)phenyl)benzamide N-(4-acetylphenyl)nicotinamide, where carbonyl group has been replaced imine group. in-vitro were consonant obtained silico results prohibited IC50 value 51 nM. Compound showed effects against MCF-7 HCT 116 cancer cell lines values 8.25 6.48 μM, revealing magnificent selectivity indexes 12.89 16.41, respectively.

Язык: Английский

---

Accurate global machine learning force fields for molecules with hundreds of atoms

Science Advances, Год журнала: 2023, Номер 9(2)

Опубликована: Янв. 11, 2023

Global machine learning force fields, with the capacity to capture collective interactions in molecular systems, now scale up a few dozen atoms due considerable growth of model complexity system size. For larger molecules, locality assumptions are introduced, consequence that nonlocal not described. Here, we develop an exact iterative approach train global symmetric gradient domain (sGDML) fields (FFs) for several hundred atoms, without resorting any potentially uncontrolled approximations. All atomic degrees freedom remain correlated sGDML FF, allowing accurate description complex molecules and materials present phenomena far-reaching characteristic correlation lengths. We assess accuracy efficiency on newly developed MD22 benchmark dataset containing from 42 370 atoms. The robustness our is demonstrated nanosecond path-integral dynamics simulations supramolecular complexes dataset.

Язык: Английский

---

Extending machine learning beyond interatomic potentials for predicting molecular properties

Nature Reviews Chemistry, Год журнала: 2022, Номер 6(9), С. 653 - 672

Опубликована: Авг. 25, 2022

Язык: Английский

---

Machine Learning Interatomic Potentials and Long-Range Physics

The Journal of Physical Chemistry A, Год журнала: 2023, Номер 127(11), С. 2417 - 2431

Опубликована: Фев. 21, 2023

Advances in machine learned interatomic potentials (MLIPs), such as those using neural networks, have resulted short-range models that can infer interaction energies with near ab initio accuracy and orders of magnitude reduced computational cost. For many atom systems, including macromolecules, biomolecules, condensed matter, model become reliant on the description short- long-range physical interactions. The latter terms be difficult to incorporate into an MLIP framework. Recent research has produced numerous considerations for nonlocal electrostatic dispersion interactions, leading a large range applications addressed MLIPs. In light this, we present Perspective focused key methodologies being used where presence physics chemistry are crucial describing system properties. strategies covered include MLIPs augmented corrections, electrostatics calculated charges predicted from atomic environment descriptors, use self-consistency message passing iterations propagated information, obtained via equilibration schemes. We aim provide pointed discussion support development learning-based systems contributions only nearsighted deficient.

Язык: Английский

---

Machine Learning-Assisted Low-Dimensional Electrocatalysts Design for Hydrogen Evolution Reaction

Nano-Micro Letters, Год журнала: 2023, Номер 15(1)

Опубликована: Окт. 13, 2023

Abstract Efficient electrocatalysts are crucial for hydrogen generation from electrolyzing water. Nevertheless, the conventional "trial and error" method producing advanced is not only cost-ineffective but also time-consuming labor-intensive. Fortunately, advancement of machine learning brings new opportunities discovery design. By analyzing experimental theoretical data, can effectively predict their evolution reaction (HER) performance. This review summarizes recent developments in low-dimensional electrocatalysts, including zero-dimension nanoparticles nanoclusters, one-dimensional nanotubes nanowires, two-dimensional nanosheets, as well other electrocatalysts. In particular, effects descriptors algorithms on screening investigating HER performance highlighted. Finally, future directions perspectives electrocatalysis discussed, emphasizing potential to accelerate electrocatalyst discovery, optimize performance, provide insights into electrocatalytic mechanisms. Overall, this work offers an in-depth understanding current state its research.

Язык: Английский

---

Uncertainty-driven dynamics for active learning of interatomic potentials

Nature Computational Science, Год журнала: 2023, Номер 3(3), С. 230 - 239

Опубликована: Март 6, 2023

Machine learning (ML) models, if trained to data sets of high-fidelity quantum simulations, produce accurate and efficient interatomic potentials. Active (AL) is a powerful tool iteratively generate diverse sets. In this approach, the ML model provides an uncertainty estimate along with its prediction for each new atomic configuration. If passes certain threshold, then configuration included in set. Here we develop strategy more rapidly discover configurations that meaningfully augment training The uncertainty-driven dynamics active (UDD-AL), modifies potential energy surface used molecular simulations favor regions space which there large uncertainty. performance UDD-AL demonstrated two AL tasks: sampling conformational glycine promotion proton transfer acetylacetone. method shown efficiently explore chemically relevant space, may be inaccessible using regular dynamical at target temperature conditions.

Язык: Английский

---