Current and emerging deep-learning methods for the simulation of fluid dynamics DOI Creative Commons
Mario Lino, Stathi Fotiadis, Anil A. Bharath

и другие.

Proceedings of the Royal Society A Mathematical Physical and Engineering Sciences, Год журнала: 2023, Номер 479(2275)

Опубликована: Июль 1, 2023

Over the last decade, deep learning (DL), a branch of machine learning, has experienced rapid progress. Powerful tools for tasks that have been traditionally complex to automate developed, such as image synthesis and natural language processing. In context simulating fluid dynamics, this led series novel DL methods replacing or augmenting conventional numerical solvers. We broadly classify these into physics- data-driven methods. Physics-driven methods, generally, tune model provide an analytical differentiable solution given dynamics problem by minimizing residuals governing partial differential equations. Data-driven fast approximate any shares some physical properties with observations used when tuning model’s parameters. Meanwhile, symbiosis solvers promising results in turbulence modelling accelerating iterative However, present challenges. Exclusively flow simulators often suffer from poor extrapolation, error accumulation time-dependent simulations, well difficulties training against turbulent flows. Substantial effort is, therefore, being invested approaches may improve current state art.

Язык: Английский

Self-Driving Laboratories for Chemistry and Materials Science DOI Creative Commons
Gary Tom, Stefan P. Schmid, Sterling G. Baird

и другие.

Chemical Reviews, Год журнала: 2024, Номер 124(16), С. 9633 - 9732

Опубликована: Авг. 13, 2024

Self-driving laboratories (SDLs) promise an accelerated application of the scientific method. Through automation experimental workflows, along with autonomous planning, SDLs hold potential to greatly accelerate research in chemistry and materials discovery. This review provides in-depth analysis state-of-the-art SDL technology, its applications across various disciplines, implications for industry. additionally overview enabling technologies SDLs, including their hardware, software, integration laboratory infrastructure. Most importantly, this explores diverse range domains where have made significant contributions, from drug discovery science genomics chemistry. We provide a comprehensive existing real-world examples different levels automation, challenges limitations associated each domain.

Язык: Английский

Процитировано

39
Invalid SMILES are beneficial rather than detrimental to chemical language models DOI Creative Commons
Michael A. Skinnider

Nature Machine Intelligence, Год журнала: 2024, Номер 6(4), С. 437 - 448

Опубликована: Март 29, 2024

Abstract Generative machine learning models have attracted intense interest for their ability to sample novel molecules with desired chemical or biological properties. Among these, language trained on SMILES (Simplified Molecular-Input Line-Entry System) representations been subject the most extensive experimental validation and widely adopted. However, these what is perceived be a major limitation: some fraction of strings that they generate are invalid, meaning cannot decoded structure. This shortcoming has motivated remarkably broad spectrum work designed mitigate generation invalid correct them post hoc. Here I provide causal evidence produce outputs not harmful but instead beneficial models. show provides self-corrective mechanism filters low-likelihood samples from model output. Conversely, enforcing valid produces structural biases in generated molecules, impairing distribution limiting generalization unseen space. Together, results refute prevailing assumption reframe as feature, bug.

Язык: Английский

Процитировано

17
Harnessing the Power of Artificial Intelligence in Pharmaceuticals: Current Trends and Future Prospects DOI Creative Commons
Saha Aritra, Indu Singh

Intelligent Pharmacy, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

2
Machine learning and chemometrics for electrochemical sensors: moving forward to the future of analytical chemistry DOI
Pumidech Puthongkham, Supacha Wirojsaengthong, Akkapol Suea‐Ngam

и другие.

The Analyst, Год журнала: 2021, Номер 146(21), С. 6351 - 6364

Опубликована: Янв. 1, 2021

Electrochemical sensors and biosensors have been successfully used in a wide range of applications, but systematic optimization nonlinear relationships compromised for electrode fabrication data analysis. Machine learning experimental designs are chemometric tools that proved to be useful method development This minireview summarizes recent applications machine electroanalytical chemistry. First, designs, e.g., full factorial, central composite, Box-Behnken discussed as approaches optimize consider the effects from individual variables their interactions. Then, principles algorithms, including linear logistic regressions, neural network, support vector machine, introduced. These models implemented extract complex between chemical structures electrochemical properties analyze complicated improve calibration analyte classification, such electronic tongues. Lastly, future is outlined. strategies will accelerate enhance performance devices point-of-care diagnostics commercialization.

Язык: Английский

Процитировано

78
Beam Search for Automated Design and Scoring of Novel ROR Ligands with Machine Intelligence** DOI
Michaël Moret,

Moritz Helmstädter,

Francesca Grisoni

и другие.

Angewandte Chemie International Edition, Год журнала: 2021, Номер 60(35), С. 19477 - 19482

Опубликована: Июнь 24, 2021

Chemical language models enable de novo drug design without the requirement for explicit molecular construction rules. While such have been applied to generate novel compounds with desired bioactivity, actual prioritization and selection of most promising computational designs remains challenging. Herein, we leveraged probabilities learnt by chemical beam search algorithm as a model-intrinsic technique automated molecule scoring. Prospective application this method yielded inverse agonists retinoic acid receptor-related orphan receptors (RORs). Each was synthesizable in three reaction steps presented low-micromolar nanomolar potency towards RORγ. This sampling eliminates strict need external compound scoring functions, thereby further extending applicability generative artificial intelligence data-driven discovery.

Язык: Английский

Процитировано

75
Artificial intelligence-enhanced drug design and development: Toward a computational precision medicine DOI
Philippe Moingeon, Mélaine A. Kuenemann, Mickaël Guedj

и другие.

Drug Discovery Today, Год журнала: 2021, Номер 27(1), С. 215 - 222

Опубликована: Сен. 21, 2021

Язык: Английский

Процитировано

71
Chemical Space Exploration with Active Learning and Alchemical Free Energies DOI Creative Commons
Yuriy Khalak, Gary Tresadern, David F. Hahn

и другие.

Journal of Chemical Theory and Computation, Год журнала: 2022, Номер 18(10), С. 6259 - 6270

Опубликована: Сен. 23, 2022

Drug discovery can be thought of as a search for needle in haystack: searching through large chemical space the most active compounds. Computational techniques narrow experimental follow up, but even they become unaffordable when evaluating numbers molecules. Therefore, machine learning (ML) strategies are being developed computationally cheaper complementary navigating and triaging libraries. Here, we explore how an protocol combined with first-principles based alchemical free energy calculations to identify high affinity phosphodiesterase 2 (PDE2) inhibitors. We first calibrate procedure using set experimentally characterized PDE2 binders. The optimized is then used prospectively on library navigate toward potent In cycle, at every iteration small fraction compounds probed by obtained affinities train ML models. With successive rounds, binders identified explicitly only subset library, thus providing efficient that robustly identifies true positives.

Язык: Английский

Процитировано

59
Machine learning in bioprocess development: from promise to practice DOI
Laura M. Helleckes, Johannes Hemmerich, Wolfgang Wiechert

и другие.

Trends in biotechnology, Год журнала: 2022, Номер 41(6), С. 817 - 835

Опубликована: Ноя. 29, 2022

Язык: Английский

Процитировано

58
When machine learning meets molecular synthesis DOI
João C. A. Oliveira, Johanna Frey, Shuo‐Qing Zhang

и другие.

Trends in Chemistry, Год журнала: 2022, Номер 4(10), С. 863 - 885

Опубликована: Авг. 31, 2022

Язык: Английский

Процитировано

42
A knowledge-guided pre-training framework for improving molecular representation learning DOI Creative Commons
Han Li,

Ruotian Zhang,

Yaosen Min

и другие.

Nature Communications, Год журнала: 2023, Номер 14(1)

Опубликована: Ноя. 21, 2023

Abstract Learning effective molecular feature representation to facilitate property prediction is of great significance for drug discovery. Recently, there has been a surge interest in pre-training graph neural networks (GNNs) via self-supervised learning techniques overcome the challenge data scarcity prediction. However, current learning-based methods suffer from two main obstacles: lack well-defined strategy and limited capacity GNNs. Here, we propose Knowledge-guided Pre-training Graph Transformer (KPGT), framework alleviate aforementioned issues provide generalizable robust representations. The KPGT integrates transformer specifically designed graphs knowledge-guided strategy, fully capture both structural semantic knowledge molecules. Through extensive computational tests on 63 datasets, exhibits superior performance predicting properties across various domains. Moreover, practical applicability discovery validated by identifying potential inhibitors antitumor targets: hematopoietic progenitor kinase 1 (HPK1) fibroblast growth factor receptor (FGFR1). Overall, can powerful useful tool advancing artificial intelligence (AI)-aided process.

Язык: Английский

Процитировано

37