A systematic study of key elements underlying molecular property prediction DOI Creative Commons
Jianyuan Deng, Zhibo Yang, Hehe Wang

et al.

Nature Communications, Journal Year: 2023, Volume and Issue: 14(1)

Published: Oct. 13, 2023

Artificial intelligence (AI) has been widely applied in drug discovery with a major task as molecular property prediction. Despite booming techniques representation learning, key elements underlying prediction remain largely unexplored, which impedes further advancements this field. Herein, we conduct an extensive evaluation of representative models using various representations on the MoleculeNet datasets, suite opioids-related datasets and two additional activity from literature. To investigate predictive power low-data high-data space, series descriptors varying sizes are also assembled to evaluate models. In total, have trained 62,820 models, including 50,220 fixed representations, 4200 SMILES sequences 8400 graphs. Based experimentation rigorous comparison, show that learning exhibit limited performance most datasets. Besides, multiple can affect results. Furthermore, cliffs significantly impact model Finally, explore into potential causes why fail dataset size is essential for excel.

Language: Английский

Automation and computer-assisted planning for chemical synthesis DOI
Yuning Shen, Julia E. Borowski, Melissa A. Hardy

et al.

Nature Reviews Methods Primers, Journal Year: 2021, Volume and Issue: 1(1)

Published: March 18, 2021

Language: Английский

Citations

133
Artificial intelligence for drug discovery: Resources, methods, and applications DOI Creative Commons
Wei Chen, Xuesong Liu, Sanyin Zhang

et al.

Molecular Therapy — Nucleic Acids, Journal Year: 2023, Volume and Issue: 31, P. 691 - 702

Published: Feb. 18, 2023

Conventional wet laboratory testing, validations, and synthetic procedures are costly time-consuming for drug discovery. Advancements in artificial intelligence (AI) techniques have revolutionized their applications to Combined with accessible data resources, AI changing the landscape of In past decades, a series AI-based models been developed various steps These used as complements conventional experiments accelerated discovery process. this review, we first introduced widely resources discovery, such ChEMBL DrugBank, followed by molecular representation schemes that convert into computer-readable formats. Meanwhile, summarized algorithms develop Subsequently, discussed pharmaceutical analysis including predicting toxicity, bioactivity, physicochemical property. Furthermore, de novo design, drug-target structure prediction, interaction, binding affinity prediction. Moreover, also highlighted advanced synergism/antagonism prediction nanomedicine design. Finally, challenges future perspectives on

Language: Английский

Citations

131
Artificial intelligence in drug discovery: applications and techniques DOI
Jianyuan Deng, Zhibo Yang, Iwao Ojima

et al.

Briefings in Bioinformatics, Journal Year: 2021, Volume and Issue: 23(1)

Published: Sept. 21, 2021

Artificial intelligence (AI) has been transforming the practice of drug discovery in past decade. Various AI techniques have used many applications, such as virtual screening and design. In this survey, we first give an overview on discuss related which can be reduced to two major tasks, i.e. molecular property prediction molecule generation. We then present common data resources, representations benchmark platforms. As a part are dissected into model architectures learning paradigms. To reflect technical development over years, surveyed works organized chronologically. expect that survey provides comprehensive review discovery. also provide GitHub repository with collection papers (and codes, if applicable) resource, is regularly updated.

Language: Английский

Citations

114
Small molecules and their impact in drug discovery: A perspective on the occasion of the 125th anniversary of the Bayer Chemical Research Laboratory DOI Creative Commons
Hartmut Beck,

Michael Härter,

B. Hass

et al.

Drug Discovery Today, Journal Year: 2022, Volume and Issue: 27(6), P. 1560 - 1574

Published: Feb. 22, 2022

The year 2021 marks the 125th anniversary of Bayer Chemical Research Laboratory in Wuppertal, Germany. A significant number prominent small-molecule drugs, from Aspirin to Xarelto, have emerged this research site. In review, we shed light on historic cornerstones drug research, discussing current and future trends discovery as well providing a personal outlook with focus small molecules.

Language: Английский

Citations

112
Molecular design in drug discovery: a comprehensive review of deep generative models DOI
Yu Cheng, Yongshun Gong, Yuansheng Liu

et al.

Briefings in Bioinformatics, Journal Year: 2021, Volume and Issue: 22(6)

Published: Aug. 4, 2021

Abstract Deep generative models have been an upsurge in the deep learning community since they were proposed. These are designed for generating new synthetic data including images, videos and texts by fitting approximate distributions. In last few years, shown superior performance drug discovery especially de novo molecular design. this study, reviewed to witness recent advances of design discovery. addition, we divide those into two categories based on representations silico. Then these classical types reported detail discussed about both pros cons. We also indicate current challenges De automatically is promising but a long road be explored.

Language: Английский

Citations

108
A review of graph neural networks: concepts, architectures, techniques, challenges, datasets, applications, and future directions DOI Creative Commons

Bharti Khemani,

Shruti Patil, Ketan Kotecha

et al.

Journal Of Big Data, Journal Year: 2024, Volume and Issue: 11(1)

Published: Jan. 16, 2024

Abstract Deep learning has seen significant growth recently and is now applied to a wide range of conventional use cases, including graphs. Graph data provides relational information between elements standard format for various machine deep tasks. Models that can learn from such inputs are essential working with graph effectively. This paper identifies nodes edges within specific applications, as text, entities, relations, create structures. Different applications may require neural network (GNN) models. GNNs facilitate the exchange in graph, enabling them understand dependencies edges. The delves into GNN models like convolution networks (GCNs), GraphSAGE, attention (GATs), which widely used today. It also discusses message-passing mechanism employed by examines strengths limitations these different domains. Furthermore, explores diverse GNNs, datasets commonly them, Python libraries support offers an extensive overview landscape research its practical implementations.

Language: Английский

Citations

108
Generative machine learning for de novo drug discovery: A systematic review DOI
Dominic D. Martinelli

Computers in Biology and Medicine, Journal Year: 2022, Volume and Issue: 145, P. 105403 - 105403

Published: March 13, 2022

Language: Английский

Citations

99
Artificial intelligence to bring nanomedicine to life DOI
Nikita Serov, Vladimir V. Vinogradov

Advanced Drug Delivery Reviews, Journal Year: 2022, Volume and Issue: 184, P. 114194 - 114194

Published: March 10, 2022

Language: Английский

Citations

96
Extending machine learning beyond interatomic potentials for predicting molecular properties DOI
Nikita Fedik, R.I. Zubatyuk, Maksim Kulichenko

et al.

Nature Reviews Chemistry, Journal Year: 2022, Volume and Issue: 6(9), P. 653 - 672

Published: Aug. 25, 2022

Language: Английский

Citations

94
New Opportunity: Machine Learning for Polymer Materials Design and Discovery DOI
Pengcheng Xu,

Huimin Chen,

Minjie Li

et al.

Advanced Theory and Simulations, Journal Year: 2022, Volume and Issue: 5(5)

Published: Feb. 12, 2022

Abstract Under the guidance of material genome initiative (MGI), use data‐driven methods to discover new materials has become an innovation science. The polymer have been one most important parts in science for excellent physical and chemical properties as well corresponding complex structures. Machine learning, core methods, taken place design discovery. In this review, authors introduced applications machine learning discovery materials. development tendency published papers about materials, commonly used algorithms, descriptors, workflow recent progresses are summarized. Then, detail how assist is fully discussed combined with two cases. Finally, opportunities challenges on future prospects field proposed.

Language: Английский

Citations

82