Accelerated rational PROTAC design via deep learning and molecular simulations

Zheng, Shuangjia; Tan, Youhai; Wang, Zhenyu; Li, Chengtao; Zhang, Zhiqing; Sang, Xu; Chen, Hongming; Yang, Yuedong

doi:10.1038/s42256-022-00527-y

Cited by 62 publications

(63 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Yang group proposed a novel deep generative model for the rational design of PROTACs with optimal pharmacokinetics. Importantly, as a proof of concept, they experimentally tested 6 PROTACs, and 1 candidate demonstrated favorable pharmacokinetics in mice . Therefore, greater expediency and lower cost are achieved for the structure-based iterative design of novel PROTACs.…”

Section: Outlook For Protac Developmentmentioning

confidence: 99%

Advancing Strategies for Proteolysis-Targeting Chimera Design

Zhi

Liu

et al. 2023

J. Med. Chem.

View full text Add to dashboard Cite

Proteolysis-targeting chimeras (PROTACs) have shown great therapeutic potential by degrading various disease-causing proteins, particularly those related to tumors. Therefore, the introduction of PROTACs has ushered in a new chapter of antitumor drug development, marked by significant advances over recent years. Herein, we describe recent developments in PROTAC technology, focusing on design strategy, development workflow, and future outlooks. We also discuss potential opportunities and challenges for PROTAC research.

show abstract

Section: Outlook For Protac Developmentmentioning

confidence: 99%

Advancing Strategies for Proteolysis-Targeting Chimera Design

Zhi

Liu

et al. 2023

J. Med. Chem.

View full text Add to dashboard Cite

show abstract

“…Therefore, it's urgent to discover new methods to improve the discovery efficiency of PROTACs. To accelerate the design progress of rational PROTACs, Zheng et al created a novel depthgenerating model (PROTAC-RL) [179]. A pair of E3 ligands and warheads are input into the model, and the designed linkers are output along with chemically feasible PROTACs having specific properties under the guidance of Reinforcement Learning (RL) [179].…”

Section: Computer Simulation Accelerates Protac Designmentioning

confidence: 99%

“…To accelerate the design progress of rational PROTACs, Zheng et al created a novel depthgenerating model (PROTAC-RL) [179]. A pair of E3 ligands and warheads are input into the model, and the designed linkers are output along with chemically feasible PROTACs having specific properties under the guidance of Reinforcement Learning (RL) [179]. Specifically, they first pre-trained a linker generation model (Proformer) based on transformer neural network.…”

Section: Computer Simulation Accelerates Protac Designmentioning

confidence: 99%

An overview of PROTACs: a promising drug discovery paradigm

et al. 2022

Mol Biomed

108

View full text Add to dashboard Cite

Proteolysis targeting chimeras (PROTACs) technology has emerged as a novel therapeutic paradigm in recent years. PROTACs are heterobifunctional molecules that degrade target proteins by hijacking the ubiquitin–proteasome system. Currently, about 20–25% of all protein targets are being studied, and most works focus on their enzymatic functions. Unlike small molecules, PROTACs inhibit the whole biological function of the target protein by binding to the target protein and inducing subsequent proteasomal degradation. PROTACs compensate for limitations that transcription factors, nuclear proteins, and other scaffolding proteins are difficult to handle with traditional small-molecule inhibitors. Currently, PROTACs have successfully degraded diverse proteins, such as BTK, BRD4, AR, ER, STAT3, IRAK4, tau, etc. And ARV-110 and ARV-471 exhibited excellent efficacy in clinical II trials. However, what targets are appropriate for PROTAC technology to achieve better benefits than small-molecule inhibitors are not fully understood. And how to rationally design an efficient PROTACs and optimize it to be orally effective poses big challenges for researchers. In this review, we summarize the features of PROTAC technology, analyze the detail of general principles for designing efficient PROTACs, and discuss the typical application of PROTACs targeting different protein categories. In addition, we also introduce the progress of relevant clinical trial results of representative PROTACs and assess the challenges and limitations that PROTACs may face. Collectively, our studies provide references for further application of PROTACs.

show abstract

“…To identify or design linkers from a larger space, particularly with desired druggable properties, many deep generative models have been proposed for solving the structure generation challenge in drug design. − Though earlier studies were devoted to de novo molecular design, − lead optimization, − and prediction of molecular properties, − they are moving to fragment linking. − Although directly generating linkers in 3D space is still a challenge, a 2D generative model could be an alternative solution for its influence on the 3D structure. For example, Imrie and co-workers reported a graph-based deep generator DeLinker to link fragments.…”

Section: Introductionmentioning

confidence: 99%

DRlinker: Deep Reinforcement Learning for Optimization in Fragment Linking Design

Tan

Dai

Huang

et al. 2022

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

Fragment-based drug discovery is a widely used strategy for drug design in both academic and pharmaceutical industries. Although fragments can be linked to generate candidate compounds by the latest deep generative models, generating linkers with specified attributes remains underdeveloped. In this study, we presented a novel framework, DRlinker, to control fragment linking toward compounds with given attributes through reinforcement learning. The method has been shown to be effective for many tasks from controlling the linker length and log P, optimizing predicted bioactivity of compounds, to various multiobjective tasks. Specifically, our model successfully generated 91.0% and 93.9% of compounds complying with the desired linker length and log P and improved the 7.5 pChEMBL value in bioactivity optimization. Finally, a quasi-scaffold-hopping study revealed that DRlinker could generate nearly 30% molecules with high 3D similarity but low 2D similarity to the lead inhibitor, demonstrating the benefits and applicability of DRlinker in actual fragment-based drug design.

show abstract

Accelerated rational PROTAC design via deep learning and molecular simulations

Cited by 62 publications

References 61 publications

Advancing Strategies for Proteolysis-Targeting Chimera Design

Advancing Strategies for Proteolysis-Targeting Chimera Design

An overview of PROTACs: a promising drug discovery paradigm

DRlinker: Deep Reinforcement Learning for Optimization in Fragment Linking Design

Contact Info

Product

Resources

About