Structure-based <i>de novo</i> drug design using 3D deep generative models

Li, Yibo; Pei, Jianfeng; Lai, Luhua

doi:10.1039/d1sc04444c

Cited by 111 publications

(117 citation statements)

References 52 publications

Supporting

Mentioning

117

Contrasting

Order By: Relevance

“…The current models mostly generate 1D or 2D molecules, with less exploration of generating 3D molecules. Although recently proposed neural networks such as L-Net [ 18 ] and ConfVAE [ 19 ] represent good attempts for generating 3D molecules, generating small molecules with activity directly for the target binding site remains a challenge. DL has unique advantages in acquiring small molecule features and balancing the effectiveness between targets, which can help medicinal chemists design multitargeted small molecules.…”

mentioning

confidence: 99%

Design of SARS-CoV-2 Mpro, PLpro Dual-Target Inhibitors Based on Deep Reinforcement Learning and Virtual Screening

et al. 2022

View full text Add to dashboard Cite

Background: Since December 2019, SARS-CoV-2 has continued to spread rapidly around the world. The effective drugs may provide a long-term strategy to combat this virus. The main protease (Mpro) and papain-like protease (PLpro) are two important targets for the inhibition of SARS-CoV-2 virus replication and proliferation. Materials & methods: In this study, deep reinforcement learning, covalent docking and molecular dynamics simulations were used to identify novel compounds that have the potential to inhibit both Mpro and PLpro. Results and conclusion: Three compounds were identified that can effectively occupy the Mpro protein cavity with the PLpro protein cavity and form high frequency contacts with key amino acid residues (Mpro: His41, Cys145, Glu166, PLpro: Cys111). These three compounds can be further investigated as potential lead compounds for SARS-CoV-2 inhibitors.

show abstract

mentioning

confidence: 99%

Design of SARS-CoV-2 Mpro, PLpro Dual-Target Inhibitors Based on Deep Reinforcement Learning and Virtual Screening

et al. 2022

View full text Add to dashboard Cite

show abstract

“…To process the lead optimization task more efficiently, many machine learning approaches have been studied recently: (i) atom modification reinforcement learning models that add or delete atoms or bonds [90,91], (ii) generative reinforcement learning which generates similar but modified structure [92], (iii) generative machine learning with controlled chemical properties that also generates similar modified structure with preserved predictive properties [93], and (iv) a 3D structure-based ligand design model that uses a 3D crystal structure of protein and ligand to generate novel molecules [94]. These applications demonstrated the possibility of lead optimization or de novo small molecule-focused machine learning methods.…”

Section: Future Perspectivementioning

confidence: 99%

A Brief Review of Machine Learning-Based Bioactive Compound Research

et al. 2022

View full text Add to dashboard Cite

Bioactive compounds are often used as initial substances for many therapeutic agents. In recent years, both theoretical and practical innovations in hardware-assisted and fast-evolving machine learning (ML) have made it possible to identify desired bioactive compounds in chemical spaces, such as those in natural products (NPs). This review introduces how machine learning approaches can be used for the identification and evaluation of bioactive compounds. It also provides an overview of recent research trends in machine learning-based prediction and the evaluation of bioactive compounds by listing real-world examples along with various input data. In addition, several ML-based approaches to identify specific bioactive compounds for cardiovascular and metabolic diseases are described. Overall, these approaches are important for the discovery of novel bioactive compounds and provide new insights into the machine learning basis for various traditional applications of bioactive compound-related research.

show abstract

“…To be more specific, given the 3D binding pocket of the target protein, these models are aware of the geometric information of the 3D pocket and generate molecules to bind to the pockets accordingly. Early approaches modify the pocket-free models by integrating evaluation functions like docking scores between sampled molecules and pockets to guide the candidate searching (Li et al, 2021). Another types of models transform the 3D pocket structures to molecular SMILES strings or 2D molecular graph (Skalic et al, 2019;Xu et al, 2021) without modeling the interactions between the small molecular structures and 3D pockets explicitly.…”

Section: Introductionmentioning

confidence: 99%

“…Li et al (2021) proposed a ligand neural network to generate 3D molecular structures and leverage Monte Carlo Tree Search to optimize candidate molecules binding to a specific pocket. Instead of involving 3D pockets in the training process, the function of 3D pockets is to evaluate the quality of drug candidates and guide the generation process.…”

mentioning

confidence: 99%

Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pockets

Peng¹,

Luo²,

Guan³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep generative models have achieved tremendous success in designing novel drug molecules in recent years. A new thread of works have shown the great potential in advancing the specificity and success rate of in silico drug design by considering the structure of protein pockets. This setting posts fundamental computational challenges in sampling new chemical compounds that could satisfy multiple geometrical constraints imposed by pockets. Previous sampling algorithms either sample in the graph space or only consider the 3D coordinates of atoms while ignoring other detailed chemical structures such as bond types and functional groups. To address the challenge, we develop Pocket2Mol, an E(3)-equivariant generative network composed of two modules: 1) a new graph neural network capturing both spatial and bonding relationships between atoms of the binding pockets and 2) a new efficient algorithm which samples new drug candidates conditioned on the pocket representations from a tractable distribution without relying on MCMC. Experimental results demonstrate that molecules sampled from Pocket2Mol achieve significantly better binding affinity and other drug properties such as druglikeness and synthetic accessibility.

show abstract

Structure-based de novo drug design using 3D deep generative models

Abstract: Deep generative models are gaining much attention in the field of de novo molecule design. Compared to traditional methods, deep generative models can be trained in a fully data-driven way...

Cited by 111 publications

References 52 publications

Design of SARS-CoV-2 Mpro, PLpro Dual-Target Inhibitors Based on Deep Reinforcement Learning and Virtual Screening

Design of SARS-CoV-2 Mpro, PLpro Dual-Target Inhibitors Based on Deep Reinforcement Learning and Virtual Screening

A Brief Review of Machine Learning-Based Bioactive Compound Research

Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pockets

Contact Info

Product

Resources

About