Abstract:Recently, molecular generation models based on deep learning have attracted significant attention in drug discovery. However, most existing molecular generation models have serious limitations in the context of drug design wherein they do not sufficiently consider the effect of the three-dimensional (3D) structure of the target protein in the generation process. In this study, we developed a new deep learning-based molecular generator, SBMolGen, that integrates a recurrent neural network, a Monte Carlo tree search, and docking simulations. The results of an evaluation using four target proteins (two kinases and two G protein-coupled receptors) showed that the generated molecules had a better binding affinity score (docking score) than the known active compounds, and the generated molecules possessed a broader chemical space distribution. SBMolGen not only generates novel binding active molecules but also presents 3D docking poses with target proteins, which will be useful in subsequent drug design. The code is available at <a class="extLink" href="https://github.com/clinfo/SBMolGen">https://github.com/clinfo/SBMolGen</a>.The Supporting Information is available free of charge at <a class="ext-link" href="/doi/10.1021/acs.jcim.1c00679?goto=supporting-info">https://pubs.acs.org/doi/10.1021/acs.jcim.1c00679</a>.Number of generated molecules, example of generated molecules for target EGFR at C = 1.0, example of generated molecules for target AA2AR at C = 1.0, example of generated molecules for target ADRB2 at C = 1.0; distribution of the similarity score of the generated molecules and distribution of the scaffold similarity score of the generated molecules; distribution of the SA scores of the generated molecules; distributions of the QED scores of the generated molecules; plots of IFIE<sub>sum</sub> against experimentally determined Δ<i>G</i> for (A) nine CDK2 inhibitors and (B) generated molecules; distribution of the docking scores of the generated molecules against DDR1; chemical structures of nine CDK2 inhibitors and their properties (<a class="ext-link" href="/doi/suppl/10.1021/acs.jcim.1c00679/suppl_file/ci1c00679_si_001.pdf">PDF</a>)This article has not yet been cited by other publications.

A Knowledge-Driven Self-Supervised Approach for Molecular Generation

An effective self-supervised framework for learning expressive molecular global representations to drug discovery

Molecular Graph Generation with Deep Reinforced Multitask Network and Adversarial Imitation Learning

Motif-based Graph Self-Supervised Learning for Molecular Property Prediction

Advances in Deep Learning-Based 3D Molecular Generative Models

Chemical-Reaction-Aware Molecule Representation Learning

Learn molecular representations from large-scale unlabeled molecules for drug discovery

A knowledge-guided pre-training framework for improving molecular representation learning

Multi-label Metabolic Pathway Prediction with Auto Molecular Structure Representation Learning

MultiModal-Learning for Predicting Molecular Properties: A Framework Based on Image and Graph Structures

Training recurrent neural networks as generative neural networks for molecular structures: how does it impact drug discovery?

Molecular Graph Representation Learning via Structural Similarity Information

Graph-based Molecular Representation Learning

ASGN: An Active Semi-supervised Graph Neural Network for Molecular Property Prediction

Knowledge-aware Contrastive Molecular Graph Learning

ChemRL-GEM: Geometry Enhanced Molecular Representation Learning for Property Prediction

Learning Neural Generative Dynamics for Molecular Conformation Generation

Structure-Based de Novo Molecular Generator Combined with Artificial Intelligence and Docking Simulations

Molecular Joint Representation Learning via Multi-modal Information of SMILES and Graphs