Deep learning for molecular generation and optimization - a review of the state of the art

  title={Deep learning for molecular generation and optimization - a review of the state of the art},
  author={Daniel C. Elton and Zois Boukouvalas and Mark D. Fuge and Peter W. Chung},
We review a recent groundswell of work which uses deep learning techniques to generate and optimize molecules. 

Figures and Tables from this paper

A Deep Learning Based Scaffold Hopping Strategy for the Design of Kinase Inhibitors
A fragment-based deep learning strategy for scaffold hopping towards the conserved hinge binding motif of kinase inhibitors in a large scale is proposed.
On failure modes in molecule generation and optimization.
A Review of Deep Learning Methods for Antibodies
A brief background of deep learning as it applies to antibody drug development, and an in-depth explanation of several deep learning algorithms that have been proposed to solve aspects of both protein design in general, and antibody design in particular are given.
Deep learning enables rapid identification of potent DDR1 kinase inhibitors
A machine learning model allows the identification of new small-molecule kinase inhibitors in days and is used to discover potent inhibitors of discoidin domain receptor 1 (DDR1), a kinase target implicated in fibrosis and other diseases, in 21 days.
Diversity oriented Deep Reinforcement Learning for targeted molecule generation
The results reveal that the model can effectively adjust the newly generated molecules towards the wanted direction, and it was possible to find promising sets of unique and diverse molecules, which was the main purpose of the newly implemented strategy.
Molecular design in drug discovery: a comprehensive review of deep generative models.
In this study, deep generative models are reviewed to witness the recent advances of de novo molecular design for drug discovery and divided into two categories based on molecular representations in silico.
Generative chemistry: drug discovery with deep learning generative models
The purpose of this paper is to review the latest advances in generative chemistry which relies on generative modeling to expedite the drug discovery process and detailed discussions on utilizing cutting-edge generative architectures are focused.
Scaffold-constrained molecular generation
This work builds on the well-known SMILES-based Recurrent Neural Network (RNN) generative model, with a modified sampling procedure, to achieve scaffold-constrained generation, and directly benefit from the associated reinforcement learning methods, allowing to design molecules optimized for different properties while exploring only the relevant chemical space.
Enabling rapid COVID-19 small molecule drug design through scalable deep learning of generative models
The quality and time to produce machine learned models for use in small molecule antiviral design were improved and the time reduced.


Optimization of Molecules via Deep Reinforcement Learning
Inspired by problems faced during medicinal chemistry lead optimization, the MolDQN model is extended with multi-objective reinforcement learning, which maximizes drug-likeness while maintaining similarity to the original molecule.
Machine learning for molecular and materials science
A future in which the design, synthesis, characterization and application of molecules and materials is accelerated by artificial intelligence is envisaged.
Molecular generative model based on conditional variational autoencoder for de novo molecular design
A molecular generative model based on the conditional variational autoencoder that can be used to generate drug-like molecules with five target properties and to adjust a single property without changing the others and to manipulate it beyond the range of the dataset is proposed.
MoleculeNet: a benchmark for molecular machine learning† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c7sc02664a
A large scale benchmark for molecular machine learning consisting of multiple public datasets, metrics, featurizations and learning algorithms.
De novo drug design with deep generative models : an empirical study
The results suggest that traditional stochastic optimizers, that rely on random perturbations or random sampling, can be improved by using generative models trained on unlabeled data, to perform knowledge-driven optimization.
Application of Generative Autoencoder in De Novo Molecular Design
The results show that the latent space preserves chemical similarity principle and thus can be used for the generation of analogue structures in autoencoder for de novo molecular design.
Shape-Based Generative Modeling for de Novo Drug Design
This work proposes a machine learning approach to generate novel molecules starting from a seed compound, its three-dimensional shape, and its pharmacophoric features, and represents a first example of the de novo design of lead-like molecules guided by shape-based features.
Conditional molecular design with deep generative models
A conditional molecular design method that facilitates generating new molecules with desired properties is presented, built as a semisupervised variational autoencoder trained on a set of existing molecules with only a partial annotation.