Abstract:Generative approaches to molecular design are an area of intense study in recent years as a method to generate new pharmaceuticals with desired properties. Often though, these types of efforts are constrained by limited experimental activity data, resulting in either models that generate molecules with poor performance or models that are overfit and produce close analogs of known molecules. In this paper, we reduce this data dependency for the generation of new chemotypes by incorporating docking scores of known and de novo molecules to expand the applicability domain of the reward function and diversify the compounds generated during reinforcement learning. Our approach employs a deep generative model initially trained using a combination of limited known drug activity and an approximate docking score provided by a second machine learned Bayes regression model, with final evaluation of high scoring compounds by a full docking simulation. This strategy results in molecules with docking scores improved by 10-20% compared to molecules of similar size, while being 130 × faster than a docking only approach on a typical GPU workstation. We also show that the increased docking scores correlate with (1) docking poses with interactions similar to known inhibitors and (2) result in higher MM-GBSA binding energies comparable to the energies of known DDR1 inhibitors, demonstrating that the Bayesian model contains sufficient information for the network to learn to efficiently interact with the binding pocket during reinforcement learning. This outcome shows that the combination of the learned latent molecular representation along with the feature-based docking regression is sufficient for reinforcement learning to infer the relationship between the molecules and the receptor binding site, which suggest that our method can be a powerful tool for the discovery of new chemotypes with potential therapeutic applications.

Bayesian Optimization in the Latent Space of a Variational Autoencoder for the Generation of Selective FLT3 Inhibitors

De novo drug design through gradient-based regularized search in information-theoretically controlled latent space

Pathway-Guided Optimization of Deep Generative Molecular Design Models for Cancer Therapy

Discovery of VEGFR-2 Inhibitors employing Junction Tree Variational Encoder with Local Latent Space Bayesian Optimization and Gradient Ascent Exploration

Constrained Bayesian Optimization for Automatic Chemical Design

Multi-Objective Latent Space Optimization of Generative Molecular Design Models

Latent Molecular Optimization for Targeted Therapeutic Design

Rethinking Molecular Design: Integrating Latent Variable and Auto-Regressive Models for Goal Directed Generation

Balancing Exploration and Exploitation: Disentangled $β$-CVAE in De Novo Drug Design

Leveraging Latent Evolutionary Optimization for Targeted Molecule Generation

Enhancing De Novo Drug Design across Multiple Therapeutic Targets with CVAE Generative Models

Bayesian Optimization on Large Graphs via a Graph Convolutional Generative Model: Application in Cardiac Model Personalization

Latent Space Bayesian Optimization with Latent Data Augmentation for Enhanced Exploration

Leveraging Active Subspaces to Capture Epistemic Model Uncertainty in Deep Generative Models for Molecular Design

High-Dimensional Bayesian Optimisation with Variational Autoencoders and Deep Metric Learning

Analysis of training and seed bias in small molecules generated with a conditional graph-based variational autoencoder -- Insights for practical AI-driven molecule generation

Molecular generative model based on conditional variational autoencoder for de novo molecular design

Generative Model for Small Molecules with Latent Space RL Fine-Tuning to Protein Targets

CLigOpt: controllable ligand design through target-specific optimization

Variational Autoencoding Molecular Graphs with Denoising Diffusion Probabilistic Model

Improving drug discovery with a hybrid deep generative model using reinforcement learning trained on a Bayesian docking approximation