Latent Molecular Optimization for Targeted Therapeutic Design

Abstract

We devise an approach for targeted molecular design, a problem of interest in computational drug discovery: given a target protein site, we wish to generate a chemical with both high binding affinity to the target and satisfactory pharmacological properties. This problem is made difficult by the enormity and discreteness of the space of potential therapeutics, as well as the graph-structured nature of biomolecular surface sites. Using a dataset of protein-ligand complexes, we surmount these issues by extracting a signature of the target site with a graph convolutional network and by encoding the discrete chemical into a continuous latent vector space. The latter embedding permits gradient-based optimization in molecular space, which we perform using learned differentiable models of binding affinity and other pharmacological properties. We show that our approach is able to efficiently optimize these multiple objectives and discover new molecules with potentially useful binding properties, validated via docking methods.

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Semi-Supervised Classification with Graph Convolutional Networks
• Learning to Plan Chemical Syntheses
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Syntax-Directed Variational Autoencoder for Structured Data
• Grammar Variational Autoencoder
• ChemGAN challenge for drug discovery: can AI reproduce natural chemical diversity?
• Toxicity Prediction using Deep Learning
• Constrained Bayesian Optimization for Automatic Chemical Design
• Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control

Cited by in corpus (9)

• Autonomous discovery in the chemical sciences part II: Outlook
• We Should at Least Be Able to Design Molecules That Dock Well
• Constrained Bayesian Optimization for Automatic Chemical Design
• All SMILES Variational Autoencoder
• Controllable Data Generation by Deep Learning: A Review
• DeepScaffold: a comprehensive tool for scaffold-based de novo drug discovery using deep learning
• Achieving Robustness to Aleatoric Uncertainty with Heteroscedastic Bayesian Optimisation
• Leveraging binding-site structure for drug discovery with point-cloud methods
• PaccMann${}^{RL}$ on SARS-CoV-2: Designing antiviral candidates with conditional generative models