3DMolNet: A Generative Network for Molecular Structures

Abstract

With the recent advances in machine learning for quantum chemistry, it is now possible to predict the chemical properties of compounds and to generate novel molecules. Existing generative models mostly use a string- or graph-based representation, but the precise three-dimensional coordinates of the atoms are usually not encoded. First attempts in this direction have been proposed, where autoregressive or GAN-based models generate atom coordinates. Those either lack a latent space in the autoregressive setting, such that a smooth exploration of the compound space is not possible, or cannot generalize to varying chemical compositions. We propose a new approach to efficiently generate molecular structures that are not restricted to a fixed size or composition. Our model is based on the variational autoencoder which learns a translation-, rotation-, and permutation-invariant low-dimensional representation of molecules. Our experiments yield a mean reconstruction error below 0.05 Angstrom, outperforming the current state-of-the-art methods by a factor of four, and which is even lower than the spatial quantization error of most chemical descriptors. The compositional and structural validity of newly generated molecules has been confirmed by quantum chemical methods in a set of experiments.

References in corpus (15)

• Auto-Encoding Variational Bayes
• The information bottleneck method
• SchNet - a deep learning architecture for molecules and materials
• Semi-Supervised Learning with Deep Generative Models
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• Machine learning prediction errors better than DFT accuracy
• Learning Deep Generative Models of Graphs
• Constrained Graph Variational Autoencoders for Molecule Design
• Learning Multimodal Graph-to-Graph Translation for Molecular Optimization
• Multi-Level Variational Autoencoder: Learning Disentangled Representations from Grouped Observations
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• Learning Latent Subspaces in Variational Autoencoders
• Generating valid Euclidean distance matrices
• Adversarial Information Factorization
• Inverse Learning of Symmetries

Cited by in corpus (6)

• A Comprehensive Survey on Deep Graph Representation Learning
• Inverse design of 3d molecular structures with conditional generative neural networks
• ChemSpaceAL: An Efficient Active Learning Methodology Applied to Protein-Specific Molecular Generation
• Ab initio machine learning in chemical compound space
• Learning Conditional Invariance through Cycle Consistency
• MOLUCINATE: A Generative Model for Molecules in 3D Space