Hierarchical, rotation-equivariant neural networks to select structural models of protein complexes

Abstract

Predicting the structure of multi-protein complexes is a grand challenge in biochemistry, with major implications for basic science and drug discovery. Computational structure prediction methods generally leverage pre-defined structural features to distinguish accurate structural models from less accurate ones. This raises the question of whether it is possible to learn characteristics of accurate models directly from atomic coordinates of protein complexes, with no prior assumptions. Here we introduce a machine learning method that learns directly from the 3D positions of all atoms to identify accurate models of protein complexes, without using any pre-computed physics-inspired or statistical terms. Our neural network architecture combines multiple ingredients that together enable end-to-end learning from molecular structures containing tens of thousands of atoms: a point-based representation of atoms, equivariance with respect to rotation and translation, local convolutions, and hierarchical subsampling operations. When used in combination with previously developed scoring functions, our network substantially improves the identification of accurate structural models among a large set of possible models. Our network can also be used to predict the accuracy of a given structural model in absolute terms. The architecture we present is readily applicable to other tasks involving learning on 3D structures of large atomic systems.

arXiv comment11 pages, 5 figures + SI: Updated based on the published version in PROTEINS. Presented at NeurIPS 2019 workshop Learning Meaningful Representations of Life

References in corpus (13)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Protein-Ligand Scoring with Convolutional Neural Networks
• Machine learning prediction errors better than DFT accuracy
• Group Equivariant Convolutional Networks
• Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Horovod: fast and easy distributed deep learning in TensorFlow
• SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
• Cormorant: Covariant Molecular Neural Networks
• Deep convolutional networks for quality assessment of protein folds
• N-body Networks: a Covariant Hierarchical Neural Network Architecture for Learning Atomic Potentials
• Energy-based Graph Convolutional Networks for Scoring Protein Docking Models
• End-to-End Learning on 3D Protein Structure for Interface Prediction

Cited by in corpus (5)

• Learning from Protein Structure with Geometric Vector Perceptrons
• Equivariant Graph Neural Networks for 3D Macromolecular Structure
• Protein model quality assessment using rotation-equivariant, hierarchical neural networks
• An In-depth Summary of Recent Artificial Intelligence Applications in Drug Design
• Capacity of Group-invariant Linear Readouts from Equivariant Representations: How Many Objects can be Linearly Classified Under All Possible Views?