Learning from Protein Structure with Geometric Vector Perceptrons

TL;DR

This paper introduces geometric vector perceptrons, a novel neural network layer that effectively captures geometric and relational information in 3D protein structures, improving performance on protein quality assessment and design tasks.

Contribution

The paper presents geometric vector perceptrons, extending dense layers to handle Euclidean vectors, enabling neural networks to better learn from 3D biomolecular structures.

Findings

Improved accuracy over existing methods in protein quality assessment.

Enhanced performance in computational protein design tasks.

Demonstrated effectiveness on two key protein learning problems.

Abstract

Learning on 3D structures of large biomolecules is emerging as a distinct area in machine learning, but there has yet to emerge a unifying network architecture that simultaneously leverages the graph-structured and geometric aspects of the problem domain. To address this gap, we introduce geometric vector perceptrons, which extend standard dense layers to operate on collections of Euclidean vectors. Graph neural networks equipped with such layers are able to perform both geometric and relational reasoning on efficient and natural representations of macromolecular structure. We demonstrate our approach on two important problems in learning from protein structure: model quality assessment and computational protein design. Our approach improves over existing classes of architectures, including state-of-the-art graph-based and voxel-based methods. We release our code at https://github.com/drorlab/gvp.