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

By Stephan Eismann, Raphael J. L. Townshend, Nathaniel Thomas, Milind Jagota, Bowen Jing, Ron O. Dror. 2020

68 citations recorded by the source corpus.

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 predefined 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 precomputed 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.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology, 3D Vision & Pose, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Materials Chemistry

In Molecular Biology

In 3D Vision & Pose

In Computational Biology & Bioinformatics

Citations in the corpus

EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction (2022, methodology): “…other protein representations have been proposed for (graph) DL methods for individual structure prediction (Jing et al., 2020), protein-protein interactions (Dai & Bailey-Kellogg, 2021; Eismann et al., 2020; Townshend et al., 2019), or protein function prediction (Gligorijevi´c et al., 2021).”
Learning from Protein Structure with Geometric Vector Perceptrons (2020, methodology): “These methods allow for equivariant representations of higher-order tensors, but due to their complexity and computational cost, their applications have until recently been limited to small molecules (Eismann et al., 2020).”
Equivariant Graph Neural Networks for 3D Macromolecular Structure (2021, unknown): “These networks have shown promising results on small molecule datasets (Thomas et al., 2018; Kondor, 2018; Anderson et al., 2019; Fuchs et al., 2020; Batzner et al., 2021) and recently also in the context of learning from macromolecular structure (Eismann et al., 2020a;b; Townshend et al., 2020).”
Geometric Prediction: Moving Beyond Scalars (2020, background): “There has recently been significant excitement around equivariant networks that have the ability to reason about arbitrary tensors [3], [16], [4], [5], [17], [18], and though they do not predict tensors, such methods have demonstrated significant results on scalar prediction problems in medical imag”
Independent SE(3)-Equivariant Models for End-to-End Rigid Protein Docking (2021, methodology): “Motivated by symmetries naturally occurring in different data types, architectures are tailored to explicitly incorporate such properties (Cohen and Welling, 2016a;b; Thomas et al., 2018; Fuchs et al., 2020; Finzi et al., 2020; Eismann et al., 2020; Satorras et al., 2021).”