Improved protein structure prediction using potentials from deep learning

Abstract

Protein structure prediction aims to determine the three-dimensional shape of a protein from its amino acid sequence 1. This problem is of fundamental importance to biology as the structure of a protein largely determines its function 2 but can be hard to determine experimentally. In recent years, considerable progress has been made by leveraging genetic information: analysing the co-variation of homologous sequences can allow one to infer which amino acid residues are in contact, which in turn can aid structure prediction 3. In this work, we show that we can train a neural network to accurately predict the distances between pairs of residues in a protein which convey more about structure than contact predictions. With this information we construct a potential of mean force 4 that can accurately describe the shape of a protein. We find that the resulting potential can be optimised by a simple gradient descent algorithm, to realise structures without the need for complex sampling procedures. The resulting system, named AlphaFold, has been shown to achieve high accuracy, even for sequences with relatively few homologous sequences. In the most recent Critical Assessment of Protein Structure Prediction 5 (CASP13), a blind assessment of the state of the field of protein structure prediction, AlphaFold created high-accuracy structures (with TM-scores † of 0.7 or higher) for 24 out of 43 free modelling domains whereas the next best method, using sampling and contact information, achieved such accuracy for only 14 out of 43 domains. AlphaFold represents a significant advance in protein structure prediction. We expect the increased accuracy of structure predictions for proteins to enable insights in understanding the function and malfunction of these proteins, especially in cases where no homologous proteins have been experimentally determined 7. Proteins are at the core of most biological processes. Since the function of a protein is dependent on its structure, understanding protein structure has been a grand challenge in biology for decades. While several experimental structure determination techniques have been developed † Template Modelling score 6 , between 0 and 1, measures the degree of match of the overall (backbone) shape of a proposed structure to a native structure.

Protein structure prediction using multiple deep neural networks in the 13th Critical Assessment of Protein Structure Prediction (CASP13)

“…The distributions are predicted with a deep, dilated residual convolutional network 13 described in detail in another paper. 24 The network consists of 220 two-dimensional (2D) residual blocks with 128 channels and dilated 3 × 3 convolutions, elu nonlinearity with dropout and batch normalization.…”

Modeling of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Proteins by Machine Learning and Physics-Based Refinement

“…However, for some of the proteins, template-based modeling is not possible because of lack of experimentally determined close homologs. Recently, the prediction of tertiary structures for proteins where no template structures are available, has been advanced significantly via novel machine learning methods 3 . This approach predicts interresidue distances from multiple sequence alignment via deep learning.…”

Refinement of protein structure via contact based potentials in replica exchange simulations

“…Computer simulations are commonly used to interpret and complement experimental data. Novel, purely data-driven approaches can predict protein structures of high quality [9,10] but lack insight into the physical processes driving structure adoption and cannot be easily complemented by experimental information. Depending on the method, local structural motifs are often less resolved [9] and could benefit from additional refinement.…”