Method of the Year 2021: Protein structure prediction

Deep Learning based approaches for protein structure prediction have sent shock waves through the structural biology community. We anticipate far-reaching and long-lasting impact.

Nature Methods has named protein structure prediction the Method of the Year 2021.

AlphaFold is a neural-network-based approach to predicting protein structures with high accuracy. We describe how it works in general terms and discuss some anticipated impacts on the field of structural biology.

Deep learning has transformed protein structure modeling. Here we relate AlphaFold and RoseTTAFold to classical physically based approaches to protein structure prediction, and discuss the many areas of structural biology that are likely to be affected by further advances in deep learning.

The greatly improved prediction of protein 3D structure from sequence achieved by the second version of AlphaFold in 2020 has already had a huge impact on biological research, but challenges remain; the protein folding problem cannot be considered solved. We expect fierce competition to improve the method even further and new applications of machine learning to help illuminate proteomes and their many interactions.

The release of protein structure predictions from AlphaFold will increase the number of protein structural models by almost three orders of magnitude. Structural biology and bioinformatics will never be the same, and the need for incisive experimental approaches will be greater than ever. Combining these advances in structure prediction with recent advances in cryo-electron microscopy suggests a new paradigm for structural biology.

The splendid computational success of AlphaFold and RoseTTAFold in solving the 60-year-old problem of protein folding raises an obvious question: what new avenues should structural biology explore? We propose a strong pivot toward the goal of reading mechanism and function directly from the amino acid sequence. This ambitious goal will require new data analytical tools and an extensive database of the atomic-level structural trajectories traced out on energy landscapes as proteins perform their function.

Lineage tracing methods help unravel complexities of cell state and differentiation programs.

The diverse microbial world provides unprecedented potential for exploring genome editing tools.

High-throughput and ultrasensitive mass spectrometry–based approaches are gaining ground in subcellular compositional analysis.

Viral tools are a mainstay in mammalian neuroscience, but there is room for improvement.

Method development is indispensable for obtaining complete genomes, and deciphering them.

Emerging algorithms are extracting information about macromolecular motions from cryo-EM data.

Advances in label-free microscopy are expanding experimental observations, from biophysics to cell biology and beyond.

Probing spatially organized DNA and its interacting elements in single cells will deepen our understanding of cell-type-specific gene regulation.

AlphaFold predicts protein structures with an accuracy competitive with experimental structures in the majority of cases using a novel deep learning architecture.

AlphaFold is used to predict the structures of almost all of the proteins in the human proteome—the availability of high-confidence predicted structures could enable new avenues of investigation from a structural perspective.

AlphaFold predicts the distances between pairs of residues, is used to construct potentials of mean force that accurately describe the shape of a protein and can be optimized with gradient descent to predict protein structures.

The trRosetta neural network was used to iteratively optimise model proteins from random 100-amino-acid sequences, resulting in ‘hallucinated’ proteins, which when expressed in bacteria closely resembled the model structures.

The trRosetta server, a web-based platform for fast and accurate protein structure prediction, is powered by deep learning and Rosetta. This protocol includes procedures for using the web-based server as well as the standalone package.

This Perspective reviews tools developed over the past five years in the macromolecular modeling, docking and design software Rosetta.

Protein structures predicted using artificial intelligence will aid medical research, but the greatest benefit will come if clinical data can be similarly used to better understand human disease.

The full might of a world-leading artificial-intelligence laboratory has been brought to bear on protein-structure prediction. The resulting method, AlphaFold2, promises to transform our understanding of proteins.

Two threads of research in the quest for methods that predict the 3D structures of proteins from their amino-acid sequences have become fully intertwined. The result is a leap forward in the accuracy of predictions.

Machine learning algorithms are fast surpassing human abilities in multiple tasks, from image recognition to medical diagnostics. Now, machine learning algorithms have been shown to be capable of accurately predicting the folded structures of proteins.

Recent advances in AI-based 3D protein structure prediction could help address health-related questions, but may also have far-reaching implications for evolution. Here we discuss the advantages and limitations of high-quality 3D structural predictions by AlphaFold2 in unravelling the relationship between protein properties and their impact on fitness, and emphasize the need to integrate in silico structural predictions with functional genomic studies.

Machine learning will transform our understanding of protein folding. And it’s essential that all data be open.

A potential breakthrough in our ability to predict protein structure from sequence could have a major impact on evolutionary biology.

Deep-learning algorithms such as AlphaFold2 and RoseTTAFold can now predict a protein’s 3D shape from its linear sequence — a huge boon to structural biologists.

AlphaFold neural network produced a ‘totally transformative’ database of more than 350,000 structures from Homo sapiens and 20 model organisms.

Machine-learning systems from the company and from a rival academic group are now open source and freely accessible.

AlphaFold and RoseTTAFold have delivered a revolutionary advance for protein structure predictions, but the implications for drug discovery are more incremental. For now.

Google’s deep-learning program for determining the 3D shapes of proteins stands to transform biology, say scientists.