Highly accurate protein structure prediction with AlphaFold
By John M. Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin Žídek, Anna Potapenko and 24 more. 2021
37,531 citations recorded by the source corpus.
Proteins are essential to life, and understanding their structure can facilitate a mechanistic understanding of their function. Through an enormous experimental effort1–4, the structures of around 100,000 unique proteins have been determined5, but this represents a small fraction of the billions of known protein sequences6,7. Structural coverage is bottlenecked by the months to years of painstaking effort required to determine a single protein structure. Accurate computational approaches are needed to address this gap and to enable large-scale structural bioinformatics. Predicting the three-dimensional structure that a protein will adopt based solely on its amino acid sequence—the structure prediction component of the ‘protein folding problem’8—has been an important open research problem for more than 50 years9. Despite recent progress10–14, existing methods fall far short of atomic accuracy, especially when no homologous structure is available. Here we provide the first computational method that can regularly predict protein structures with atomic accuracy even in cases in which no similar structure is known. We validated an entirely redesigned version of our neural network-based model, AlphaFold, in the challenging 14th Critical Assessment of protein Structure Prediction (CASP14)15, demonstrating accuracy competitive with experimental structures in a majority of cases and greatly outperforming other methods. Underpinning the latest version of AlphaFold is a novel machine learning approach that incorporates physical and biological knowledge about protein structure, leveraging multi-sequence alignments, into the design of the deep learning algorithm. AlphaFold predicts protein structures with an accuracy competitive with experimental structures in the majority of cases using a novel deep learning architecture.
Subfields: Materials Chemistry, Molecular Biology, 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 Materials Chemistry
- Yang Zhang
- Jinbo Xu
- David Baker
- Andriy Kryshtafovych
- Torsten Schwede
- Jianlin Cheng
- Robin Pearce
- Sheng Wang
In Molecular Biology
- Yang Zhang
- Jinbo Xu
- David Baker
- Robin Pearce
- David T. Jones
- Sheng Wang
- Björn Wallner
- Torsten Schwede
In Computational Biology & Bioinformatics
- Yang Zhang
- Jinbo Xu
- Sheng Wang
- David T. Jones
- Arne Elofsson
- Björn Wallner
- Robin Pearce
- Burkhard Rost
Citations in the corpus
- CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis (2022, methodology): “It has been successfully applied to natural language processing (Devlin et al., 2019; Lewis et al., 2020; Raffel et al., 2020), computer vision (Dosovitskiy et al., 2021), and many other areas (Oord et al., 2018; Jumper et al., 2021).”
- Sustainable AI: Environmental Implications, Challenges and Opportunities (2021, background): “It has the potential to revolutionize biological sciences by providing in-silico methods for tasks only possible in a physical laboratory setting [1].”
- Galactica: A Large Language Model for Science (2022, methodology): “In recent years, deep learning has been used to represent Nature, such as proteins and molecules (Jumper et al., 2021; Ross et al., 2021).”
- Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling (2023, background): “…made big splashes as generative models in areas as diverse as text-to-image synthesis (Ramesh et al., 2022; Crowson et al., 2022; Rombach et al., 2022), protein modeling (Jumper et al., 2021; Ahdritz et al., 2022), and computer programming (Chen et al., 2021; Xu et al., 2022; Fried et al., 2022).”
- Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation (2021, methodology): “…models tested on arithmetic tasks (Nogueira et al., 2021, Appendix C), models trained with reinforcement learning (Lampinen et al., 2021), image, speech recognition, and machine translation models (Likhomanenko et al., 2021), and protein structure prediction (Jumper et al., 2021, Appendix 1.5).”
- Learning to Learn with Generative Models of Neural Network Checkpoints (2022, neutral): “…language modeling (Radford et al., 2018; 2019; Brown et al., 2020), coding (Chen et al., 2021b; Li et al., 2022), , reinforcement learning (Chen et al., 2021a; Janner et al., 2021), image synthesis (Esser et al., 2020; Ramesh et al., 2021; Yu et al., 2022) and protein folding (Jumper et al., 2021).”
- Robust deep learning based protein sequence design using ProteinMPNN (2022, positive): “…passing neural network (MPNN) with 3 encoder and 3 decoder layers and 128 hidden dimensions which predicts protein sequences in an autoregressive manner from N to C terminus using protein backbone features – distances between Ca-Ca atoms, relative Ca-Ca-Ca frame orientations and rotations,…”
- Generalized Biomolecular Modeling and Design with RoseTTAFold All-Atom (2023, positive): “InAF2andRF2,thesequence orderofaminoacidsandbasesisrepresented byarelativepositionencoding;foratoms,we omittedsuchanencodingandleveragedthe permutationinvarianceofthenetworkatten-tionmechanisms.”
- AFsample: improving multimer prediction with AlphaFold using massive sampling (2023, methodology): “Given enough evolutionary-related sequences, AF2 predicts monomers with very high accuracy even without using structural templates (Jumper et al. 2021).”
- Critical assessment of protein intrinsic disorder prediction (CAID) ‐ Results of round 2 (2023, methodology): “To establish a comparison baseline, we also incorporated the AlphaFold-disorder package, 9 which utilizes AlphaFold2 predicted structures publicly available in databases 10,11 to infer predictions for disorder and binding.”
- In silico protein interaction screening uncovers DONSON’s role in replication initiation (2023, methodology): “To discover novel DONSON interactors within DNA replication pathways, we performed an in silico screen using the AF-M program developed by DeepMind ( 35 , 61 ).”
- Programmable protein delivery with a bacterial contractile injection system (2023, positive): “We used AlphaFold 35–37 to predict the 3D structure of the putative distal tip of Pvc13—the region that we predicted would make the initial contact with target cells (Fig.”
- AlphaFold2 and its applications in the fields of biology and medicine (2023, methodology): “12 At present, there are many modeling methods based on DL, among which AlphaFold, 1,47 RoseTTAFold, 48 ESMFold 49 (ESMFold also offers an extensive database of protein structural predictions, which include 617 million metagenomic protein structures) and the recent language model by Chowdhury et al.”
- End-to-end protein–ligand complex structure generation with diffusion-based generative models (2022, methodology): “Recent studies in protein structure prediction have reported improvements in prediction accuracy and generalizability through refinement of model architecture [17], training data [47], and scaling of pre-trained protein language models [18].”
- Evolutionary-scale prediction of atomic level protein structure with a language model (2022, background): “[12] John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin ŽÂıdek, Anna Potapenko, Alex Bridgland, Clemens Meyer, Simon A.”
- TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction (2022, methodology): “Physically, the inter-molecular trigonometry module, inspired by the intra-molecular Evoformer module used in AlphaFold2 Jumper et al. [2021], ensures that our energy landscape disfavors configurations of protein-ligand complexes that are prohibited by laws of nature, for instance, no two atoms coul”
- DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking (2022, methodology): “For example, the pLDDT confidence score of AlphaFold2 [Jumper et al., 2021] has had a very significant impact in many applications [Necci et al.”
- State-specific protein–ligand complex structure prediction with a multiscale deep generative model (2022, positive): “…and associated pair representations using an attention-based network, motivated by recent vision-language models and folding prediction networks 1,38 ; and (3) an equivariant structure denoising module (ESDM) to generate the binding complex atomistic structure conditioned on outputs from the…”