High-resolution de novo structure prediction from primary sequence

By R. Wu, Fan Ding, Rui Wang, Rui Shen, Xiwen Zhang, Shitong Luo, Chenpeng Su, Zuofan Wu, Qi Xie, Bonnie Berger and 2 more. 2022

529 citations recorded by the source corpus.

Recent breakthroughs have used deep learning to exploit evolutionary information in multiple sequence alignments (MSAs) to accurately predict protein structures. However, MSAs of homologous proteins are not always available, such as with orphan proteins or fast-evolving proteins like antibodies, and a protein typically folds in a natural setting from its primary amino acid sequence into its three-dimensional structure, suggesting that evolutionary information and MSAs should not be necessary to predict a protein’s folded form. Here, we introduce OmegaFold, the first computational method to successfully predict high-resolution protein structure from a single primary sequence alone. Using a new combination of a protein language model that allows us to make predictions from single sequences and a geometry-inspired transformer model trained on protein structures, OmegaFold outperforms RoseTTAFold and achieves similar prediction accuracy to AlphaFold2 on recently released structures. OmegaFold enables accurate predictions on orphan proteins that do not belong to any functionally characterized protein family and antibodies that tend to have noisy MSAs due to fast evolution. Our study fills a much-encountered gap in structure prediction and brings us a step closer to understanding protein folding in nature.

Subfields: 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 Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

State-specific protein–ligand complex structure prediction with a multiscale deep generative model (2022, neutral): “On the contrary, 3D coordinates are solely updated for atomic nodes (PredictDrift, Supplementary Algorithm 11), while the rigid-body frames ( t , R ) are passively reconstructed according to the updated atomic coordinates, circumventing numerical issues regarding fitting quaternion or axis-angle var”
Protein structure generation via folding diffusion (2022, methodology): “OmegaFold is designed without reliance on multiple sequence alignments (MSAs), and performs similarly to AlphaFold while generalizing better to orphan proteins that may not have such evolutionary neighbors 48 .”
OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization (2023, unknown): “These include the ESMFold method for protein structure prediction ([54]), which replaces MSAs with protein language models ([55], [56], [57]), and FastFold, a community effort that has implemented significant improvements including fast model-parallel training and inference ([46]).”
Illuminating protein space with a programmable generative model (2022, methodology): “We sought to evaluate the fidelity of sequence–structure pairs generated by Chroma by measuring their agreement with three state-of-the-art methods for structure prediction 9,39,40 .”