Improved protein structure refinement guided by deep learning based accuracy estimation

By Naozumi Hiranuma, Hahnbeom Park, Minkyung Baek, Ivan V. Anishchenko, J. Dauparas, David Baker. 2020

230 citations recorded by the source corpus.

We develop a deep learning framework (DeepAccNet) that estimates per-residue accuracy and residue-residue distance signed error in protein models and uses these predictions to guide Rosetta protein structure refinement. The network uses 3D convolutions to evaluate local atomic environments followed by 2D convolutions to provide their global contexts and outperforms other methods that similarly predict the accuracy of protein structure models. Overall accuracy predictions for X-ray and cryoEM structures in the PDB correlate with their resolution, and the network should be broadly useful for assessing the accuracy of both predicted structure models and experimentally determined structures and identifying specific regions likely to be in error. Incorporation of the accuracy predictions at multiple stages in the Rosetta refinement protocol considerably increased the accuracy of the resulting protein structure models, illustrating how deep learning can improve search for global energy minima of biomolecules. Here the authors present DeepAccNet, a deep learning framework that estimates per-residue accuracy and residue-residue distance signed error in protein models, which are used to guide Rosetta protein structure refinement. Benchmarking suggests an improvement of accuracy prediction and refinement compared to other related state of the art methods.

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

1. Robin Pearce
2. Yang Zhang
3. Chengxin Zhang
4. Yang Li
5. Wei Zheng
6. Eric W. Bell
7. Chaok Seok
8. Xiaogen Zhou

In Molecular Biology

1. Robin Pearce
2. Yang Zhang
3. Chengxin Zhang
4. Yang Li
5. Eric W. Bell
6. Wei Zheng
7. Xiaogen Zhou
8. Andriy Kryshtafovych

In Computational Biology & Bioinformatics

1. Robin Pearce
2. Chengxin Zhang
3. Yang Li
4. Wei Zheng
5. Yang Zhang
6. Eric W. Bell
7. Xiaogen Zhou
8. Chaok Seok

Citations in the corpus

• Accurate prediction of protein structures and interactions using a 3-track neural network (2021, positive): “The increased prediction accuracy was critical for success in all cases; models made with trRosetta did not yield MR solutions.”
• AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models (2021, methodology): “Several other protein structure prediction resources also use lDDT-based metrics (15,16).”
• Protein tertiary structure prediction and refinement using deep learning and Rosetta in CASP14 (2021, unknown): “The standard protocol integrates DeepAccNet-MSA into our standard Rosetta refinement protocol, 19 and resulting models were submitted for the group “ BAKER.”
• State-of-the-Art Estimation of Protein Model Accuracy Using AlphaFold (2022, unknown): “We compared AlphaFold’s ability to assess the quality of decoy structures with the Rosetta energy function, as well as DeepAccNet, which is a state-of-the-art machine learning model for estimating the accuracy of protein structure models (17).”
• I-TASSER-MTD: a deep-learning-based platform for multi-domain protein structure and function prediction (2022, methodology): “Users can also estimate the total local distance difference test (lDDT) of the model for reference by using deep-learning based quality assessment methods, such as DeepAccNet 98 and DeepUMQA 99 .”
• Assessment of Protein Model Structure Accuracy Estimation in CASP14: Old and New Challenges (2021, methodology): “The best EMA method in CASP14 was from the Baker group, 11 and it also performed better than other methods in CASP-COVID 19 .”
• Improving de novo protein binder design with deep learning (2022, unknown): “There has also been progress in accuracy prediction for protein structure models; for example Dee-pAccuracyNet (DAN), which uses a representation consisting of 3D convolutions of local atomic environments 20 , achieved state-of-the-art performance in accuracy prediction in CASP14.”