Literature DB >> 33494711

DeepDist: real-value inter-residue distance prediction with deep residual convolutional network.

Tianqi Wu1, Zhiye Guo1, Jie Hou2, Jianlin Cheng3.   

Abstract

BACKGROUND: Driven by deep learning, inter-residue contact/distance prediction has been significantly improved and substantially enhanced ab initio protein structure prediction. Currently, most of the distance prediction methods classify inter-residue distances into multiple distance intervals instead of directly predicting real-value distances. The output of the former has to be converted into real-value distances to be used in tertiary structure prediction.
RESULTS: To explore the potentials of predicting real-value inter-residue distances, we develop a multi-task deep learning distance predictor (DeepDist) based on new residual convolutional network architectures to simultaneously predict real-value inter-residue distances and classify them into multiple distance intervals. Tested on 43 CASP13 hard domains, DeepDist achieves comparable performance in real-value distance prediction and multi-class distance prediction. The average mean square error (MSE) of DeepDist's real-value distance prediction is 0.896 Å2 when filtering out the predicted distance ≥ 16 Å, which is lower than 1.003 Å2 of DeepDist's multi-class distance prediction. When distance predictions are converted into contact predictions at 8 Å threshold (the standard threshold in the field), the precision of top L/5 and L/2 contact predictions of DeepDist's multi-class distance prediction is 79.3% and 66.1%, respectively, higher than 78.6% and 64.5% of its real-value distance prediction and the best results in the CASP13 experiment.
CONCLUSIONS: DeepDist can predict inter-residue distances well and improve binary contact prediction over the existing state-of-the-art methods. Moreover, the predicted real-value distances can be directly used to reconstruct protein tertiary structures better than multi-class distance predictions due to the lower MSE. Finally, we demonstrate that predicting the real-value distance map and multi-class distance map at the same time performs better than predicting real-value distances alone.

Entities:  

Keywords:  Contact prediction; Deep learning; Protein distance prediction; Protein structure prediction

Mesh:

Substances:

Year:  2021        PMID: 33494711      PMCID: PMC7831258          DOI: 10.1186/s12859-021-03960-9

Source DB:  PubMed          Journal:  BMC Bioinformatics        ISSN: 1471-2105            Impact factor:   3.169


  31 in total

1.  Protein secondary structure prediction based on position-specific scoring matrices.

Authors:  D T Jones
Journal:  J Mol Biol       Date:  1999-09-17       Impact factor: 5.469

2.  HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignment.

Authors:  Michael Remmert; Andreas Biegert; Andreas Hauser; Johannes Söding
Journal:  Nat Methods       Date:  2011-12-25       Impact factor: 28.547

3.  DeepMSA: constructing deep multiple sequence alignment to improve contact prediction and fold-recognition for distant-homology proteins.

Authors:  Chengxin Zhang; Wei Zheng; S M Mortuza; Yang Li; Yang Zhang
Journal:  Bioinformatics       Date:  2020-04-01       Impact factor: 6.937

4.  Improved protein structure prediction using potentials from deep learning.

Authors:  Andrew W Senior; Richard Evans; John Jumper; James Kirkpatrick; Laurent Sifre; Tim Green; Chongli Qin; Augustin Žídek; Alexander W R Nelson; Alex Bridgland; Hugo Penedones; Stig Petersen; Karen Simonyan; Steve Crossan; Pushmeet Kohli; David T Jones; David Silver; Koray Kavukcuoglu; Demis Hassabis
Journal:  Nature       Date:  2020-01-15       Impact factor: 49.962

5.  Analysis of distance-based protein structure prediction by deep learning in CASP13.

Authors:  Jinbo Xu; Sheng Wang
Journal:  Proteins       Date:  2019-09-13

6.  PconsFold: improved contact predictions improve protein models.

Authors:  Mirco Michel; Sikander Hayat; Marcin J Skwark; Chris Sander; Debora S Marks; Arne Elofsson
Journal:  Bioinformatics       Date:  2014-09-01       Impact factor: 6.937

7.  CCMpred--fast and precise prediction of protein residue-residue contacts from correlated mutations.

Authors:  Stefan Seemayer; Markus Gruber; Johannes Söding
Journal:  Bioinformatics       Date:  2014-07-26       Impact factor: 6.937

8.  Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model.

Authors:  Sheng Wang; Siqi Sun; Zhen Li; Renyu Zhang; Jinbo Xu
Journal:  PLoS Comput Biol       Date:  2017-01-05       Impact factor: 4.475

9.  Deep learning extends de novo protein modelling coverage of genomes using iteratively predicted structural constraints.

Authors:  Joe G Greener; Shaun M Kandathil; David T Jones
Journal:  Nat Commun       Date:  2019-09-04       Impact factor: 14.919

10.  Prediction of interresidue contacts with DeepMetaPSICOV in CASP13.

Authors:  Shaun M Kandathil; Joe G Greener; David T Jones
Journal:  Proteins       Date:  2019-07-27
View more
  18 in total

1.  Multi-head attention-based U-Nets for predicting protein domain boundaries using 1D sequence features and 2D distance maps.

Authors:  Sajid Mahmud; Zhiye Guo; Farhan Quadir; Jian Liu; Jianlin Cheng
Journal:  BMC Bioinformatics       Date:  2022-07-19       Impact factor: 3.307

2.  Study of Real-Valued Distance Prediction for Protein Structure Prediction with Deep Learning.

Authors:  Jin Li; Jinbo Xu
Journal:  Bioinformatics       Date:  2021-05-07       Impact factor: 6.937

3.  Protein model accuracy estimation empowered by deep learning and inter-residue distance prediction in CASP14.

Authors:  Xiao Chen; Jian Liu; Zhiye Guo; Tianqi Wu; Jie Hou; Jianlin Cheng
Journal:  Sci Rep       Date:  2021-05-25       Impact factor: 4.379

4.  High-Performance Deep Learning Toolbox for Genome-Scale Prediction of Protein Structure and Function.

Authors:  Mu Gao; Peik Lund-Andersen; Alex Morehead; Sajid Mahmud; Chen Chen; Xiao Chen; Nabin Giri; Raj S Roy; Farhan Quadir; T Chad Effler; Ryan Prout; Subil Abraham; Wael Elwasif; N Quentin Haas; Jeffrey Skolnick; Jianlin Cheng; Ada Sedova
Journal:  Workshop Mach Learn HPC Environ       Date:  2021-12-27

5.  Combination of deep neural network with attention mechanism enhances the explainability of protein contact prediction.

Authors:  Chen Chen; Tianqi Wu; Zhiye Guo; Jianlin Cheng
Journal:  Proteins       Date:  2021-02-16

6.  Ultrafast end-to-end protein structure prediction enables high-throughput exploration of uncharacterized proteins.

Authors:  Shaun M Kandathil; Joe G Greener; Andy M Lau; David T Jones
Journal:  Proc Natl Acad Sci U S A       Date:  2022-01-25       Impact factor: 12.779

7.  Enhancing protein inter-residue real distance prediction by scrutinising deep learning models.

Authors:  Julia Rahman; M A Hakim Newton; Md Khaled Ben Islam; Abdul Sattar
Journal:  Sci Rep       Date:  2022-01-17       Impact factor: 4.379

8.  MULTICOM2 open-source protein structure prediction system powered by deep learning and distance prediction.

Authors:  Tianqi Wu; Jian Liu; Zhiye Guo; Jie Hou; Jianlin Cheng
Journal:  Sci Rep       Date:  2021-06-23       Impact factor: 4.379

Review 9.  Deep Learning-Based Advances in Protein Structure Prediction.

Authors:  Subash C Pakhrin; Bikash Shrestha; Badri Adhikari; Dukka B Kc
Journal:  Int J Mol Sci       Date:  2021-05-24       Impact factor: 5.923

Review 10.  Recent Advances in Protein Homology Detection Propelled by Inter-Residue Interaction Map Threading.

Authors:  Sutanu Bhattacharya; Rahmatullah Roche; Md Hossain Shuvo; Debswapna Bhattacharya
Journal:  Front Mol Biosci       Date:  2021-05-11
View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.