Improving prediction of protein secondary structure, backbone angles, solvent accessibility and contact numbers by using predicted contact maps and an ensemble of recurrent and residual convolutional neural networks.

Journal: Bioinformatics (Oxford, England)
Published Date: Jul 15, 2019

Abstract

MOTIVATION: Sequence-based prediction of one dimensional structural properties of proteins has been a long-standing subproblem of protein structure prediction. Recently, prediction accuracy has been significantly improved due to the rapid expansion of protein sequence and structure libraries and advances in deep learning techniques, such as residual convolutional networks (ResNets) and Long-Short-Term Memory Cells in Bidirectional Recurrent Neural Networks (LSTM-BRNNs). Here we leverage an ensemble of LSTM-BRNN and ResNet models, together with predicted residue-residue contact maps, to continue the push towards the attainable limit of prediction for 3- and 8-state secondary structure, backbone angles (θ, τ, ϕ and ψ), half-sphere exposure, contact numbers and solvent accessible surface area (ASA).

Authors

  • Jack Hanson
    Signal Processing Laboratory, Griffith University, Brisbane 4111, Australia.
  • Kuldip Paliwal
    Signal Processing Laboratory, School of Engineering, Griffith University, Brisbane, Australia.
  • Thomas Litfin
    School of Information and Communication Technology, Griffith University, Gold Coast 4222, Australia.
  • Yuedong Yang
    Institute for Glycomics and School of Information and Communication Technique, Griffith University, Parklands Dr. Southport, QLD 4222, Australia.
  • Yaoqi Zhou
    Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, Guangdong, 518106, China. Electronic address: zhouyq@szbl.ac.cn.

Keywords

External Resources

Popular Topics
Recent Journals