A sparse autoencoder-based deep neural network for protein solvent accessibility and contact number prediction

BMC Bioinformatics

Research

Zhiwen Zeng¹  Lei Deng²  Chao Fan² 

[1] School of Information Science and Engineering, Central South University, No.932 South Lushan Road, 410083, Changsha, China;School of Software, Central South University, No.22 Shaoshan South Road, 410075, Changsha, China;

关键词: Solvent accessibility;    Contact number;    Deep neural network;    Sequence-derived features;   

DOI  :  10.1186/s12859-017-1971-7

来源: Springer

【 摘 要 】

BackgroundDirect prediction of the three-dimensional (3D) structures of proteins from one-dimensional (1D) sequences is a challenging problem. Significant structural characteristics such as solvent accessibility and contact number are essential for deriving restrains in modeling protein folding and protein 3D structure. Thus, accurately predicting these features is a critical step for 3D protein structure building.ResultsIn this study, we present DeepSacon, a computational method that can effectively predict protein solvent accessibility and contact number by using a deep neural network, which is built based on stacked autoencoder and a dropout method. The results demonstrate that our proposed DeepSacon achieves a significant improvement in the prediction quality compared with the state-of-the-art methods. We obtain 0.70 three-state accuracy for solvent accessibility, 0.33 15-state accuracy and 0.74 Pearson Correlation Coefficient (PCC) for the contact number on the 5729 monomeric soluble globular protein dataset. We also evaluate the performance on the CASP11 benchmark dataset, DeepSacon achieves 0.68 three-state accuracy and 0.69 PCC for solvent accessibility and contact number, respectively.ConclusionsWe have shown that DeepSacon can reliably predict solvent accessibility and contact number with stacked sparse autoencoder and a dropout approach.

【 授权许可】

CC BY   
© The Author(s) 2017

【 图 表 】

Fig. 1