Communications in Information and Systems

Volume 19 (2019)

Number 3

Inference of RNA structural contacts by direct coupling analysis

Pages: 279 – 297

DOI: https://dx.doi.org/10.4310/CIS.2019.v19.n3.a3

Authors

Xiaoling He (School of Physics, Huazhong University of Science & Technology, Wuhan, China)

Shuaimin Li (School of Physics, Huazhong University of Science & Technology, Wuhan, China)

Xiujuan Ou (School of Physics, Huazhong University of Science & Technology, Wuhan, China)

Jun Wang (School of Physics, Huazhong University of Science & Technology, Wuhan, China)

Yi Xiao (School of Physics, Huazhong University of Science & Technology, Wuhan, China)

Abstract

Direct coupling analysis (DCA) has been widely used to infer residue-residue contacts in protein structures but rarely to those in RNA structures. Here we analyze the performances of two popular algorithms of DCA, DCA under mean-field approximation (mfDCA) and pseudo-likelihood maximization approximation (plmDCA), in the inference of RNA contacts and found that, unlike proteins, their performances are similar in this case. Furthermore, a deep learning model of fully convolutional neural network (FCN) is used to improve the performance of DCA and the result is better than that of the original DCA.

Full Text (PDF format)

X. He and S. Li made equal contributions to this paper.

The research of Y. Xiao was supported by the NSFC under Grant No. 31570722 and 11874162.

Received 30 August 2019

Published 6 December 2019