RPI-MCNNBLSTM: BLSTM Networks Combining With Multiple Convolutional Neural Network Models to Predict RNA-Protein Interactions Using Multiple Biometric Features Codes

• Main Authors: Ruibo Gao, Tianhua Yang, Yaxue Shen, Yifan Rong, Kang Ye, Junlan Nie
• Format: Article
• Language: English
• Published: IEEE 2020-01-01
• Series: IEEE Access
• Subjects: Multiple convolutional neural network; LSTM networks; multiple biometric features codes; RNA-protein interactions
• Online Access: https://ieeexplore.ieee.org/document/9224615/

RNA plays an important role in many biological processes, and RNA functions are primarily achieved by binding with a variety of proteins. But with the increasing complexity of RPIs networks, high-throughput biological techniques are usually expensive and time consuming. Therefore, there is an urgent need for high speed and reliably computational methods to predict RNA-protein interactions. In this study, we propose a hybrid deep learning model: RPI-MCNNBLSTM, which combines three convolutional neural networks (CNN) with a BLSTM network, to predict RNA-protein interactions using many-sided biological information including protein sequences, RNA sequence and structure. Firstly, we adopt a filling method to pad sequence and structure into equal length, and perform numerical encoding for the sequence and structure of the above equal length, respectively, which are appropriate for subsequent convolution operations. Secondly, we establish the three CNNs to learn the three biological information, separately, then use the BLSTM to capture the long range dependencies among the three features identified by the CNNs. The learned weighted representations are fed into a classification layer to predict ncRNA-protein interactions. Finally, the experimental results indicate that the proposed method achieves superior performance with the accuracy of 98.37% on the RPI1807 dataset, 92.99% on the RPI2241 dataset, 95.47% on the RPI369 dataset, 90.0% on the RPI448 and 87.4% on the RPI1446 dataset, respectively. The code of RPI-MCNNBLSTM and the datasets used in this work are available at https://github.com/xiaopang136/RPI for academic users.