Protein Secondary Structure Prediction Using CNN and Random Forest

Abstract

Protein structure prediction is an important problem in computational biology. Protein secondary structure prediction is the basis of protein three-dimensional structure prediction. In order to find an efficient algorithm for protein secondary structure prediction, this paper predicted the secondary structure of protein based on the depth learning algorithm and random forest algorithm. This method improves the model structure of convolutional neural networks (CNN). The Rectified Linear Units (ReLU) activation layer is added after each convolution layer to solve the gradient disappearance problem. In order to preserve the important features of the original data to the maximum extent, the feature data is used as the input of the Random Forest (RF) classifier to classify and predict the protein secondary structure. Compared with the traditional convolution neural network method, this method improves the prediction accuracy. Experiments show that the prediction accuracy of the ensemble learner composed of convolution neural network (CNN) and Random Forest (RF) model is higher than that of the traditional convolution neural network model the 25PDB data set. Therefore, the combination of deep learning algorithm and random forest model can improve the prediction accuracy of protein secondary structure better.