Bartletts principal regressive and arbitrary African buffalo optimizatized three-dimensional protein structure prediction

Protein sequencing and structure prediction is analytic to farther predict the tertiary structure, realize protein function and design drugs. However, experimental techniques are used to determine the structure of proteins which are time-consuming and expensive, and thus it's very intense to design significant computational methods for predicting protein structure based on optimization techniques. Existing deep learning-based methods have accomplished exceptional accomplishments of protein structure prediction, but the methods often utilize the features from prior knowledge with less focus on the error rate. To address this issue in this work, a Bartlett's Principal Regressive and Arbitrary African Buffalo Optimization (BPR-AABO) for performing secondary Protein Structure Prediction is proposed. The BPR-AABO method has one input layer, three hidden layers and output layer for protein structure prediction with higher accuracy and minimal time consumption. In BPR-AABO method, protein data is considered as an input and transmitted to the input layer. Relevant features are extracted from input protein data in hidden layer 1 by applying Bartlett's specificity test. The extracted features are transmitted to the hidden layer 2 where Principal Component Regression Analysis is applied with the chosen features for protein structure identification. Then, protein structure identification results are transmitted to the hidden layer 3. In that layer, Improved African Buffalo Optimization Model with sigmoid activation function is used for positioning the amino acids to form the protein structure and therefore performing protein structure prediction with higher accuracy and lesser time consumption. Experimental evaluation is carried out on factors such as, prediction accuracy, prediction time and ROC with respect to number of protein data.