The health of things for classification of protein structure using improved grey wolf optimization

In the field of computational biology, prediction of high-resolution protein structure is regarded as a major challenge. Physical and chemical properties of the protein structure determine its quality and differentiate native structures from predicted structures. Various machine learning classification models are studied with six physical and chemical properties to classify the root mean square deviation of the protein structure. This work proposes an improved version of a meta-heuristic technique named grey wolf optimization (IGWO), which is an extension of traditional grey wolf optimization (GWO) for the feature selection. The proposed novel IGWO ascertains optimal subset of features, and further, four machine learning classifiers have been used for efficient prediction of protein structure. Artificial neural network classifier predicts the protein structure with a maximum approximate accuracy of 91%. The experimental result reveals that the proposed meta-heuristic technique is stable enough to maximize the accuracy and minimize the number of optimal features. In this paper, the result of the proposed technique has been compared with other related evolutionary techniques and the proposed optimizer outperforms all other techniques. The dataset used in the study is available at http://bit.ly/RMSDClassification-DS.