Prediction of cellulose synthase using conserved protein domain through Machine learning-based Approach

Abstract

Abstract Cellulose synthase, a pivotal enzyme involved in cellulose and hemicellulose synthesis, plays a crucial role in plant cell wall maintenance. This study aims to deepen our understanding of cellulose synthase by developing a highly sensitive computational method for its characterization. By leveraging a synergistic approach that combines the utilization of Pfam domains and a dipeptide composition-based support vector machine model, we address the challenges associated with conventional characterization methods. Our computational approach focuses on identifying unique domains within the cellulose synthase protein sequence through proximity-based detection. In the absence of such domains, predictions are generated using a support vector machine model. The accuracy of our method was evaluated, yielding an overall accuracy of 89.92% and a specificity of 90.37%. The outcomes of this research provide valuable insight into the structure and function of cellulose synthase, shedding light on the glycosyltransferase activities critical for cellulose synthesis. Furthermore, the implications of this study extend to the cotton and textile industry, where cellulose-based materials are widely employed. The developed computational approach demonstrates its potential as an effective tool for further exploration and characterization of cellulose synthase.