Data Fusion Analysis for Determining Localization of Proteins Associated to Escherichia coli

Abstract

In recent years, the interest in protein analysis based on biomolecular features has rapidly grown. This has led to explore the use of machine learning models, as they could hold an important alternative to contribute to the problems associated to these analyses. Models as support vector machines, artificial neural networks and random forest were compared for the prediction of protein localization. Two main sources of data were used to train the models: the information from targeting signal and from the protein sequences to determine the localization sites of the protein. A third scenario with a fusion of both sources of data was employed. Four classes were established according to the subcellular localization of the protein: cytoplasm, periplasmatic space, outer and inner membranes. Results reached values between 77% and 92% in terms of balanced accuracy. The models with better performance were based on random forest and support vector machines.