Recurrent Neural Network based Classification of Protein-Protein Interactions

Abstract

Proteomics is an attempt to describe or explain biological state and qualitative and quantitative changes of protein content of cells and extracellular biological materials under different conditions to further understand biological processes. Protein-Protein interaction prediction and classification is a very important task. Prediction and classification of protein-protein interactions can help in improving the understanding of diseases and can provide the basis for new therapeutic approaches. In this work a model is proposed to classify protein-protein interactions. Jordan Recurrent Neural Network is used to classify the proteinprotein interactions. The model developed gives 97.25% of accuracy which is 8.7% more than Back-Propagation Neural Network.