Function Prediction from Protein Primary Structure using Deep Learning LSTM Algorithm

Abstract

Biological information of protein primary structure is responsible for finding the protein function, extracting features and function of a protein in the biology lab is challenging and time-consuming. Identification of protein function provides essential information for the treatment of various diseases and drug design. Therefore, extracting the protein knowledge from primary structure alone has been a diverse field in the study of bioinformatics data mining and computational biology. This study aimed to function prediction of protein primary structure using the LSTM methods. PRNP(prion protein )most of the nervous system tissues express by prion protein, this is generally to protease-resistant from disease, due to this reasons, the human codon PRNP is most closely associated with Alzheimer disease. The PRNP protein data trained with Hemo sapiens PRNP selection, classification was implemented with network layer perceptron. The learning algorithms are frame by the nervous system. The training results observation indicate that the learning success of prion protein classification leads positively.