Epitope Prediction of Antigen Protein Using Attention-based LSTM Network

Abstract B-cells inducing antigen-specific immune responses in vivo produce large amounts of antibodies by recognizing the subregions (epitope regions) antigen proteins. They can inhibit their functioning binding to Predicting epitope regions is beneficial for design and development vaccines aimed induce antibody production. However, prediction accuracy requires improvement. The conventional region methods have focused only on target sequence amino acid sequences an entire protein not thoroughly considered its features as a whole. In present paper, we propose deep learning method based short-term memory with attention mechanism consider characteristics whole addition sequence. proposed achieves better compared experimental using data from database.