Recurrent Neural Networks for Linear B-Epitope Prediction in Antigens

• DOI: 10.21533/scjournal.v6i2.140
• Publication Date: 2018-05-11T06:59:05Z
• Authors: Betul Akcesme; Faruk B. Akcesme; Muhamed Adilovic; Mehmet Can
• Keywords: Linear epitope

Experimental methods used for characterizing epitopes that play a vital role in the development of peptide vaccines, diagnosis diseases, and also allergy research are time consuming need huge resources. There many online epitope prediction tools can help experimenters short listing candidate peptides. To predict B-cell an antigenic sequence, Jordan recurrent neural network (JRNN) found to be more successful. train test networks, 262.583 B retrieved from IEDB database. 99.9% these have lengths interval 6-25 amino acids. For each lengths, committees 11 expert networks trained. experts alongside epitopes, non-epitopes needed. Non-epitopes created as random sequences acids same length followed by filtering process. distinguish non-epitopes, votes eleven aggregated majority vote. An overall accuracy 97.23% is achieved. Then linear b antigen, ESAT6 (Tuberculosis).