Prediction of Rheumatoid Arthritis-Associated Antigen Peptides Using a Novel Mixed Algorithm

• Published in: Information Computing and Applications pp. 528 - 535
• Main Authors: Liu, Tao, Zhang, Jiuyun, Lei, Tianyu
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg
• Series: Communications in Computer and Information Science
• Subjects: antigen peptide; Dynamic immune algorithm; human leukocyte antigen; Rheumatoid arthritis; Support vector machine regression
• ISBN: 9783642340406; 3642340407
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-642-34041-3_74

Identification of peptides binding to human leukocyte antigen (HLA) class II molecules is critical for understanding the basis of immunity, and for the development of vaccines and immunotherapeutic for autoimmune disease. A variety of methods have been successfully introduced into this field. However, lacking data is their common obstacle inducing a poor performance of prediction of peptides binding to some HLA class II molecules. For improving present condition, a coarse-graining idea, integrating peptides interacting with HLA class II molecules associated rheumatoid arthritis (RA) as the RA-associated dataset, was proposed. Then a new approach a novel mixed approach, combining dynamic immune algorithm (DIA) with support vector machine regression (SVR), was employed to build a prediction model of RA-associated antigen peptides on the RA-associated dataset. The rationality and advantage of this proposal is seen from the promising results (Acc = 93.1%, MCC = 0.863, AUC=0.966), and from the actual application of the human type II collagen.