Sequential antigenic regions identification in proteins that can stimulate immune system is one of the major challenges in peptide-based vaccine design. These antigenic regions that motivate B-cell response, are known as B-cell epitope (BCE).
The BCEs involve two types of cells i.e., plasma and memory B-cell. The plasma B-cells secrete and produce soluble antibodies. The memory B-cells remain in the organism and can re-expose to antigen. Therefore, to understand the structural and sequence features of BCEs is critical both for the design of operative vaccines and development of sensitive diagnostic tests.
In this work, we have developed a novel predictor, iLBE (Identification of B-Cell Epitope), by integrating evolutionary and sequence-based features through a random forest (RF) classifier. After optimization of each consecutive feature vector through Wilcoxon rank-sum test, we combined the RF scores by the logistic regression (LR) to enhance the prediction accuracy.