![]() | Only 14 pages are availabe for public view |
Abstract This work aims to enhance the antibodies production for diagnostic kits using epitopes prediction approaches to design immunogenic peptides that can be injected into animal to produce antibodies that cross react with target protein. Epitopes prediction using machine learning approach was adopted due to the increasing availability of experimentally identified epitopes in addition to low performance of previous methods. In this work, we use support vector machine (SVM) and string kernel to build models for predicting linear B-cell epitopes The obtained models were tested by 10 fold cross validation method then applied to the coat protein of Potato leaf roll virus (PLRV) and one of the predicted epitopes was chemically synthesised and injected into mice and the obtained antibodies cross react successfully with PLRV infected plant tissue. Keywords: Immunoinformatic, Epitopes prediction, Support vector machine, String kernel. |