| Summary: | 碩士 === 國立中興大學 === 基因體暨生物資訊學研究所 === 101 === Immune reaction is the most important defense mechanism for destroying invading-pathogens in our body, and the epitope is the position of antigen-antibody interaction on pathogen proteins. Most of epitopes belong to structural epitope, but the existing sequence-based predicting website still has a lot of ways to improve the predicting performance. Therefore, in this study used SVM as machine learning tool to predict epitope by protein sequences. First, we built five SVM models in the first layer according to five features include binary, composition, position-specific scoring matrix, secondary structure, accessible surface area and association rule, then choosing the patterns which have the best performance in each model. Second, using the confidence score of first layer models as the input value for the SVM model in the second layer that SVM model integrated first layer SVM models for improving the predicting accuracy. The final prediction model can achieve up to 63% of accuracy in epitope predicting result, the predicting performance is batter then existing predicting website. Finally, a case study used two-subunit cytochrome c oxidase of Paracoccus denitrificans as testing, the accuracy can achieve up to 66%.
|
|---|
