A Meta Decision Tree Approach for B-cell Epitope Mining

檔案

本條目沒有附加檔案。

日期

2016-01-01

期刊名稱

2016 IEEE CONFERENCE ON COMPUTATIONAL INTELLIGENCE IN BIOINFORMATICS AND COMPUTATIONAL BIOLOGY (CIBCB)

摘要

The ability of antibodies to respond to an antigen depends on the antibodies' specific recognition of epitopes, which are sites of the antigen to which antibodies bind. An increase in the availability of protein sequences and structures has enabled the identification of conformational epitopes, using various computational methods. The meta learner, among various approaches, has proved its feasibility and comparable accuracy in B-cell epitope prediction in previous studies. Nevertheless, its performance highly depends on the classification results of its multiple epitope base predictors within the meta learning architecture. We here propose bagging meta decision trees for epitope prediction to avoid the dependence on epitope prediction tools, and introduce 3D sphere-based attributes to improve prediction accuracy. Our experimental results demonstrate the superior performance of the bagging meta decision tree approach in comparison with single epitope predictors.

所屬典藏

會議論文
顯示完整詮釋資料 統計資料