B-cell epitope prediction
Patent Information
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- NEC ONCOIMMUNITY AS
- Filing Date
- 2024-06-13
- Publication Date
- 2026-04-22
AI Technical Summary
Current methods for predicting B-cell epitopes, especially conformational B-cell epitopes, face challenges due to limited data defining true epitopes and the requirement for 3D protein structure information, which is often not available, leading to inaccurate predictions and difficulties in distinguishing between different epitopes.
A computer-implemented method using a trained machine learning model that predicts B-cell epitopes based on the protein's structure and surface characteristics in an unbound state, without requiring the full 3D structure, by accessing secondary structure, relative solvent accessibility, and half-sphere exposure, and incorporating physiochemical characteristics to improve prediction accuracy.
This approach enables accurate prediction of B-cell epitopes, including conformational ones, without the need for experimental 3D structure data, providing improved confidence in identifying true epitopes and their relationships, thus enhancing vaccine design and diagnostic applications.
Smart Images

Figure EP2024066470_19122024_PF_FP_ABST