BiCNN: A Highly Efficient and Lightweight Model for TCR Affinity Prediction
In this study, we propose a lightweight and efficient model for TCR affinity prediction, BiCNN, which integrates a bidirectional LSTM and a convolutional neural network to effectively capture the complex interactions between TCRs and antigens. Experimental results demonstrate that BiCNN outperforms traditional models in terms of prediction accuracy under limited data and short training time, offering a practical solution for immunological research.
Authors
- Zhonghuang Wang (ORCID: https://orcid.org/0000-0002-1268-883X)
- Wenming Zhao (ORCID: https://orcid.org/0000-0002-4396-8287)
- Zongyou Yu (ORCID: https://orcid.org/0009-0005-1098-9793)
- Ruoshen Yang
- Jian Cao (ORCID: https://orcid.org/0000-0001-8903-2903)
Institutions
- Beijing Technology and Business University (CN)
- Beijing Institute of Genomics (CN)
- Beijing Academy of Artificial Intelligence (CN)
Publication Details
- Published
- 2025-05-23
- DOI
- https://doi.org/10.1109/cvidl65390.2025.11085611
- Primary Topic
- Gene expression and cancer classification
- Type
- article
- Field-Weighted Citation Impact
- 0.00