Related Experiment Video
Updated: Jan 6, 2026
06:50
Author Spotlight: A Computational Approach to Decipher Amino Acid Preferences in Multispecific Protein-Protein Interactions
Published on: January 26, 2024
2.5K
SiaScoreNet: a siamese neural network-based model integrating prediction scores for HLA-peptide interaction
Mahsa Saadat1,2, Fatemeh Zare-Mirakabad1,2, Milad Besharatifard1,2
1Computational Biology Research Center (CBRC), Amirkabir University of Technology, Tehran, Iran.
Bioinformatics Advances
|November 26, 2025
Summary
SiaScoreNet improves cancer immunotherapy by accurately predicting Human Leukocyte Antigen (HLA)-peptide interactions using a novel ensemble method. This tool enhances personalized treatment development with improved performance and efficiency.
Area of Science:
- Computational biology
- Immunology
- Bioinformatics
Background:
- Cancer immunotherapy relies on Human Leukocyte Antigen (HLA)-peptide interactions for tumor recognition.
- Accurate prediction of these interactions is crucial for developing personalized immunotherapies.
- Existing allele-specific models lack scalability, while pan-specific and ensemble methods offer improvements.
Purpose of the Study:
- To develop SiaScoreNet, a novel predictive pipeline for enhanced HLA-peptide interaction prediction.
- To improve the accuracy and efficiency of predicting HLA-peptide binding for immunotherapy applications.
- To address the limitations of existing models in scalability and capturing complex interactions.
Main Methods:
- Utilized ESM, a transformer-based protein language model, for sequence embedding.
- Integrated predicted scores from state-of-the-art models into a feature vector.
- Employed a nonlinear ensemble strategy to combine features and enhance prediction performance.
Main Results:
- SiaScoreNet demonstrated superior accuracy compared to existing models, achieving performance comparable to leading methods.
- The model offers improved runtime efficiency over current state-of-the-art predictors.
- Evaluations were conducted using HPV virus data for HLA-peptide prediction.
Conclusions:
- SiaScoreNet provides a powerful and efficient tool for HLA-peptide interaction prediction.
- The developed pipeline advances the development of personalized cancer immunotherapies.
- Publicly available code and data facilitate further research and application.
