Transfer Learning for T-Cell Response Prediction

2024-03-18

arXiv.org

AI Summary

Predicting T-cell responses to peptide sequences is critical for personalized cancer vaccine development, yet models suffer from “shortcut learning” due to limited sample sizes and heterogeneous, multi-source data—leading them to rely on spurious domain-specific cues (e.g., biological origin) rather than genuine immunogenic signals. To address this, we propose a domain-aware evaluation paradigm that systematically identifies and quantifies such bias for the first time. We further introduce a cross-species independent fine-tuning strategy, integrating a Transformer architecture with strict domain-isolated evaluation to enable robust domain-adaptive transfer learning. Our method achieves state-of-the-art performance on human peptide immunogenicity prediction, significantly outperforming existing baselines. Moreover, it demonstrates strong generalization to non-human peptides—including murine and viral sequences—establishing a novel paradigm for cross-species immunogenicity modeling.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

We study the prediction of T-cell response for specific given peptides, which could, among other applications, be a crucial step towards the development of personalized cancer vaccines. It is a challenging task due to limited, heterogeneous training data featuring a multi-domain structure; such data entail the danger of shortcut learning, where models learn general characteristics of peptide sources, such as the source organism, rather than specific peptide characteristics associated with T-cell response. Using a transformer model for T-cell response prediction, we show that the danger of inflated predictive performance is not merely theoretical but occurs in practice. Consequently, we propose a domain-aware evaluation scheme. We then study different transfer learning techniques to deal with the multi-domain structure and shortcut learning. We demonstrate a per-source fine tuning approach to be effective across a wide range of peptide sources and further show that our final model is competitive with existing state-of-the-art approaches for predicting T-cell responses for human peptides.

Problem

• Predict T-cell response for peptides
• Address limited heterogeneous training data
• Prevent shortcut learning in models

Innovation

• Transformer model application
• Domain-aware evaluation scheme
• Per-source fine tuning approach