Driving Accurate Allergen Prediction with Protein Language Models and Generalization-Focused Evaluation
Abstract
Allergens, typically proteins capable of triggering adverse immune responses, represent a significant public health challenge. To accurately identify allergen proteins, we introduce Applm (Allergen Prediction with Protein Language Models), a computational framework that leverages the 100-billion parameter xTrimoPGLM protein language model. We show that Applm consistently outperforms seven state-of-the-art methods in a diverse set of tasks that closely resemble difficult real-world scenarios. These include identifying novel allergens that lack similar examples in the training set, differentiating between allergens and non-allergens among homologs with high sequence similarity, and assessing functional consequences of mutations that create few changes to the protein sequences. Our analysis confirms that xTrimoPGLM, originally trained on one trillion tokens to capture general protein sequence characteristics, is crucial for Applm's performance by detecting important differences among protein sequences. In addition to providing Applm as open-source software, we also provide our carefully curated benchmark datasets to facilitate future research.
Cite
@article{arxiv.2508.10541,
title = {Driving Accurate Allergen Prediction with Protein Language Models and Generalization-Focused Evaluation},
author = {Brian Shing-Hei Wong and Joshua Mincheol Kim and Sin-Hang Fung and Qing Xiong and Kelvin Fu-Kiu Ao and Junkang Wei and Ran Wang and Dan Michelle Wang and Jingying Zhou and Bo Feng and Alfred Sze-Lok Cheng and Kevin Y. Yip and Stephen Kwok-Wing Tsui and Qin Cao},
journal= {arXiv preprint arXiv:2508.10541},
year = {2025}
}
Comments
59 pages, 5 main figures, 15 supplementary figures, 2 supplementary tables