Applm

Applm是一个基于蛋白质语言模型的过敏原预测计算框架，利用1000亿参数的xTrimoPGLM蛋白质语言模型，在多种任务中表现出色，包括识别训练集中缺乏相似示例的新过敏原、区分高序列相似性的同源中的过敏原和非过敏原，以及评估导致蛋白质序列变化较小的突变的后果。Applm在准确预测过敏原方面取得了新的突破，为未来过敏原研究提供了有价值的基准数据集和评估流程。

Applm is a computational framework for allergen prediction based on protein language models. It utilizes the 100-billion-parameter xTrimoPGLM protein language model and excels in multiple tasks, including identifying novel allergens that lack similar examples in the training set, distinguishing allergens from non-allergens among homologs with high sequence similarity, and evaluating the consequences of mutations that induce minimal changes to protein sequences. Applm has achieved new breakthroughs in accurate allergen prediction, providing a valuable benchmark dataset and evaluation workflow for future allergen research.