Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

103 Views

2024 Technical Program

Protein and Co-Products

Session: Anti-Nutritional Factors to Allergies: The Negative Connotation of Alternative Proteins

Predicting allergenic proteins and peptides based on AI language models

Wednesday, May 1, 2024

8:52 AM – 9:09 AM EDT

Room: 520b

Presenting Author(s)

Yonghui Li, PhD

Associate Professor
Kansas State University
Manhattan, Kansas, United States

Co-Author(s)

Zhenjiao Du

Graduate student
Kansas State University
Manhattan, Kansas, United States
Changqi Liu (he/him/his)

Associate Professor
San Diego State University
San Diego, California, United States

Abstract: Allergic diseases encompass a range of disorders driven by both innate and adaptive immune responses, in collaboration with epithelial cells. The majority of allergens are related to protein and peptides, which are widely present in foods, personal care products, and the environment. The growing demand for identifying new allergens and allergenic risk assessment drives the development of efficient and accurate in silico prediction models. Recently, inspired by the extraordinary performance of transformer-based models in natural language processing tasks, a series of attempts have been made in pretrained protein language models (pLMs) (e.g., ProtTrans, ESM, UniRep, etc.) for downstream protein sequence tasks. This study aims to develop robust models for allergenic protein/peptide prediction using five pLMs of varying sizes and systematically assess their performance through fine-tuning with a convolutional neural network. The developed pLM4Alg models have achieved state-of-the-art performance with accuracy, Matthews correlation coefficient, and area under the curve scoring 94.1-95.5%, 0.882-0.911, and 98.3-99%, respectively. A positive relationship between the prediction performance and the size of pLMs was observed. Moreover, pLM4Alg is the first model capable of handling prediction tasks involving residue-missed sequences and sequences containing non-standard amino acid residues. To facilitate easy access, a user-friendly web server has been established (https://f6wxpfd3sh.us-east-1.awsapprunner.com ). pLM4Alg is expected to become the leading machine learning-based prediction model for allergenic peptides and proteins. Its collaboration with other predictors holds great promise in accelerating allergy research.