Related Experiment Video
Updated: Jun 15, 2025
08:58
Label-Free Identification of Lymphocyte Subtypes Using Three-Dimensional Quantitative Phase Imaging and Machine Learning
Published on: November 19, 2018
12.4K
PredIL13: Stacking a variety of machine and deep learning methods with ESM-2 language model for identifying
Hiroyuki Kurata1, Md Harun-Or-Roshid1, Sho Tsukiyama1
1Department of Bioscience and Bioinformatics, Kyushu Institute of Technology, Kawazu, Iizuka, Fukuoka, Japan.
Plos One
|August 22, 2024
Summary
We developed PredIL13, an advanced computational tool to identify Interleukin (IL)-13 inducing peptides. This method accelerates the discovery of peptides crucial for understanding COVID-19 severity and other biological processes.
Area of Science:
- Biochemistry and Molecular Biology
- Computational Biology
- Immunology
Background:
- Interleukin (IL)-13 is a recently identified cytokine implicated in the severity of COVID-19.
- Identifying IL-13 inducing peptides is crucial for understanding its biological roles.
- Experimental methods for peptide identification are time-consuming and costly.
Purpose of the Study:
- To develop a novel computational tool for predicting IL-13 inducing peptides.
- To enhance the accuracy and efficiency of IL-13 peptide identification.
- To provide a valuable resource for researchers in immunology and virology.
Main Methods:
- Developed PredIL13, an ensemble learning method utilizing the ESM-2 protein language model.
- Stacked probability scores from 168 single-feature machine/deep learning models.
- Employed a logistic regression meta-classifier and sequential deletion based on iterative AWCLR ranking (SDIWC) to select top models.
Main Results:
- The PredIL13 method significantly outperformed existing state-of-the-art predictors.
- Identified the top 16 single-feature models contributing to prediction accuracy.
- Demonstrated the efficacy of the ESM-2 model and AWCLR in feature selection.
Conclusions:
- PredIL13 is a highly accurate and efficient tool for detecting IL-13 inducing peptides.
- This computational approach accelerates the discovery of biologically significant peptides.
- The findings contribute to a better understanding of IL-13's role in diseases like COVID-19.
