Related Experiment Video
Updated: Aug 22, 2025
06:19
Constructing and Visualizing Models using Mime-based Machine-learning Framework
Published on: July 22, 2025
270
Decoding Diabetes Biomarkers and Related Molecular Mechanisms by Using Machine Learning, Text Mining, and Gene
Amira M Elsherbini1, Alsamman M Alsamman2, Nehal M Elsherbiny3,4
1Department of Oral Biology, Faculty of Dentistry, Mansoura University, Mansoura 35116, Egypt.
International Journal of Environmental Research and Public Health
|November 11, 2022
Summary
This study identified key genes associated with diabetes using bioinformatics and machine learning. It highlights 39 potential biomarkers, including HLA-DQB1, for early diabetes detection and diagnosis.
Area of Science:
- Genomics
- Bioinformatics
- Molecular Biology
Background:
- The molecular underpinnings of diabetes mellitus remain incompletely understood.
- Identifying key genes and pathways is crucial for understanding diabetes pathogenesis.
Purpose of the Study:
- To identify frequently reported and differentially expressed genes (DEGs) in diabetes using bioinformatics.
- To discover novel biomarkers for diabetes diagnosis and early detection.
Main Methods:
- Text mining of 40,225 diabetes literature abstracts to identify prevalent genes.
- Gene expression analysis of three datasets (44 patients, 57 controls) to find DEGs.
- Machine learning algorithms (decision tree, extra-tree regressor, random forest) applied to identify diagnostic biomarkers.
Main Results:
- Text mining identified 5939 diabetes-related genes, with 10 genes (e.g., HNF4A, PPARA, VEGFA) mentioned in over 200 articles.
- Gene expression analysis revealed 135 significant DEGs, including CEACAM6 and ENPP4.
- Machine learning models achieved accuracies from 0.6364 to 0.88, identifying 39 potential biomarkers, with HLA-DQB1 highlighted for early detection.
Conclusions:
- Bioinformatics and machine learning approaches effectively identified key genes and potential biomarkers for diabetes.
- The study provides valuable insights into the genetic landscape of diabetes and highlights specific genes for diagnostic applications.
