Pathway to precision: machine learning and bioinformatics in diabetes gene expression studies.

Area of Science:

• Biomedical Informatics
• Genomics
• Metabolic Disorders

Background:

• Diabetes mellitus is a critical metabolic disorder requiring in-depth understanding of its pathogenic mechanisms.
• Identifying molecular causes of diabetes is essential, with biomarker analysis, machine learning, and bioinformatics playing crucial roles.

Purpose of the Study:

• To summarize current advancements in diabetes research.
• To highlight progress in bioinformatics, gene expression analysis, and machine learning applications for diabetes.
• To identify key techniques and applications in recent diabetes studies.

Main Methods:

• A systematic literature search was performed on Google Scholar, PubMed, and Scopus.
• Keywords included "diabetes," "biomarkers," "bioinformatics," and "machine learning."
• PRISMA guidelines were followed for article selection, focusing on open-access English articles published between 2020 and 2024, resulting in 96 selected studies.

Main Results:

• Microarrays were the predominant data acquisition technique (70 occurrences).
• KEGG pathway analysis was the most common bioinformatics method (84 instances), followed by LASSO (43) and PCA (47) for machine learning.
• Diagnostic applications (51) were more frequent than prognostic applications (45).

Conclusions:

• KEGG pathway analysis and microarray data are vital in current diabetes research.
• Underutilization of newer technologies like RNA-seq and single-cell RNA-seq was noted.
• Integrating advanced bioinformatics, machine learning, and clinical data is crucial for precision medicine and personalized diabetes treatment.