Identification of Key Genes in Diabetic Retinopathy Using Bioinformatics and Machine Learning Approaches

Area of Science:

• Genomics and Bioinformatics
• Ophthalmology
• Molecular Biology

Background:

• Diabetic retinopathy (DR) poses a significant threat to vision, necessitating improved diagnostic and therapeutic strategies.
• Identifying key molecular players in DR pathogenesis is crucial for advancing treatment options.

Purpose of the Study:

• To identify key genes associated with diabetic retinopathy (DR) using bioinformatics and machine learning.
• To evaluate the diagnostic performance and explore biological functions of identified genes.
• To elucidate upstream regulatory mechanisms for potential therapeutic targets in DR.

Main Methods:

• Bioinformatic analysis of DR transcriptomic datasets (GSE94019, GSE60436) from the Gene Expression Omnibus (GEO) database.
• Application of Least Absolute Shrinkage and Selection Operator (LASSO) regression and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) for feature gene selection.
• Receiver operating characteristic (ROC) curve analysis for diagnostic performance evaluation and construction of a lncRNA-miRNA-mRNA regulatory network.

Main Results:

• Identification of Collagen Type VI Alpha 2 Chain (COL6A2) as a key upregulated gene in DR with high diagnostic accuracy (AUC=1.00 training, 0.89 validation).
• Functional enrichment analysis revealed COL6A2-associated differentially expressed genes (DEGs) are involved in extracellular matrix (ECM) organization, cell adhesion, angiogenesis, and inflammation.
• Prediction of potential regulatory roles for hsa-miR-762 and hsa-miR-29a-3p in COL6A2 expression via a ceRNA network.

Conclusions:

• COL6A2 is identified as a significant molecular marker for diabetic retinopathy (DR).
• COL6A2's involvement in ECM remodeling, cell adhesion, and angiogenesis highlights its role in DR pathogenesis.
• These findings offer potential molecular targets for early DR diagnosis and targeted therapies.