Aqueous humor proteomics analyzed by bioinformatics and machine learning in PDR cases versus controls

Area of Science:

• Ophthalmology
• Proteomics
• Biomarker Discovery

Background:

• Proliferative diabetic retinopathy (PDR) involves complex pathophysiological mechanisms.
• Early detection and monitoring of PDR are crucial for effective management.
• The diagnostic value of aqueous humor (AH) in PDR onset needs further investigation.

Purpose of the Study:

• To comprehend the molecular events contributing to PDR.
• To evaluate the diagnostic utility of AH biomarkers for PDR.
• To identify reliable indicators for monitoring PDR progression.

Main Methods:

• Proteomics analysis of aqueous humor from PDR and cataract patients.
• Bioinformatics and machine learning (inference of biomolecular combinations with minimal bias) were employed.
• Identification of functional relevance, hub proteins, and potential biomarkers.

Main Results:

• Identified potential surrogate protein biomarkers for PDR monitoring (e.g., SIAE, SEMA7A, GNS, IGKV3D-15; ATP6AP1, SPARCL1, SERPINA7).
• Discovered hub proteins (ALB, FN1, ACTB, SERPINA1, C3, VTN) in AH proteomes.
• SERPINA1 showed high correlation with best corrected visual acuity (BCVA) and diabetes duration; "Complement and coagulation cascades" pathway implicated in PDR.

Conclusions:

• Aqueous humor proteomics offers stable and accurate biomarkers for PDR early warning and diagnosis.
• The study enhances understanding of PDR molecular mechanisms.
• Provides a valuable resource for optimizing PDR management strategies.