Principal component analysis based feature extraction approach to identify circulating microRNA biomarkers

Area of Science:

• Biomarker Discovery
• Molecular Diagnostics
• Genomics

Background:

• Blood biomarkers are crucial for disease diagnosis due to ease of collection.
• Blood reflects overall patient status, complicating specific disease biomarker identification.
• MicroRNAs (miRNAs) are promising biomarker candidates due to their recent discovery, limited number, and critical regulatory roles in biological processes.

Purpose of the Study:

• To identify combinations of miRNAs that can discriminate between healthy controls and 14 different diseases, including 5 types of cancer.
• To develop a novel feature selection method for biomarker discovery that does not require prior knowledge of sample status (diseased or healthy).

Main Methods:

• A new principal component analysis-based feature selection method was employed.
• The method was applied to identify miRNA combinations capable of distinguishing 14 diseases from healthy controls.
• KEGG pathway enrichment analysis was performed on target genes of identified miRNAs.

Main Results:

• Twelve specific miRNAs (hsa-miR-425, hsa-miR-15b, hsa-miR-185, hsa-miR-92a, hsa-miR-140-3p, hsa-miR-320a, hsa-miR-486-5p, hsa-miR-16, hsa-miR-191, hsa-miR-106b, hsa-miR-19b, and hsa-miR-30d) were identified as potential biomarkers.
• Combinations of 10 of these miRNAs effectively discriminated each of the 14 diseases from healthy controls.
• These miRNAs showed significant up- or downregulation in various diseases, and their target genes were enriched in disease-specific KEGG pathways.

Conclusions:

• The identified 12 miRNAs serve as potential blood-based biomarkers for early disease detection.
• The novel feature selection method is effective for identifying disease-specific miRNA signatures.
• These miRNA biomarkers show promise for distinguishing between different diseases and healthy states.