Multi-Omics and Machine Learning Integration Identifies Key Endothelial Modules in Carotid Artery Stenosis

Area of Science:

• Biomarkers and diagnostics
• Genomics and transcriptomics
• Vascular biology

Background:

• Carotid artery stenosis (CAS) is a significant risk factor for ischemic stroke.
• Current diagnostic methods for CAS lack sufficient sensitivity for early detection.
• There is a critical need for reliable molecular biomarkers for early CAS identification.

Purpose of the Study:

• To identify novel endothelial cell-derived molecular biomarkers for early diagnosis of CAS.
• To validate these biomarkers across multiple independent datasets and molecular levels.

Main Methods:

• Integrated single-cell RNA sequencing (scRNA-seq), bulk transcriptomics, and multi-omics data.
• Applied Weighted Gene Co-expression Network Analysis (WGCNA) and machine learning algorithms.
• Validated findings across independent GEO and ZZ cohorts at single-cell, transcriptomic, and proteomic levels.

Main Results:

• scRNA-seq revealed endothelial cell enrichment in CAS and identified 836 potential markers.
• Integrative analysis narrowed down candidates to 80, with NRP1 and XAF1 selected by machine learning.
• NRP1 and XAF1 demonstrated consistent upregulation and strong diagnostic performance across multi-omics validation datasets.

Conclusions:

• NRP1 and XAF1 are identified as novel endothelial cell-derived biomarkers for CAS.
• These biomarkers show significant potential for early screening and clinical diagnosis of carotid artery stenosis.
• The findings highlight the utility of multi-omics integration and machine learning in biomarker discovery.