Integrative Machine Learning Analysis of Programmed Cell Death Pathways Identifies Novel Diagnostic Biomarkers for...

Area of Science:

• Cardiovascular Biology
• Molecular Mechanisms of Disease
• Genomics and Bioinformatics

Background:

• Atrial fibrillation (AF) is a major cause of stroke, heart failure, and mortality.
• The precise molecular mechanisms underlying AF remain poorly understood.

Purpose of the Study:

• To map programmed cell death (PCD) pathways in AF.
• To identify diagnostic genes associated with AF.
• To develop a risk stratification tool for AF patients.

Main Methods:

• Integrated bulk transcriptomes with weighted gene co-expression network analysis and consensus clustering.
• Employed a machine-learning pipeline with 66 model combinations to analyze PCD pathways.
• Utilized CIBERSORT, xCell, and ssGSEA for immune infiltration profiling.
• Validated hub-gene expression in an HL-1 atrial pacing model and patient PBMCs.

Main Results:

• Identified four hub genes: SGPL1, NPC2, PTGDS, and RCAN1.
• Developed a nomogram and PCD-based risk score (PCDscore) with robust discrimination.
• Observed increased immune-cell infiltration and dysregulated immune modulators in high PCDscore patients, with macrophage enrichment.
• Validated gene expression changes in AF cell models and patient PBMCs.

Conclusions:

• The study reveals PCD-linked remodeling in AF.
• SGPL1, NPC2, PTGDS, and RCAN1 are nominated as candidate diagnostic biomarkers.
• A PCD-based nomogram and risk score may aid in patient stratification and targeted interventions.