Machine learning identifies a NETs-related four-gene diagnostic signature for abdominal aortic aneurysm

Area of Science:

• Vascular Biology
• Immunology
• Genomics

Background:

• Abdominal aortic aneurysm (AAA) is a degenerative vascular disease marked by inflammation.
• Neutrophil extracellular traps (NETs) are implicated in AAA progression, but their transcriptomic profile is not well understood.
• Identifying diagnostic biomarkers and understanding immune heterogeneity are crucial for AAA risk stratification.

Purpose of the Study:

• To identify key NETs-related genes (NRGs) dysregulated in AAA.
• To develop a diagnostic signature for early AAA detection.
• To explore molecular subtypes and potential therapeutic targets in AAA.

Main Methods:

• Integrated transcriptomic datasets from Gene Expression Omnibus (GEO).
• Employed weighted gene co-expression network analysis and differential expression profiling.
• Utilized machine learning algorithms (LASSO, SVM-RFE, Random Forest) for feature selection and signature derivation.
• Validated the diagnostic model in an independent cohort.

Main Results:

• Identified a robust 4-gene diagnostic signature (CXCR4, GZMB, ITGA6, CD47) for AAA.
• The signature achieved an AUC of 0.920 in the training cohort and showed consistent performance in validation.
• Consensus clustering revealed two molecular subtypes, with one showing high NETosis activity.
• Identified Eugenol and Tretinoin as potential therapeutic compounds.

Conclusions:

• NETs play a critical role in AAA pathogenesis.
• The 4-gene signature provides a reliable tool for early AAA diagnosis.
• The study offers a framework for precision risk stratification and developing novel nonsurgical therapies for AAA.