Multi-omics and Machine Learning Analysis of Abdominal Aortic Aneurysm Biomarkers

551 genes were differentially expressed in abdominal aortic aneurysm (AAA) tissue compared to controls. An integrated analysis of expression and protein quantitative trait loci (eQTL/pQTL) via Mendelian randomization identified 14 overlapping candidates, with four genes—PLAU, CD58, PCYOX1, and THBS4—consistently prioritized by machine learning. The dataset supports biomarker discovery and validation, including experimental confirmation in a murine AAA model.

Use Cases

• Prioritizing candidate biomarkers for AAA based on integrated genetic, transcriptomic, and proteomic evidence.
• Validating machine learning models for gene selection using differential expression and Mendelian randomization results.
• Investigating shared genetic architecture between gene expression regulators and AAA susceptibility via Bayesian colocalization.
• Benchmarking discriminatory performance of biomarkers using receiver operating characteristic (ROC) curve analysis.

Strengths

• Integrates data from three public microarray datasets (GSE47472, GSE57691, GSE7084) with batch correction applied.
• Analysis framework combines multiple methodologies: differential expression, two-sample Mendelian randomization, and multiple machine learning approaches.
• Key findings, such as the role of PLAU, were validated in an independent murine AAA model.

Limitations

• Column names and exact row count are not specified in the provided metadata.
• The dataset size is reported inconsistently across entries (16680 and 9600).
• The description is truncated, limiting a full understanding of the dataset's complete contents and conclusions.

Provenance

Source

Integrated analysis of public microarray datasets (GSE47472, GSE57691, GSE7084).

Collection Method

Microarray data merged with batch correction, analyzed via Mendelian randomization, machine learning, and validated in a murine model.

Freshness

Last updated 2026-05-29