Exploring IRGs as a Biomarker of Pulmonary Hypertension Using Multiple Machine Learning Algorithms

Area of Science:

• Genomics and bioinformatics
• Biomarker discovery
• Cardiovascular research

Background:

• Pulmonary arterial hypertension (PAH) presents a severe health challenge with high mortality and a lack of effective diagnostic biomarkers.
• Current diagnostic methods for PAH are limited in sensitivity and simplicity for clinical application.

Purpose of the Study:

• To identify novel diagnostic biomarkers for pulmonary arterial hypertension (PAH) through comprehensive genomics research.
• To develop and validate a predictive diagnostic model for PAH utilizing identified biomarkers.

Main Methods:

• Analysis of a large transcriptome dataset incorporating PAH and inflammatory response genes (IRGs).
• Integration and evaluation of 113 machine learning models for diagnostic potential assessment.
• Development of a clinical diagnostic model based on identified hub genes and validation in an animal PAH model.

Main Results:

• The Lasso + LDA machine learning model showed the highest AUC of 0.741.
• Five hub genes (CTGF, DDR2, FGFR2, MYH10, YAP1) exhibited differential expression in PAH patients.
• A diagnostic model using these hub genes achieved an AUC of 0.87, with MYH10 showing an AUC of 0.8.

Conclusions:

• A robust diagnostic model for PAH was successfully developed using inflammatory response genes (IRGs).
• The identified hub genes (CTGF, DDR2, FGFR2, MYH10, YAP1) show promise as novel molecular diagnostic markers for PAH.
• The findings support the potential of genomics and machine learning in advancing PAH diagnostics.