Machine learning analytics identify gene expression signatures of severe hypertension in African-Americans

Authors

• C Oguz
• AR Davis
• GH Gibbons

Abstract

Hypertension (HTN) is a cardiovascular disease risk factor that substantially increases the risk for heart attack, chronic kidney disease, and stroke, if left untreated. In U.S., HTN is significantly more prevalent among African Americans in comparison with other ethnicities. While increased age and obesity are the strongest risk factors leading to treatment-resistant HTN, the differences between controlled and treatment-resistant HTN in terms of their underlying molecular pathophysiology remain unclear. We implemented a predictive modeling approach to study hypertensive cases and normotensive controls among 180 African-American patients from the Minority Health Genomics and Translational Research Bio-Repository Database (MH-GRID) Network Study. Model inputs included clinical data and RNA-Seq based gene expression data. By integrating these data sets into machine learning models and interaction networks, we identified the anatomical, adaptive, neural, hemodynamic, endocrine, and humoral processes predictive of HTN, and its treatment-resistant and controlled subphenotypes. The presence of several stress response related pathways and processes within the molecular signature of treatment-resistant HTN suggested that the resistance against antihypertensive treatment could be a result of advanced "vascular age". Our findings shed light into distinct molecular signatures and biological processes associated with the controlled and treatment-resistant HTN subphenotypes. Biological pathways enriched within the most predictive network modules suggested that different mechanisms and groups of genes were predictive of these two subphenotypes. This study describes a machine learning based framework that uses genomic data to generate biological insights, which can guide future clinical and pharmacogenomic studies in targeting specific pathways to improve treatment of HTN.

Scientific Focus Area: Systems Biology

This page was last updated on Friday, March 26, 2021