Developing machine-learning techniques to discover a system-view on Hypertension
Abstract Winner: Precision Medicine & Drug Discovery
Author: Alon Botzer
Background: Essential hypertension is a condition affecting over 1 billion people worldwide, commonly associated with serious illnesses such as stroke, cardiovascular disease, diabetes mellitus and renal disease; hence it is considered "the silent killer". Since hypertension is a complex trait disease, its causes and underlying mechanisms remain poorly understood. The consequences of this condition and its growing prevalence are primary drivers for innovative methods required to further explore new directions that can promote effective treatment and reduce the likelihood of occurrence of this complex condition.
Goal: To study essential hypertension on a system perspective, making use of bioinformatics tools in order to gain new insights that are not pronounced when focusing on a details-based resolution. This can provide a basis for accurate treatment and targeted drug discovery.
Methods: A] Developing unsupervised machine-learning techniques to construct a protein-protein interaction network graph for hypertension associated genes, in order to identify key elements that may play a significant role in this constellation, based on graph centrality analysis. Enriched gene regulatory elements (transcription factors and microRNAs) were extracted by motif finding techniques and knowledge-based tools that could enhance the understanding of the nature of hypertension system regulation. B] Creation of an organism-to-gene orthology map based on pairwise ortholog scores relative to human. Implementing a two-way hierarchical clustering on map scores suggests the evolutionary order by which the circulatory system has evolved. C] Comparing gene targets of hypertension-indication drugs with targets of hypertension-induced side-effect drugs. This analysis made by supervised machine-learning algorithms clarifies the clinical pathways and mechanisms that influence hypertension.
Results: This combined approach yielded a new insight regarding elements that regulate the hypertension gene-network on the translation and post-transcription levels – SP1, EZH2, miRNA27 and miRNA548C, as well as highlighting the central role that Insulin may play in inducing hypertension. The evolutionary analysis confirms the developmental order of the blood circulatory mechanism in respect to affiliated genes. The pharmacological study confirms the role of dopaminergic and adrenergic receptors as major pathways affecting hypertension and strengthens the understanding regarding the causes of hypertension as a prevalent consequential side-effect among many common drugs.
Discussion: We view blood pressure regulation module as a system-of-systems comprising several contributing sub-systems and pathways rather than a single mechanism. Such a regulatory pattern is affected by a large number of distributed elements that could be considered for therapeutic purposes. To our surprise, Insulin proved to be significantly central, suggesting a primary role in hypertension. This finding also highlights the tight link between essential hypertension and the metabolic syndrome diseases.
Co Author/Co-Investigator Names/Professional Title: Eng. Alon Botzer, Ph.D. candidate, Bar-Ilan University, Faculty of Life Sciences; Prof. John Moult, Ph.D., University of Maryland, Institute for Bioscience and Biotechnology Research; Prof. Ron Unger, Ph.D., Bar-Ilan University, Head of Biomedical Informatics Department