Premium
HeteroPath: A Pathway‐based Computational Modeling Approach to Identify Tissue‐Specific Gene Expression Networks
Author(s) -
Jambusaria Ankit,
Klomp Jeff,
Hong Zhigang,
Rafii Shahin,
Malik Asrar B,
Rehman Jalees
Publication year - 2017
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.31.1_supplement.927.4
Subject(s) - biology , transcriptome , gene expression , regulation of gene expression , gene expression profiling , gene regulatory network , wnt signaling pathway , gene , microbiology and biotechnology , computational biology , transcription factor , signal transduction , genetics
Computational analysis of tissue‐specific gene expression may provide important insights into disease mechanisms and organ specific therapeutic targets. Here we applied our novel approach, HeteroPath, to characterize endothelial cell (EC) heterogeneity from an inherently unbiased perspective. Endothelial cells (ECs) in distinct vascular beds exhibit significant heterogeneity in structure and function as well as propensity for vascular disease. Understanding the molecular basis of tissue‐specific endothelial gene expression signatures and networks may provide important insights into vascular bed‐specific disease mechanisms. In a cohort of freshly isolated mouse ECs from three distinct tissues (heart, brain, and lung) we analyzed the transcriptomic heterogeneity of endothelial cells by comparing a traditional parametric gene set enrichment analysis (PGSEA) to our novel gene interaction network model that assesses the bi‐directional gene expression state (HeteroPath). Ranking gene expression levels in the context of established signaling pathways allowed us to discover transcriptomic signatures specific to a vascular bed. We further characterized the tissue‐specific signatures by constructing transcriptional networks consisting of the identified heterogeneous genes and their regulatory transcription factors as determined by motif enrichment analysis. PGSEA predominantly identified upregulation of amino acid metabolism pathways in the brain endothelium such as phenylalanine, tryptophan and tyrosine metabolism. HeteroPath analysis, on the other hand, identified transcriptomic upregulation of signaling pathways in the brain such as the Wnt signaling and the adherens junction pathways when compared to lung or heart endothelium. A novel unbiased computational approach centered on analyzing gene expression within the context of pathways establishes tissue‐specific gene expression signatures. As HeteroPath assesses bi‐directional gene expression within pathways, it provides a more comprehensive description of cell heterogeneity and identifies previously unrecognized tissue‐specific therapeutic targets. Support or Funding Information This work was supported in part by the National Institutes of Health: R01‐GM094220, R01‐HL118068, T32HL007829 and R01HL090152. Abstract FigureFirst, we preprocess and normalize the freshly isolated endothelial gene expression data (Series GSE47067) and then we integrate in the pathway/gene set data. Then, we statistically evaluated each KEGG pathway using PGSEA and the novel HeteroPath algorithms to identify tissue‐specific pathways.