Abstract 21229: Systems Biology Based Approach for Identification of Markers of Unstable Plaque
Background: Atherosclerotic plaque can be detected early in the progression of coronary artery disease; however, identification of rupture prone plaque has been difficult. In an effort to identify candidate genes whose gene products may be used as biomarkers of plaque vulnerability we used a systems biology based approach.
Methods and Results: We constructed an AP Protein-Protein Interaction (AP-PPI) network, using known AP-relevant genes that were identified together with validated PPIs of their encoded proteins according to the Human Protein Reference Database (HPRD). Differentially expressed genes from microarray data encoding molecular profiles of subjects with stable vs. unstable plaques were identified. Proteins encoded by these significantly differentially expressed genes were mapped onto the AP-PPI network. Key statistical and topological measures between significantly differentially expressed genes and the interaction network were computed with the use of automated literature mining tools to provide further annotation to the nodes. Nodes in the network were also classified according to the degree of connectivity; superhubs are represented by nodes with connectivity degree >100, hubs are nodes with connectivity degree > 20 and 2 and < 20; and peripheral-B nodes represent proteins with one interacting partner only. In-silico analysis of the upstream sequences of the differentially expressed genes was conducted to identify key regulatory nodes in the network.
Conclusions: We identified 7 hubs that while not in the curated list, were found to be associated with atherosclerosis based on text and literature mining (CCL18, GRB2, TP53, EGFR, ID3, SGK1, ENO1). Four hubs not associated with the candidate gene list (MYH10, EEF1A1, MCC, RPN2) warrant further study. Genes that are significantly differentially expressed are not always represented by highly-connected nodes, indicating that use of microarray data alone could lead to a higher false non-discovery rate with regard to identification of putative therapeutic targets.
- © 2010 by American Heart Association, Inc.