Abstract 16548: A Multiscale Approach to Genotype-Phenotype Relationships of Arrhythmias Reveals Novel Disease Genes and Emergent Properties
Significant efforts have been dedicated to the identification of genes underlying complex, common diseases with the goal of predicting disease risk and propensity for disease progression. Although significant advances have been made connecting genes to phenotypes, our understanding of the mechanisms of complex diseases is incomplete and our ability to predict disease risk is inadequate. Here we combine genomics with electrophysiology models, using protein interaction networks as the intermediary to explain genotype-phenotype relationships of arrhythmias. Furthermore, we provide experimental validation of novel metabolic genes contributing to arrhythmogenesis using zebrafish and reverse genetics.
We identified previously unknown disease modules and genes associated with arrhythmias using both computational and experimental approaches. Genome-wide association and Mendelian genetic studies were used to identify genes associated with disease. These genes were used as seed nodes to construct interaction networks representing various arrhythmias and ECG states. Integrated analysis of the genome scale networks and dynamic action potential models revealed novel emergent phenomenon which enabled us to classify disease genes with a p-value of 6.3x10-11 and final quantization and topographic errors of 1.06 and 0.03. We predicted several additional genes that are likely to result in arrhythmia, in particular members of specific metabolic pathways. We also confirmed a significant role for several members of the propanoate pathway using the zebrafish model.
Separately, we discovered fragile electrophysiological parameters using sensitivity analysis techniques applied to simulations of dynamic ordinary differential equation models of cardiac action potentials. This modeling efficiently classifies many genes as fragile with previously determined gene specific estimated predictive values > 90% and demonstrates substantial enrichment for genomic regions associated with cardiac electrical parameters over the remainder of the genome. These data support the hypothesis that variants in genes associated with fragile electrophysiological responses are responsible for the common heritable variation in these parameters.
- © 2013 by American Heart Association, Inc.