Simulation Studies Informed by RNA‐seq Data Suggest the Utility of a Multi‐network Bayesian Graphical Model Algorithm for the Study of Hypertension in the Dahl S Rat

Alex Dayton,Francesco Stingo,Chun Yang,Pengyuan Liu,Nadezhda Zheleznova,John Bukowy,Kwang-Woo Ahn,Purushottam Laud,Marina Vannucci,Mingyu Liang,Allen Cowley
DOI: https://doi.org/10.1096/fasebj.29.1_supplement.814.14
2015-01-01
Abstract:In the past, we used static Bayesian graphical models to investigate the role of individual genes in the development of hypertension. Here, we present a simulation study for an experimental design that will allow us to directly compare the networks of gene-gene interactions in the Dahl S (SS) rat and in SS rats with GWAS genes associated with hypertension knocked out (KO). RNA-seq was performed on 8 tissues from an SS rat and two KO rats, which had been fed either a 0.4% or 4% salt diet for 7 days. The list of transcripts whose expression values met a fold change cutoff in the SS but not the KO rats underwent ontology analysis using DAVID, which found pathways that were overrepresented in the list and the number of differentially regulated genes in those pathways; this number was used to modify networks from our previous graphical network study. For each pathway, three networks were created: one which mirrored the previous pathway and two in which five or ten connections were changed. Data was then simulated from these networks and analyzed using a Bayesian network inference algorithm. The algorithm was sufficiently powered by our simulated data set to accurately describe differences between simulated pathways and to elucidate the structure of those pathways, implying that our experimental design has the potential to yield valuable insights into the role of GWAS genes in the onset of hypertension. (HL082798)
What problem does this paper attempt to address?