Using mechanistic Bayesian networks to identify downstream targets of the Sonic Hedgehog pathway

Abstract

BackgroundThe topology of a biological pathway provides clues as to how a pathway operates, but rationally using this topology information with observed gene expression data remains a challenge.ResultsWe introduce a new general-purpose analytic method called Mechanistic Bayesian Networks (MBNs) that allows for the integration of gene expression data and known constraints within a signal or regulatory pathway to predict new downstream pathway targets. The MBN framework is implemented in an open-source Bayesian network learning package, the Python Environment for Bayesian Learning (PEBL). We demonstrate how MBNs can be used by modeling the early steps of the sonic hedgehog pathway using gene expression data from different developmental stages and genetic backgrounds in mouse. Using the MBN approach we are able to automatically identify many of the known downstream targets of the hedgehog pathway such as Gas1 and Gli1, along with a short list of likely targets such as Mig12.ConclusionsThe MBN approach shown here can easily be extended to other pathways and data types to yield a more mechanistic framework for learning genetic regulatory models.

Spatiotemporal regulation of GLI target genes in the mammalian limb bud

“…In contrast, if genes required continued exposure to SHH, they would remain downregulated in wild-type limb buds that received only a brief exposure to SHH. Although previous studies have identified SHH-responsive genes at E10.5 using microarrays (Probst et al, 2011; Shah et al, 2009), we decided to perform a new analysis using RNA-seq on E10.25 limb buds in order to make direct quantitative comparisons (Fig. 4A).…”

“…Whole-genome approaches using DNA microarrays have been used to identify SHH-responsive genes during limb development (Bangs et al, 2010; Hu et al, 2012; McGlinn et al, 2005; Probst et al, 2011; Shah et al, 2009; Vokes et al, 2008). Combining this approach with three-dimensional spatial information from whole-mount embryo in situ hybridization has been successfully used to identify genes regulated in distinct domains (Bangs et al, 2010; Probst et al, 2011; Welten et al, 2011).…”

A Gene Regulatory Network for Root Epidermis Cell Differentiation in Arabidopsis

“…The Bayesian network analysis was performed using the Pebl software environment datasets ( http://jmlr.csail.mit.edu/papers/v10/shah09a.html ) [89] , [90] . All of the microarray datasets from wild-type and knockout mutants (n = 66) were used to define interactions in a 219 node network, which included the 208 core root epidermal genes, 7 genes representing knockout mutants but not part of the core genes, 2 hormone treatments (ACC and IAA), and 2 morphological characters (hair length and degree of hair branching).…”

Computational modelling of the inflammatory response in trauma, sepsis and wound healing: implications for modelling resilience

“…In order to better discern organizational aspects of interacting networks of inflammatory mediators, such as coregulation or autoinduction, a variety of methods have been developed. Hierarchical clustering and Bayesian methods use high-throughput genomic or proteomic data of several time-points and/or conditions to correlate gene expression patterns with function and infer regulatory networks of correlated genes [34][35][36][37]. Focusing on the dynamics of inflammation, we used a simple network analysis method used over discrete intervals of data to analyse the commonality and differences between experimental surgical cannulation trauma þ haemorrhage in mice versus the sham procedure (surgical cannulation only) [25].…”

“…A cohort of 10 000 virtual trauma patients was generated from the 33 patients' individual inflammatory and physiological trajectories. Each virtual patient was then subjected to three insults of trauma: low/intermediate Injury Severity Score (ISS) (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20), intermediate/high ISS (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) and severe ISS (35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45)(46)(47)(48)(49)(50). The in silico distributions of model variables equated with length of stay in the intensive care unit, degree of multiple organ dysfunction and interleukin (IL)-6 area under the curve were in concordance with...…”