runSpinglass: Run Spinglass algorithm on a Scored PvalueAnnotation
In SMITE: Significance-based Modules Integrating the Transcriptome and Epigenome

Description Usage Arguments Details Value Note Author(s) References See Also Examples

This function is a function to prepare the data for calling the Spinglass network algorithm.

1 2	runSpinglass(pvalue_annotation, network, random_alpha = 0.05, gam = 0.5, node_alpha = 0.05, maxsize = 500, minsize = 8, num_iterations = 1000, simplify = TRUE)

`pvalue_annotation`	An S4 object of class PvalueAnnotation
`network`	An graph object of class graphNEL or igraph
`random_alpha`	A numeric specifying a threshold with with to determine module signficance after randomization
`gam`	A parameter used by the Spinglass algorithm
`node_alpha`	The proportion of nodes to be used as seeds for the community detection
`maxsize`	The maximum module size
`minsize`	The minimum module size
`num_iterations`	The number of randomizations that will be computed to determine whether the module is significant by chance
`simplify`	A logical (TRUE(DEFAULT)/FALSE) that specifies whether network should be simplified by removing self loops and repeated edges

In the provided Epimods reference, West et al outlined the advantages of using the spin-glass algorithm in the detection of modules. Please consult the reference for more detailed information on the spin-glass algorithm implemented in the package igraph.

Like Epimods, this function employs the spin-glass algorithm implemented in igraph and uses random permutations to assess the "modularity," the number and strength of connected nodes, of a module. However, SMITE scores are interpreted as Chi-square distributed statistics whenever possible, rather than the weighted-T-statistic in Epimods.

An S4 object of class PvalueAnnotation with modules loaded

This function was adapted from a function in the Epimods package that employs the spin-glass algorithm and uses random permutations to assess the "modularity" of a module . The original function was created by West et al.

N. Ari Wijetunga

James West, Stephan Beck, Xiangdong Wang & Andrew E. Teschendorff An integrative network algorithm identifies age-associated differential methylation interactome hotspots targeting stem-cell differentiation pathway. Scientific Reports 3, Article number: 1630 (2013)

https://code.google.com/p/epimods/

FEM runBioNet extractModules plotModule

data(test_annotation_score_data)

#load(system.file("data","Reactome.Symbol.Igraph.rda", package="SMITE"))
 
## NOTE: commented out for example.  See vignette for better explanation ##
#test_annotation <- runSpinglass(pvalue_annotation=test_annotation, 
#network=REACTOME, maxsize=50, num_iterations=10)

plotModule(test_annotation, which_network=6, layout="fr")