greedyModuleSelection: Greedy Module Selection for a Seed
In krumsieklab/MoDentify: Phenotype-driven module identification

Description Usage Arguments Value References Examples

Greedy Module Selection for a Seed

greedyModuleSelection(nodeNr, graph, data, phenotype, covars = NULL,
  alpha = 0.05, moduleCache = NULL, cacheFolder = NULL,
  better.than.components = TRUE, representative.method = "average",
  scoringFunction = linearScoring)

`nodeNr`	the number of the node, which should be used as a seed.
`graph`	an igraph object, which can be generated with `generateNetwork`. The ID of the nodes must correspond to the name of the variables.
`data`	a `data.table` with three columns: name, sampleID and z-score.
`phenotype`	a vector with the values for a phenotype of interest. It must have the same number of samples as in data.
`covars`	a `data.table` containing the covariates to correct for. The rows for the observations must be in the same order as in the phenotype vector.
`alpha`	significance level (type 1 error) for accepting the modules.
`cacheFolder`	location of the temporary files, where the cache is stored
`better.than.components`	if `TRUE`, modules will only be enlarged and accepted, if they are better than all of their components.
`representative.method`	the method used for the calculation of the module representatives. Currently implemented: "eigenmetabolite" and "average"
`scoringFunction`	a scoring function accepting parameters moduleRepresentatives, phenotype and covars. See `linearScoring`

a list containing the members of the module, the module-score, its regression coefficient for the given phenotype, the score and regression coefficient for the seed, the cache and the consecutive module-scores after adding each new node.

\insertRef

Do2017MoDentify

data(qmdiab.data)
data(qmdiab.annos)
data(qmdiab.phenos)

net.graph <- generateNetwork(data = qmdiab.data, annotations = qmdiab.annos)
data <- data.table(
  sampleID = paste0("sample", 1:dim(qmdiab.data)[1]),
  qmdiab.data
)
data <- melt(data = data, id.vars = "sampleID", variable.name = "name")
data[, z.score := scale(value), by = .(name)]

module <- greedyModuleSelection(
  graph = net.graph, nodeNr = 51, data = data,
  phenotype = qmdiab.phenos$T2D
)