Given a set of genes of interest, full bipartite networks with scores (one network for each sample), a significance cutoff for statistical testing, and a hop constraint, BLOBFISH finds a subnetwork of significant edges connecting the genes.

Given a set of genes of interest, full bipartite networks with scores (one network for each sample), a significance cutoff for statistical testing, and a hop constraint, BLOBFISH finds a subnetwork of significant edges connecting the genes.

Usage

RunBLOBFISH(
  geneSet,
  networks,
  alpha,
  hopConstraint,
  nullDistribution,
  verbose = FALSE,
  topX = NULL,
  doFDRAdjustment = TRUE,
  pValueChunks = 100,
  loadPValues = FALSE,
  pValueFile = "pvalues.RDS"
)

Arguments

geneSet

A character vector of genes comprising the targets of interest.

networks

A list of bipartite (PANDA-like) networks, where each network is a data frame with the following format: tf,gene,score

alpha

The significance cutoff for the statistical test.

hopConstraint

The maximum number of hops to be considered between gene pairs. Must be an even number.

nullDistribution

The null distribution, specified as a vector of values.

verbose

Whether or not to print detailed information about the run.

topX

Select the X lowest significant p-values for each gene. NULL by default.

doFDRAdjustment

Whether or not to perform FDR adjustment.

pValueChunks

The number of chunks to split when calculating the p-value. This parameter allows the edges to be split into chunks to prevent memory errors.

loadPValues

Whether p-values should be loaded from pValueFile or re-generated. Default is FALSE.

pValueFile

The file where the p-values should be saved. If NULL, they are not saved and need to be recalculated.

Value

A bipartite subnetwork in the same format as the original networks.