epiNEM: Epistatic NEMs - main function.

In epiNEM: epiNEM

Description

This function contains the inference algorithm to learn logical networks from knock-down data including double knock-downs.

Usage

epiNEM(
  filename = "random",
  method = "greedy",
  nIterations = 10,
  nModels = 0,
  random = list(single = 4, double = 1, reporters = 100, FPrate = 0.1, FNrate = 0.1,
    replicates = 1),
  ltype = "marginal",
  para = c(0.13, 0.05),
  init = NULL
)

Arguments

filename	A binary, tab-delimited matrix. Columns: single and double knockdowns. Rows: genes showing effect or not? Default: random; artificial data is generated to 'random' specifications
method	greedy or exhaustive search. Default: greedy
nIterations	number of iterations. Default: 10
nModels	number of Models. Default: 0
random	list specifying how the data should be generated: no. of single mutants, no. of double mutants, no. of reporterGenes, FP-rate, FN-rate, no. of replicates
ltype	likelihood either "marginal" or "maximum"
para	false positive and false negative rates
init	adjacency matrix to initialise the greedy search

Value

List object with an adjacency matrix denoting the network, the model of the silencing scheme (rows are knock-downs, columns are signalling genes), a string with the inferred logial gates, a column indices denoting position of logical gates, the log transformed likelihood and the effect reporter distribution (rows are the signalling genes including the null node).

Author(s)

Madeline Diekmann

See Also

nem

Examples

data <- matrix(sample(c(0,1), 100*4, replace = TRUE), 100, 4)
colnames(data) <- c("A", "A.B", "B", "C")
rownames(data) <- paste("E", 1:100, sep = "_")
res <- epiNEM(data, method = "exhaustive")
plot(res)

epiNEM documentation built on Nov. 8, 2020, 6:51 p.m.