estimate.null.distribution.correlation: estimate.null.distribution.correlation

In iDOS: Integrated Discovery of Oncogenic Signatures

Description

Function to estimate probability of observing correlations as high as observed using a feature list of interest

Usage

estimate.null.distribution.correlation(
  exp.data = NULL, 
  cna.data.log2 = NULL, 
  corr.threshold = 0.3, 
  corr.direction = "two.sided", 
  subtypes.metadata = NULL, 
  feature.ids = NULL, 
  observed.correlated.features = NULL, 
  iterations = 50, 
  cancer.type = NULL, 
  data.dir = NULL
  )

Arguments

exp.data	Feature by sample mRNA abundance matrix
cna.data.log2	Feature by sample CNA log ratio matrix
corr.threshold	Threshold for Spearman's Rho to consider a feature as candidate driver
corr.direction	Whether to include positively (greater), negatively (less) or both (two.sided) correlated features. Defaults to two.sided
subtypes.metadata	Subtypes metadata list. Contains at least subtype specific samples
feature.ids	Vector of features to be used to estimate correlation
observed.correlated.features	List of features that were found to be correlated for subtypes of a given cancer type
iterations	Number of random permutations for estimating p value
cancer.type	Name of the cancer type or dataset
data.dir	Path to output directory where the randomisation results will be stored

Value

1 if successful

Author(s)

Syed Haider

See Also

estimate.expression.cna.correlation

Examples

# load test data
x <- get.test.data(data.types = c("mRNA.T", "CNA"));

# temporary output directory
tmp.output.dir <- tempdir();

# estimate mRNA and CNA correlation for each cancer/disease type
correlated.features <- estimate.expression.cna.correlation(
  exp.data = x$mRNA.T$BLCA, 
  cna.data.log2 = x$CNA.log2$BLCA, 
  corr.threshold = 0.3, 
  corr.direction = "two.sided", 
  subtypes.metadata = list(
    "subtype.samples.list" = list("All" = colnames(x$mRNA.T$BLCA))
    ), 
  feature.ids = rownames(x$mRNA.T$BLCA), 
  cancer.type = "BLCA", 
  data.dir = paste(tmp.output.dir, "/data/BLCA/", sep = ""),
  graphs.dir = paste(tmp.output.dir, "/graphs/BLCA/", sep = "")
  );

# estimate NULL distribution
estimate.null.distribution.correlation(
  exp.data = x$mRNA.T$BLCA,
  cna.data.log2 = x$CNA.log2$BLCA, 
  corr.threshold = 0.3, 
  corr.direction = "two.sided", 
  subtypes.metadata = list(
    "subtype.samples.list" = list("All" = colnames(x$mRNA.T$BLCA))
    ), 
  feature.ids = rownames(x$mRNA.T$BLCA), 
  observed.correlated.features = correlated.features$correlated.genes.subtypes, 
  iterations = 50, 
  cancer.type = "BLCA", 
  data.dir = paste(tmp.output.dir, "/data/BLCA/", sep = "")
  );

iDOS documentation built on May 2, 2019, 2:28 p.m.