estimate.null.distribution.correlation {iDOS}

estimate.null.distribution.correlation

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 = "")
  );