methylKit (version 0.99.2)

getCorrelation: get correlation between samples in methylBase or methylBaseDB object

Description

The functions returns a matrix of correlation coefficients and/or a set of scatterplots showing the relationship between samples. The scatterplots will contain also fitted lines using lm() for linear regression and lowess for polynomial regression.

Usage

getCorrelation(object,method="pearson",plot=FALSE,nrow)
# S4 method for methylBase
getCorrelation(object, method, plot)
# S4 method for methylBaseDB
getCorrelation(object, method = "pearson",
  plot = FALSE, nrow = 2e+06)

Arguments

object

a methylBase or methylBaseDB object

method

a character string indicating which correlation coefficient (or covariance) is to be computed (default:"pearson", other options are "kendall" and "spearman")

plot

scatterPlot if TRUE (default:FALSE)

nrow

a numeric giving the number of lines to read in of methylBaseDB object, defaults to 2e6

Value

a correlation matrix object and plot scatterPlot

Details

The argument 'nrow' is only evaluated if the input is a methylBaseDB object. If 'nrow' is not specified getCorrelation will read the first 2M records of the given object, but if you want to read all records 'nrow' has to be NULL. You should change 'nrow' if using getCorrelation with all records of the methylBaseDB object would take too long.

If the scatter plot is plotted, the red line in the plot is from linear regression fit and the green line is from polynomial regression fit with stats::lowess.

Examples

# NOT RUN {
data(methylKit)

getCorrelation(methylBase.obj,method="pearson",plot=FALSE)

# create methylBaseDB
methylBaseDB.obj <- unite(methylRawList.obj,save.db=TRUE,dbdir="methylDB")

getCorrelation(methylBaseDB.obj,method="pearson",plot=FALSE,nrow=10000)




# remove Database again
rm(methylBaseDB.obj)
unlink("methylDB",recursive=TRUE)

# }