affy (version 1.50.0)

makeHmap: A function to create a heatmap-like object or matrix of correlations between miRNA and mRNA data.

Description

This function is intended for use when both miRNA and mRNA data are available for the same samples. In this situation it may be advantageous to compute correlations between the two RNA types, in order to detect mRNA transcripts that are targeted by miRNA.

Usage

makeHmap(mRNAdat, miRNAdat, mRNAlst, mRNAvec = NULL, miRNAvec = NULL,
  chipPkg, header, plot = TRUE, out = TRUE, lhei = c(0.05, 0.95),
  margins = c(5, 8))

Arguments

mRNAdat

An ExpressionSet, data.frame or matrix of mRNA expression values. The row.names for these data should correspond to the manufacturer's probe ID. Currently, the only manufacturer supported is Affymetrix.

miRNAdat

An ExpressionSet, data.frame or matrix of mRNA expression values. The row.names for these data should correspond to the manufacturer's probe ID. Currently, the only manufacturer supported is Affymetrix.

mRNAlst

A list of mRNA probe IDs where the names of each list item are mirBase miRNA IDs. Usually this will be the output from mirna2mrna.

mRNAvec

A numeric vector used to subset or reorder the mRNA data, by column. If NULL, this will simply be 1:ncol(mRNAdat).

miRNAvec

A numeric vector used to subset or reorder the miRNA data, by column. If NULL, this will simply be 1:ncol(miRNAdat).

chipPkg

Character. The name of the chip-specific annotation package (e.g., "hgu133plus2.db").

header

Character. The plot title if a heatmap is output.

plot

Boolean. Should a heatmap be generated?

out

Boolean. Should the matrix of correlation coefficients be output?

lhei

From heatmap.2. This controls the ratio of the heatmap to the key size. If there are very few mRNAs being plotted, you may need to change to something like (0.5, 5).

margins

From heatmap.2. This controls the right and bottom margins, respectively. Increase either value if names get cut off.

Value

• This function will output a numeric matrix if the 'out' argument is TRUE.

Details

As noted above, this function is intended to generate output from simultaneous analyses of miRNA and mRNA data for the same samples, the goal being either a heatmap like plot of correlations, or the data (or both).

If creating a plot, note that if the number of significant mRNA probes is large, the resulting heatmap will have many rows and will not plot correctly on the usual graphics device within R. In order to visualize, it is almost always better to output as a pdf. In addition, the dimensions of this pdf will have to be adjusted so the row names for the heatmap will be legible. As an example, a heatmap with 10 miRNA transcripts and 100 mRNA transcripts will likely need a pdf with a width argument of 6 and a height argument of 25 or 30. It may require some experimentation to get the correct arguments to the pdf function.

Also please note that this function by necessity outputs rectangular data. However, there will be many instances in which a given miRNA isn't thought to target a particular mRNA. Whenever this occurs, the heatmap will have a white cell, and the output data for that combination will be NA.

See Also

mirna2mrna