This function “smooths” gene expression data to assist in the identification of regional expression biases.
the expression set to analyze
an associated chromLoc annotation object
a character vector specifying the chromosomes to analyze
a vector containing the index of reference samples from which to make comparisons. Defaults to NULL (internally referenced samples
boolean - re-center gene expression matrix columns. Helpful if
a function to summarizes/aggregates gene
expression values that map to the same locations. Defaults to the
maximum absolute value
smoothing function to use - either
additional paramaters to pass along to the smoothing function
reb returns an eset that contains predictions of regional expression
bias using data smoothing approachs. The exprSet is separated into
subsets based on the
genome chromLocation object and the gene
expression data within the subsets is organized by genomic location
and smoothed. In addition, the
approx function is used to
estimate data between any missing values. This was implimented so the
function follows the ‘principles of least astonishment’.
Smoothing approachs are most straightforwardly applied by comparing a
set of test samples to a set of control samples. For single color
experiments, the control samples can be specified using the
ref argument and the comparisons are generated internal to the
reb function. This argument can also be used for two-color
experiments provided both the test and control samples were run against
a common reference.
If multiple clones map to the same genomic locus the
argument can be used to summarize the overlapping expression
values to a single summarized value. This is can be helpful in two
situtations. First, the
functions do not allow for duplicate values. Currently, if duplicate
values are found and these smoothing functions are used, the duplicate
values are simply discard. Second, if 50 copies of
the actin gene are present on a the array and actin changes expression
under a given condition, it may appear as though a regional expression
bias exists as 50 values within a region change expression.
Summarizing the 50 expression values to a single value can partially
correct for this effect.
The idiogram package can be used to plot the regional expression bias.
Furge KA, Dykema KJ, Ho C, Chen X. Comparison of array-based comparative genomic hybridization with gene expression-based regional expression biases to identify genetic abnormalities in hepatocellular carcinoma. BMC Genomics. 2005 May 9;6(1):67. PMID: 1588246
MCR eset data was obtained with permission. See PMID: 15377468
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
# The mcr.eset is a two-color gene expression exprSet # with cytogenetically complex (MCR) and normal # control (MNC) samples which are a pooled-cell line reference. data("mcr.eset") data(idiogramExample) ## Create a vector with the index of normal samples norms <- grep("MNC",colnames(mcr.eset@exprs)) ## Smooth the data using the default 'movbin' method, ## with the normal samples as reference cset <- reb(mcr.eset@exprs,vai.chr,ref=norms,center=TRUE) ## Display the results with midiogram midiogram(cset@exprs[,-norms],vai.chr,method="i",dlim=c(-5,5),col=.rwb)