genBart (version 1.0.1)

normalizeData: Data Normalization

Description

Perform various normalizations of expression data

Usage

normalizeData(meta, norm.method = "mean")

Arguments

meta

list returned by metaData

norm.method

String denoting whether to normalize to the mean or median of all samples or a control group specified in meta. Default is norm.method = "mean".

Value

y1b data frame of baseline samples normalized according to norm.method. NULL if baseline samples are not specified in meta.

y2b data frame of baseline samples normalized to controls according to norm.method. NULL if control samples are not specified in meta.

y1 data frame of all samples normalized according to norm.method.

y2 data frame of all samples normalized to controls according to norm.method. NULL if control samples are not specified in meta.

y3 data frame of all samples normalized to their baseline. NULL if study is not longitudinal or if baseline samples are not specified in meta.

norm.method string describing normalization method used.

Details

This function performs various normalizations of the expression data, depending on the study design and the parameters defined in metaData. For all study designs, the data is normalized to the mean (or median) of all the samples. For cross-sectional studies with controls, an additional normalization to the mean (or median) of the controls is performed. For longitudinal designs, baseline normalization ( subtract out each subject's baseline) and normalization to the mean (or median) of controls (if present) is performed. In addition, separate normalizations on baseline samples is performed.

Examples

# NOT RUN {
# Example data
data(tb.expr)
data(tb.design)

# Use first 100 probes to demonstrate
dat <- tb.expr[1:100,]

# Create desInfo object
meta.data <- metaData(y = dat, design = tb.design, data.type = "microarray", 
                    columnname = "columnname", long = TRUE, sample.id = "sample_id",
                    subject.id = "monkey_id", time.var = "timepoint",
                    baseline.var = "timepoint", baseline.val = 0)

# Normalize and cluster data
data.norm <- normalizeData(meta = meta.data)
# }