YuGene: YuGene: A simple method for comparing gene expression across platforms using a cumulative proportion approach.

Description

YuGene is a simple method for comparison of gene expression generated across different experiments, and on different platforms; that does not require global renormalization, and is not restricted to comparison of identical probes. YuGene works on a range of microarray dataset distributions, such as between manufacturers. The resulting output allows direct comparisons of gene expression between experiments and experimental platforms.

Usage

YuGene(data.prop, progressBar = TRUE)

Arguments

data.prop

a matrix or data.frame of log intensity values, with samples in columns and expression levels in rows. Can be probe or transcript level. Can be raw or previously (i.e. quantile) normalized data.

progressBar

set to FALSE to supress progress bar

Value

returns an object of class `YuGene': a matrix of the same dimensions with each sample transformed to the cumulative proportion (YuGene) metric.

References

Kim-Anh L<U+00EA> Cao, Florian Rohart, Leo McHugh, Othmar Korn, Christine A. Wells. YuGene: A simple approach to scale gene expression data derived from different platforms for integrated analyses. Genomics. http://dx.doi.org/10.1016/j.ygeno.2014.03.001.

Examples

Run this code

# NOT RUN {
data(ascorbate) # gene expression data available in YuGene package
# apply the transform to the data
YuGene.transformed <- YuGene(ascorbate$gene)

# show distributions before and after YuGene
opar <- par()      # make a copy of current settings
par(mfrow=c(1,2))
plot(density(ascorbate$gene[,1]),main='Expression values', xlab='log2 expr.'); 
plot(density(YuGene.transformed[,1]),main='YuGene values',xlab='YuGene value');
par(opar)     # restore original settings

# unadjusted pvals from the quantile normalized data
quant.pvals <- apply(ascorbate$gene,1,function(row){return(t.test(row[1:3],row[4:6])$p.value)})
YuGene.pvals <- apply(YuGene.transformed,1,function(row){return(t.test(row[1:3],row[4:6])$p.value)})
plot(quant.pvals,YuGene.pvals,pch='.',main='comparison of pvals before and after YuGene Transform')
text(0.8,0.2,paste("Pearson cor: ",round(cor(quant.pvals,YuGene.pvals,method='pearson'),digits=3)))

# }