Differential expression analysis of probe-set data

Perform two-group differential expression analysis using "limma".

Usage

limma.pbset(data, group.id, group.id.level = c("E", "V"), pbset.id = NULL)

Arguments

data

dataset to be analyzed. The dataset must have rows as unique probe-sets and columns as samples.

group.id

a vector of sample-group labels for each sample of the dataset. It must be a 2-level non-numeric factor vector.

group.id.level

a vector of sample-group label level. It must have two and only two elements and the first element is the reference. By default, group.id.level = c("E", "V"). That is in our study, we compare endometrial tumor samples to ovarian tumor samples, with endometrial as our reference.

pbset.id

a vector of unique probe-set names. By default, pbset.id = NULL for it to be the row names of the dataset.

Value

a data frame with differential expression analysis results, group means and group standard deviations, for each unique probe-set.

References

Ritchie M., Phipson B., Wu D., Hu Y., Law C., Shi W. and Smyth G. (2015). "limma powers differential expression analyses for RNA-sequencing and microarray studies." Nucleic Acids Research, 43(7), pp. e47.

Examples

uhdata.psl <- med.sum.pbset(data = uhdata.pl,
                            num.per.unipbset = 10)
ctrl.genes <- unique(rownames(uhdata.pl))[grep("NC", unique(rownames(uhdata.pl)))]

uhdata.psl.nc <- uhdata.psl[!rownames(uhdata.psl) %in% ctrl.genes, ]

group.id <- substr(colnames(uhdata.psl.nc), 7, 7)
group.id.level <- levels(as.factor(group.id))

limma.fit.uhdata<- limma.pbset(data = uhdata.psl.nc,
                               group.id = group.id,
                               group.id.level = group.id.level)
                               table(limma.fit.uhdata$P.Value < 0.01,
                               dnn = "DE genes")
#> DE genes
#> FALSE  TRUE 
#>   159    16