gsvaRanks: GSVA ranks and scores

In rcastelo/GSVA: Gene Set Variation Analysis for Microarray and RNA-Seq Data

GSVA ranks and scores

Description

Calculate GSVA scores in two steps: (1) calculate GSVA ranks; and (2) calculate GSVA scores using the previously calculated ranks.

Usage

## S4 method for signature 'gsvaParam'
gsvaRanks(param, verbose = TRUE, BPPARAM = SerialParam(progressbar = verbose))

## S4 method for signature 'gsvaRanksParam'
gsvaScores(param, verbose = TRUE, BPPARAM = SerialParam(progressbar = verbose))

Arguments

param	A parameter object of the gsvaRanksParam class.
verbose	Gives information about each calculation step. Default: TRUE.
BPPARAM	An object of class BiocParallelParam specifying parameters related to the parallel execution of some of the tasks and calculations within this function.

Value

In the case of the gsvaRanks() method, an object of class gsvaRanksParam.

In the case of the gsvaScores() method, a gene-set by sample matrix of GSVA enrichment scores stored in a container object of the same type as the input ranks data container. If the input was a base matrix or a dgCMatrix object, then the output will be a base matrix object with the gene sets employed in the calculations stored in an attribute called geneSets. If the input was an ExpressionSet object, then the output will be also an ExpressionSet object with the gene sets employed in the calculations stored in an attributed called geneSets. If the input was an object of one of the classes described in GsvaExprData, such as a SingleCellExperiment, then the output will be of the same class, where enrichment scores will be stored in an assay called es and the gene sets employed in the calculations will be stored in the rowData slot of the object under the column name gs.

References

Hänzelmann, S., Castelo, R. and Guinney, J. GSVA: Gene set variation analysis for microarray and RNA-Seq data. BMC Bioinformatics, 14:7, 2013. DOI

See Also

gsvaParam, gsvaRanksParam, gsva

Examples

library(GSVA)

p <- 10 ## number of genes
n <- 30 ## number of samples
nGrp1 <- 15 ## number of samples in group 1
nGrp2 <- n - nGrp1 ## number of samples in group 2

## consider three disjoint gene sets
geneSets <- list(gset1=paste0("g", 1:3),
                 gset2=paste0("g", 4:6),
                 gset3=paste0("g", 7:10))

## sample data from a normal distribution with mean 0 and st.dev. 1
y <- matrix(rnorm(n*p), nrow=p, ncol=n,
            dimnames=list(paste("g", 1:p, sep="") , paste("s", 1:n, sep="")))

## genes in set1 are expressed at higher levels in the last 'nGrp1+1' to 'n' samples
y[geneSets$set1, (nGrp1+1):n] <- y[geneSets$set1, (nGrp1+1):n] + 2

## build GSVA parameter object
gsvapar <- gsvaParam(y, geneSets)

## calculate GSVA ranks
gsvarankspar <- gsvaRanks(gsvapar)
gsvarankspar
## calculate GSVA scores
gsva_es <- gsvaScores(gsvarankspar)
gsva_es

## calculate now GSVA scores in a single step
gsva_es1 <- gsva(gsvapar)

## both approaches give the same result with the same input gene sets
all.equal(gsva_es1, gsva_es)

## however, results will be (obviously) different with different gene sets
geneSets2 <- list(gset1=paste0("g", 3:6),
                  gset2=paste0("g", c(1, 2, 7, 8)))

## note that there is no need to calculate the GSVA ranks again
geneSets(gsvarankspar) <- geneSets2
gsvaScores(gsvarankspar)

rcastelo/GSVA documentation built on Dec. 17, 2024, 4:51 p.m.