FGNet (version 3.6.2)

fea_gage: FEA - GAGE

Description

Performs the functional enrichment analysis and clustering through GAGE [1] (GSEA).

Usage

fea_gage(eset, refSamples, compSamples, geneIdType, geneLabels=NULL, 
    organism = "Hs", 
    annotations = c("GO_BP", "GO_MF", "GO_CC", "KEGG", "REACTOME"), 
    geneSets = NULL, 
    sameDirection = FALSE, 
    onlyEssentialTerms = TRUE, 
    compareType = "as.group", 
    jobName = NULL, ...)

Arguments

eset

expressionSet or expression matrix.

refSamples

numeric. Index of the samples to use as reference (control).

compSamples

numeric. Index of the samples to analyze.

geneIdType

character. Type of gene identifier should be the same as the one provided in the geneSets, or available in the organism package.

geneLabels

named character vector. Gene name or label to use in the report/plots instead of the original gene ID. The vector names should be the gene ID and the content of the vector the gene label. The resulting geneTermSets table will contain the original gene ID column (geneIDs) and the label column (Genes).

organism

two letter code for the organism. See: data(organisms);organisms

annotations

character vector. Annotation spaces to select from the provided geneSets. Set to NULL to use the geneSets as is (i.e. geneSets not split/named by annotation)

geneSets

geneSets. If NULL geneSets are calculated automatically based on the organism, gene ID and annotations. The geneSets can also be provided from a previous execution or loaded from a .gtm file. i.e.: readList("c2.cp.v4.0.symbols.gmt")

sameDirection

logical. Should all the genes in the geneSet be altered in the same direction (up/down)?

onlyEssentialTerms

logical. Wether to simplify the results and keep only the essential terms in the clusters.

compareType

character: 'as.group', 'unpaired', '1ongroup'... See GAGE for details.

jobName

character. Folder name and prefix for the files.

...

other arguments to pass to GAGE.

Value

• Invisible list with the folowing fields: queryArgs list with the arguments for the query.

  clusters data.frame containing the clusters and their information:

  • Cluster: Cluster ID.
  • nGenes: Number of genes in the cluster.
  • dir: Direction in which the term/pathway is altered (Up/Down).
  • Genes: Genes in the cluster.
  • Terms: Terms in the cluster.
  geneTermSets data.frame containing the gene-term sets that support each cluster.
  • Cluster: Number (id) of the cluster the gene-term set belongs to.
  • essentialSet: Logical. Is the pathway selected as essential?
  • dir: Direction in which the term/pathway is altered (Up/Down).
  • Terms: Term in the gene-term set.
  • Genes: Genes in the gene-term set.
  • GenesIDs: In case GeneLabels was provided, original gene ID.
  • Other stats provided byGAGE: p.geomean, stat.mean, p.val, q.val, set.size
  fileName: .txt file with the FEA results. genesFC: Fold change.

code

vignette("FGNet-vignette")

References

[1] Luo W, Friedman MS, Shedden K, Hankenson KD, Woolf PJ (2009) GAGE: generally applicable gene set enrichment for pathway analysis. BMC Bioinformatics. 10:161. URL: http://www.bioconductor.org/packages/release/bioc/html/gage.html

See Also

• fea_david()(Requires internet connection)

fea_gtLinker() & fea_gtLinker_getResults() (Requires internet connection) fea_topGO() To import results from a previous/external FEA analysis: format_david(), format_results() and readGeneTermSets().

Examples

# Load expressionSet:
library(gage)
data(gse16873)

# Load gene labels?
library(org.Hs.eg.db)
geneSymbols <- select(org.Hs.eg.db,columns="SYMBOL",keytype="ENTREZID", 
    keys=rownames(gse16873))
head(geneSymbols)
table(table(geneSymbols$ENTREZID)) # All need to be unique identifiers

geneLabels <- geneSymbols$SYMBOL
names(geneLabels) <- geneSymbols$ENTREZID
head(geneLabels)

# FEA:
results <- fea_gage(eset=gse16873, 
                    refSamples=grep('HN',colnames(gse16873), ignore.case =T), 
                    compSamples=grep('DCIS',colnames(gse16873), ignore.case=T), 
                    geneIdType="ENTREZID", geneLabels=geneLabels, organism="Hs",
                    annotations="REACTOME")


# To continue the workflow... (see help for further details)
FGNet_report(results)

incidMat <- fea2incidMat(results)
functionalNetwork(incidMat)