eisa (version 1.22.0)

ListHyperGParams-class: Classes for quick GO/KEGG/CHR/miRNA target or other enrichment calculation of multiple gene sets.

Description

These classes extend the HyperGParams class from the Category package to perform enrichment calculation quickly for multiple gene sets.

Usage

"makeValidParams"(object)
"drive"(p)
"drive"(p) <- dri
"ontology"(object)
"ontology"(object) <- go
"conditional"(r)
"conditional"(r) <- cond
"hyperGTest"(p)

Arguments

object,p,r

A ListHyperGParams object.

dri

Logical scalar, whether to store the genes that are in the intersection of the specified gene set and the annotation category.

go

Character scalar, the ontology for GO, possible values: ‘BP’, ‘CC’, ‘MF’.

cond

Logical scalar, whether to perform conditional enrichment calculation. Currently this option is ignored.

Value

makeValidParmas returns another ListHyperGParams instance that has the same class as its arguments'.ontology returns a character vector of length one.conditional returns a logical vector of length one.drive returns a logical vector of length one.

Member functions

Most of these functions are analogous to the ones defined in the Category package, the only difference is that they handle ListHyperGParams objects. makeValidParams validates ListHyperGParams object, in particular, it removes duplicate genes, both from the gene universe and the specified gene sets; and it also makes sure that all genes in the gene sets are included in the universe. ontology can be used to query or set the ontology for enrichment calculated against the GO database. conditional queries or sets whether conditional GO enrichment will be performed. This feature is not implemented yet, see the Category and GOstats packages for a working implementation and more information. drive queries or sets whether the intersections of the gene sets and the universe are stored in the result object. This information can be calculated later as well, but it is faster to store it at the same time when the hypergeometric test is performed.

Details

The ListHyperGParams abstract class extends HyperGParams and allows to specify a list of gene sets for the enrichment calculation instead of a single set.

ListHyperGParams calculates the enrichment much faster than the original HyperGParams classes in the Category package, especially if the calculation is performed against the same gene universe for many gene sets.

ListHyperGParams is an abstract class, it is not possible to instantiate objects from it. Instead, its various extensions must be used: GOListHyperGParams, KEGGListHyperGParams, CHRListHyperGParams and miRNAListHyperGParams.

The various ListHyperGParams objects can be created with the standard new command, by giving all necessary arguments. Please see the examples below.

See Also

Functions for enrichment calculation of ISA modules: ISAGO, ISAKEGG, ISACHR, ISAmiRNA. Perhaps see also the vignette in the GOstats package.

Examples

# GO enrichment, "by hand"
# Load data first
data(ALLModulesSmall)

# Create gene sets
library(hgu95av2.db)
genes <- getFeatureNames(ALLModulesSmall)
entrez <- lapply(genes, function(x) na.omit(unlist(mget(x,
              hgu95av2ENTREZID))))

# Create universe
universe <- na.omit(unlist(mget(featureNames(ALLModulesSmall),
              hgu95av2ENTREZID)))

# Create parameter object
param <- new("GOListHyperGParams", geneIds=entrez, universeGeneIds=universe,
             pvalueCutoff=0.01, drive=FALSE, ontology="BP",
             conditional=FALSE, testDirection="over",
             annotation=annotation(ALLModulesSmall))

# Do the calculation
GOBP <- hyperGTest(param)

# Inspect the result
GOBP
summary(GOBP)[[1]]

# How to create other parameter objects
paramKEGG <- new("KEGGListHyperGParams", geneIds=entrez,
                 universeGeneIds=universe, drive=FALSE,
                 annotation=annotation(ALLModulesSmall))
paramCHR <- new("CHRListHyperGParams", geneIds=entrez,
                 universeGeneIds=universe, drive=FALSE,
                 annotation=annotation(ALLModulesSmall))

# Enrichment with user-supplied categories, we use a list of
# hand-picked genes that are involved in myelin formation
mygenes <- c("YARS", "NFKB2", "NGFR", "CDH1", "NFAT5", "NDRG1", "GAP43", 
 "EGR2", "MSN", "ROCK1", "SREBF2", "SOX10", "FIG4", "EGR1", "PIK3R1", 
 "CDC42", "EDN3", "EDNRB", "NCAM1", "DHH", "OMG", "PMP22", "LAMA4", 
 "MPDZ", "MTMR2", "REL", "S100A1", "ITGA4", "GFAP", "FGF2", "RPSA",
 "CADM1", "CDH19", "DNM2", "PAX3", "SREBF1", "DAG1", "DRP2", "SDC2",
 "MBP", "RELA", "RELB", "JUN", "NAB1", "MOBP", "SKI",  "COL5A2", "RHOA",
 "NFASC", "NEFL", "MPZ", "MAG", "EDNRA", "ERBB4", "LITAF", "MMP2",
 "PLP1", "CDKN1A", "PAK1", "RDX", "GJB1", "LAMA5", "JAM3", "ITGB1",
 "PARD3", "FABP7", "LAMA2", "ERBB3", "CADM4", "FOXO4", "TSPAN31",
 "GPR126", "PTK2", "RAC1", "CDKN2A", "CLDN5", "ID2", "LAMC1", "SOX2",
 "CNTN2", "ERBB2", "NFKB1", "NAB2", "EDN2", "MMP9", "CCND1", "L1CAM",
 "MOG") 

library(org.Hs.eg.db)
myentrez <- na.omit(mapIds(org.Hs.eg.db, mygenes, 'ENTREZID', 
keytype='SYMBOL'))
categories <- list(myelin=myentrez)

data(ALLModules)
genes2 <- getFeatureNames(ALLModules)
entrez2 <- lapply(genes2, function(x) na.omit(unlist(mget(x,
              hgu95av2ENTREZID))))

# Create universe
universe2 <- na.omit(unlist(mget(featureNames(ALLModules),
              hgu95av2ENTREZID)))

paramMY <- new("GeneralListHyperGParams", geneIds=entrez2,
               universeGeneIds=universe2, drive=FALSE,
               annotation=annotation(ALLModulesSmall),
               categories=categories)
MY <- hyperGTest(paramMY)
MY
summary(MY)[[1]]