enrichment: Enrichment analysis

In bhagwataditya/importomics: Unified statistal Modeling of Omics Data

Enrichment analysis

Description

Are selected genes enriched in pathway?

Usage

enrichment(
  object,
  pathwaydt,
  fit = fits(fdt(object))[1],
  coef = coefs(fdt(object), fit = fit)[1],
  var = abstractvar(fdt(object), fit = fit, coef = coef),
  levels = fdt(object)[[var]] %>% base::levels() %>% extract(-1),
  genevar = "gene",
  genesep = "[ ,;]",
  n = 3,
  verbose = TRUE,
  genes = FALSE
)

Arguments

object	SummarizedExperiment
pathwaydt	pathway data.table
fit	string
coef	string
var	selection fvar
levels	selection levels
genevar	gene fvar
genesep	gene separator (string)
n	number
verbose	whether to msg
genes	whether to report genes

Details

Four enrichment analyses per geneset using the Fisher Exact Test (see four pvalues). Results are returned in a data.table

in	: genes in pathway
in.det	: detected genes in pathway
in.sel	: up/downregulated genes in pathway
in.up(.genes)	: upregulated genes in pathway
in.down(.genes)	: downregulated genes in pathway
out	: genes outside pathway
det	: detected genes (in + out)
sel	: up/downregulated genes (in + out)
up	: upregulated genes (in + out)
down	: downregulated genes (in + out)
p.coef.upDET	: prob to randomly select this many (or more) upregulated genes (among detected genes)
p.coef.downDET	: prob to randomly select this many (or more) downregulated genes (among detected genes)
p.coef.selDET	: prob to randomly select this many (or more) up OR downregulated genes (among detected genes)
p.coef.selGEN	: prob to randomly select this many (or more) up OR downregulated genes (among genome genes)
p.detGEN	: prob to randomly select this many (or more) detected genes (among genome genes)

Examples

# Read
    pathwaydt <- read_msigdt(collections = 'C5:GO:BP')
    file <- download_data('atkin.somascan.adat')
    object <- read_somascan(file, fit = 'limma', coefs = 't1')
    fvars(object) %<>% gsub('EntrezGeneSymbol', 'gene', .)
    object %<>% abstract_fit()
    var <- abstractvar(fdt(object))
    enrichdt1 <- enrichment(object, pathwaydt, var = var)                                   # 2:n factor 
    enrichdt2 <- enrichment(object, pathwaydt, var = var, levels = c('flat', 'down', 'up')) # 1:n factor
    enrichdt3 <-  altenrich(object, pathwaydt)                                              # alternative implementation
    cols <- intersect(names(enrichdt1), names(enrichdt3))
    all(enrichdt1[, cols, with = FALSE]  ==  enrichdt3[, cols, with = FALSE])   # identical

bhagwataditya/importomics documentation built on April 20, 2024, 11:19 p.m.