In jokergoo/rGREAT: GREAT Analysis - Functional Enrichment on Genomic Regions

library(knitr)
knitr::opts_chunk$set(
    error = FALSE,
    tidy  = FALSE,
    message = FALSE,
    warning = FALSE,
    fig.width = 5,
    fig.height = 5,
    fig.align = "center"
)
options("width" = 100)

Let's first do a combination analysis with rGREAT and simplifyEnrichment. Let's say, you have a list of genomic regions of interest (in the following example, we use a list of transcription factor binding sites). You do a GO enrichment analysis with rGREAT and visualize the enrichment results with simplifyEnrichment.

library(rGREAT)
df = read.table(url("https://raw.githubusercontent.com/jokergoo/rGREAT_suppl/master/data/tb_encTfChipPkENCFF708LCH_A549_JUN_hg19.bed"))
# convert to a GRanges object
gr = GRanges(seqnames = df[, 1], ranges = IRanges(df[, 2], df[, 3]))
res = great(gr, "BP", "hg19")
tb = getEnrichmentTable(res)
head(tb)

library(simplifyEnrichment)
sig_go_ids = tb$id[tb$p_adjust < 0.001]
cl = simplifyGO(sig_go_ids)
head(cl)
table(cl$cluster)

The plot looks not bad. Next you may want to get the gene-region associations for significant GO terms. For example, the association for the first GO term:

getRegionGeneAssociations(res, cl$id[1])

There are more than 600 significant GO terms, and perhaps you don't want to obtain such gene-region associations for every one of them.

As simplifyEnrichment already clusters significant GO terms into a smaller number of groups, you may switch your need of obtaining the associations from single GO terms to a group of GO terms where they show similar biological concept. In the new version of rGREAT (only available on GitHub currently), the second argument can be set as a vector of GO IDs. So, for example, to get the gene-region association for GO terms in the first cluster:

getRegionGeneAssociations(res, cl$id[cl$cluster == 1])

SessionInfo

sessionInfo()

jokergoo/rGREAT documentation built on March 28, 2024, 5:31 a.m.