[BUG] Set custom background?

[mkrdonovan]

Fantastic GO enrichment tool!

Wondering if the option to set your own custom background has been implemented or if there are plans to?

Thanks!

[AviMaayan]

You can try this Appyter: https://appyters.maayanlab.cloud/Independent_Enrichment_Analysis/

[mkrdonovan]

Thanks for the tip! Will give it a try--wasn't familiar with Appyters before, but it's a very cool concept. Still, would love to have bg optional in the R Enrichr package :)

[JoyOtten]

I totally agree that it would be a very nice option to have background genes as an option in the R enrichr package.

[julien-roux]

Here is a recent paper illustrating again the risks of not using the correct background for enrichment analyses: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1009935

So this feature would definitely be needed! A lot of biologists use the app without paying attention to this problem....

[d0minicO]

Agreed! I loved how easy the tool was to set up in R using the enrichR package. Now having to use something else because without a custom background it is unusable unfortunately... Anyone worth their salt will ask what background was used, and if you haven't used correct one then you come out with a red face.

[AviMaayan]

It is there. We implemented support for background. The background is already implemented and available in the R package. It will be added to the UI soon.

[d0minicO]

Great! Please could you describe how to use it? I've looked at the vignette and reference manual but can't seem to find any mention of background... Thanks for the great tool!

[lachmann12]

Hi there,

Below is some R code that demonstrates how to perform background correction with the Enrichr API:

(The input file should contain one gene symbol per line)

library(jsonlite)
library(httr)

baseurl <- "https://maayanlab.cloud/speedrichr/api/"

#1. Upload a gene set for enrichment analysis.
#This function returns an identifier that is needed for later calls (userListId)
genelist <- readLines("testinput.txt")
description <- "Example gene list"

ENRICHR_URL <- paste0(baseurl, "addList")
payload <- list(
  list = upload_file("testinput.txt"),
  description = description
)
response <- POST(ENRICHR_URL, body = payload, encode = "multipart")
data <- fromJSON(content(response, as = "text"))
print(toJSON(data, pretty = TRUE))

#2. Upload the background
#The function returns an identifier that we will need for later calls (backgroundid)
backgroundgenes <- readLines("human_genes.tsv")

ENRICHR_URL <- paste0(baseurl, "addbackground")
payload <- list(
  background = backgroundgenes
)
response <- POST(ENRICHR_URL, body = payload)
data_back <- fromJSON(content(response, as = "text"))
print(toJSON(data_back, pretty = TRUE))

#3. List available libraries, we can then select one of these libraries for the enrichment analysis calculation
ENRICHR_URL <- paste0(baseurl, "listlibs")
response <- GET(ENRICHR_URL)
data_libs <- fromJSON(content(response, as = "text"))
print(toJSON(data_libs, pretty = TRUE))

#4. Pass IDs retrieved from step 1 and step 2 and a gene set library from step 3. Here we select a library, and apply the enrichment analysis using the two previously uploaded gene sets.
library_name <- "GO_Biological_Process_2018"

ENRICHR_URL <- paste0(baseurl, "backgroundenrich")
payload <- list(
  userListId = data$userListId,
  backgroundid = data_back$backgroundid,
  backgroundType = library_name
)
response <- POST(ENRICHR_URL, body = payload)
data_ben <- fromJSON(content(response, as = "text"))
print(toJSON(data_ben, pretty = TRUE))

[lachmann12]

The input files are: "testinput.txt" and "human_genes.tsv". The "human_genes.tsv" is the background gene set and "testinput.txt" the gene set to run enrichment against.