domainScanR is a custom R function that takes gene names or Uniprot IDs and outputs an enrichment plot and a table quantifying enrichment of protein domains

Also available as a shiny app! https://dominico.shinyapps.io/domainScanR/

install.packages("tidyverse")
install.packages("magrittr")
install.packages("httr")
install.packages("viridis")

Check you can load all these libraries successfully. If you see no errors then you are good to go!

library(tidyverse)
library(magrittr)
library(httr)
library(viridis)

run this single line to load the funtion directly from github

devtools::source_url("https://github.com/d0minicO/domainScanR/blob/main/domainScanR.R?raw=TRUE")

• input: the list of genes or proteins to search for enrichment of domains in. A character vector of gene names OR uniprot IDs to search

• data_type (optional): whether gene names or uniprot IDs have been provided. One of c("gene", "uniprot) --- Default is to use gene names (data_type="gene")

• background (optional): which background to use, default is whole proteome, or provide your own list. either "proteome" or your own character vecor of Genes / IDs. --- Default is to use whole proteome (background="proteome")

• stat_test (optional): # statistical test to use. One of c("Fisher", "Chi", "Hypergeometric") --- Default is to use Fisher test (most stringent) (stat_test="Fisher")

• p_adj (optional): p-adjustment method to use. One of p.adjust.methods: c("holm","hochberg","hommel","bonferroni,"BH","BY","fdr","none") --- Default is to use "BH" (Benjamini & Hochberg method) (p_adj="BH)

• thresh (optional): threshold for p-adjusted filtering. Can be any numeric value --- Default is 0.05 (thresh=0.05)

• to_plot (optional): The top N domains to plot, ranked by p value. Can be any numeric value or use NA to turn off plotting --- Default is to plot top 10 (to_plot=10)

• plot_name (optional): character vector name for the title of the plot

Basic usage

## with gene names
names =
  c(
    "CUL3",
    "CUL4B",
    "CUL7",
    "DDRGK1",
    "DDX21",
    "DDX23",
    "DDX54",
    "DEPDC1B",
    "DGCR8",
    "DHX57",
    "ECT2"
    )

domainScanR(input=names)


## with uniprot codes
ids =
  c(
    "P19474",
    "O15164",
    "P14373",
    "Q15654",
    "P62256",
    "Q9H9Y6",
    "P30876",
    "P13727",
    "Q9Y2Y8",
    "O43900",
    "Q6VN20",
    "Q96S59",
    "Q9H871"
    )

dat = domainScanR(input=ids,
            data_type="uniprot")

# save the plot
ggsave(plot=dat[[1]],
       filename = "domainScanR_example_uniprot_IDs.pdf",
       width=8,
       height=4)

# save the table
dat[[2]] %>%
  write.table(file="domainScanR_example_uniprot_IDs_table.tsv",
              quote=F,
              sep="\t",
              row.names=F)

domainScanR(input=ids,
            data_type="uniprot",
            background=bkg_ids,
			plot_name="your_ids_custom_bkg",
			to_plot=25,
			stat_test="Chi",
			thresh="0.01",
			p_adj="FDR")

By Default, the function returns a list

• First item is the plot
• Second item is a tibble with the full table of result of the statistical test --- NOTE: if to_plot=NA is used, just the table is returned

• domain_name : The name of the protein domain.
• interpro_id : The InterPro identifier associated with the protein domain.
• count_input : The number of proteins with the domain in the input list of protein IDs / Gene names.
• total_input : The total number of proteins in the input list of protein IDs / Gene names.
• freq_input : The frequency (percentage) of proteins with the domain in the input list of protein IDs / Gene names.
• count_bkg : The number of proteins with the domain in the background list of protein IDs / Gene names.
• freq_bkg : The frequency (percentage) of proteins with the domain in the background list of protein IDs / Gene names.
• total_bkg : The total number of proteins in the background list of protein IDs / Gene names.
• domainRatio : The ratio of the proteins identified with the domain compared to the total number of proteins with that domain.
• p_adjusted : The p-value from the statistical test, adjusted for multiple comparisons.
• stat_test : The statistical test used to compare the frequency of the domain in the input list and the background list.
• Genes : The gene names associated with the domain in the input list of protein IDs / Gene names.
• IDs : The Uniprot protein IDs associated with the domain in the input list of protein IDs / Gene names.

domainScanR uses the extensive interpro database (https://www.ebi.ac.uk/interpro/) to retreive domain annotations for the reviewed human proteome available on uniprot (2023-04-19, UP000005640) Statistical tests are done in a similar manner to GO terms enrichment. Either Fisher test, Chi-square test, or Hypergeometric tests are supported. Custom backgrounds are supported. P values are by default adjusted using Benjamini & Hochberg method, though other methods can be applied.

dominic.owens ..at?? utoronto.ca