SpectroPipeR - example code - functional enrichment analysis (external package) • SpectroPipeR

SpectroPipeR does not incorporate a function for functional enrichment analysis due to the existence of numerous solutions, and users may have a preference for their own approach. However, the following code example illustrates the recommended implementation of statistical results files for a functional enrichment analysis using the Gprofiler2 package. For more information, please refer to the Gprofiler R implementation and the R vignette of Gprofiler2.

Gprofiler2 code suggestion

# install gprofiler2 package
install.packages("gprofiler2")

# load packages
library(gprofiler2)
library(tidyverse)

#set working directory or create RStudio project inside your SpectroPipeR output folder
setwd("[your working directory or SpectroPipeR output folder]")

# load SpectroPipeR statistics --------------------------------------------
#path to statistical_analysis.csv
stats <- read_csv("[path]/statistical_analysis.csv")

# filter statistics -------------------------------------------------------
stats_filtered <- stats %>%
  # filter for ≥2 peptides
  filter(n>=2) %>%
  # filter for FC & q-value
  filter(fold_change_absolute >= 2 & p.fdr <= 0.05)


# get statistical comparisons ---------------------------------------------
comparison <- unique(stats_filtered$slr_ratio_meta)


# generate output folder --------------------------------------------------
dir.create("gprofiler_query_Results",showWarnings = F)


# perform Gprofiler analysis in a for loop over comparisons ---------------
for(i in comparison){

  # filter data for comparison
  grpofiler_input <- stats_filtered %>%
    filter(slr_ratio_meta %in% i)

  # Gene list functional enrichment.
  #
  # run gprofiler
  # query         - character vector of protein IDs
  # significant   - whether all or only statistically significant results should be returned.
  # organism      - Organism names are constructed by concatenating the first letter of the name and
  #                 the family name. Example: human - 'hsapiens', mouse - 'mmusculus'

  grpofiler_res<- gost(query = grpofiler_input$PG.ProteinGroups,
                       significant = T, # only statistically significant results should be returned
                       organism = "hsapiens" # select the right organism
                       )

  # write grpofiler results container
  write_rds(x = grpofiler_res,file = paste0("gprofiler_query_Results/",
                                            str_replace_all(i,"/","_vs_"),".rds"))
  # write grpofiler results table
  write_csv(x = as_tibble(grpofiler_res$result),
            file = paste0("gprofiler_query_Results/",str_replace_all(i,"/","_vs_"),".csv"))

  # write grpofiler results plot
  publish_gostplot(
    p = gostplot(grpofiler_res, interactive = FALSE),
    highlight_terms = NULL,
    filename = paste0("gprofiler_query_Results/",str_replace_all(i,"/","_vs_"),"__plot.png"),
    width = NA,
    height = NA
  )

  # write grpofiler results table
  publish_gosttable(
    gostres = grpofiler_res,
    filename = paste0("gprofiler_query_Results/",str_replace_all(i,"/","_vs_"),"__table.png"),
    highlight_terms = grpofiler_res$result$term_id[which(grpofiler_res$result$p_value<1E-16)]
  )
}

results

Upon execution of the above mentioned code, you will obtain a comprehensive functional enrichment analysis of the statistical results of SpectroPipeR. This analysis is encapsulated within a result container file (.rds), providing a robust and reusable data structure for further exploration. Additionally, the output includes a structured table and a visually compelling plot, offering both tabular and graphical representations of the analysis results.

annotated Manhattan plot of enrichment results

table with the functional enrichment analysis results

table with the functional enrichment analysis results with a p-value below 1E-16