calculate_go_enrichment: Perform gene ontology enrichment analysis

In protti: Bottom-Up Proteomics and LiP-MS Quality Control and Data Analysis Tools

Perform gene ontology enrichment analysis

Description

Analyses enrichment of gene ontology terms associated with proteins in the fraction of significant proteins compared to all detected proteins. A two-sided Fisher's exact test is performed to test significance of enrichment or depletion. GO annotations can be provided to this function either through UniProt go_annotations_uniprot, through a table obtained with fetch_go in the go_data argument or GO annotations are fetched automatically by the function by providing ontology_type and organism_id.

Usage

calculate_go_enrichment(
  data,
  protein_id,
  is_significant,
  group = NULL,
  y_axis_free = TRUE,
  go_annotations_uniprot = NULL,
  ontology_type,
  organism_id = NULL,
  go_data = NULL,
  plot = TRUE,
  label = TRUE,
  plot_cutoff = "adj_pval top10"
)

Arguments

data	a data frame that contains at least the input variables.
protein_id	a character column in the data data frame that contains the protein accession numbers.
is_significant	a logical column in the data data frame that indicates if the corresponding protein has a significantly changing peptide. The input data frame may contain peptide level information with significance information. The function is able to extract protein level information from this.
group	a character column in the data data frame that contains information by which the analysis should be grouped. The analysis will be performed separately for each of the groups. This is most likely a column that labels separate comparisons of different conditions. In protti the asign_missingness() function creates such a column automatically.
y_axis_free	a logical value that specifies if the y-axis of the plot should be "free" for each facet if a grouping variable is provided. Default is TRUE. If FALSE is selected it is easier to compare GO categories directly with each other.
go_annotations_uniprot	recommended, a character column in the data data frame that contains gene ontology annotations obtained from UniProt using fetch_uniprot. These annotations are already separated into the desired ontology type so the argument ontology_type is not required.
ontology_type	optional, character value specifying the type of ontology that should be used. Possible values are molecular function (MF), biological process (BP), cellular component (CC). This argument is not required if GO annotations are provided from UniProt in go_annotations_uniprot. It is required if annotations are provided through go_data or automatically fetched.
organism_id	optional, character value specifying an NCBI taxonomy identifier of an organism (TaxId). Possible inputs include only: "9606" (Human), "559292" (Yeast) and "83333" (E. coli). Is only necessary if GO data is not provided either by go_annotations_uniprot or in go_data.
go_data	Optional, a data frame that can be obtained with fetch_go. If you provide data not obtained with fetch_go make sure column names for protein ID (db_id) and GO ID (go_id) are the same as for data obtained with fetch_go.
plot	a logical argument indicating whether the result should be plotted or returned as a table.
label	a logical argument indicating whether labels should be added to the plot. Default is TRUE.
plot_cutoff	a character value indicating if the plot should contain the top 10 most significant proteins (p-value or adjusted p-value), or if a significance cutoff should be used to determine the number of GO terms in the plot. This information should be provided with the type first followed by the threshold separated by a space. Example are plot_cutoff = "adj_pval top10", plot_cutoff = "pval 0.05" or plot_cutoff = "adj_pval 0.01". The threshold can be chosen freely.

Value

A bar plot displaying negative log10 adjusted p-values for the top 10 enriched or depleted gene ontology terms. Alternatively, plot cutoffs can be chosen individually with the plot_cutoff argument. Bars are colored according to the direction of the enrichment. If plot = FALSE, a data frame is returned. P-values are adjusted with Benjamini-Hochberg.

Examples

# Load libraries
library(dplyr)
library(stringr)

# Create example data
# Contains artificial de-enrichment for ribosomes.
uniprot_go_data <- fetch_uniprot_proteome(
  organism_id = 83333,
  columns = c(
    "accession",
    "go_f"
  )
)

if(!is(data, "character")){

data <- uniprot_go_data %>%
  mutate(significant = c(
    rep(TRUE, 1000),
    rep(FALSE, n() - 1000)
  )) %>%
  mutate(significant = ifelse(
    str_detect(
      go_f,
      pattern = "ribosome"
    ),
    FALSE,
    significant
  ))

# Plot gene ontology enrichment
calculate_go_enrichment(
  data,
  protein_id = accession,
  go_annotations_uniprot = go_f,
  is_significant = significant,
  plot = TRUE,
  plot_cutoff = "pval 0.01"
)

# Calculate gene ontology enrichment
go_enrichment <- calculate_go_enrichment(
  data,
  protein_id = accession,
  go_annotations_uniprot = go_f,
  is_significant = significant,
  plot = FALSE,
)

head(go_enrichment, n = 10)
}

protti documentation built on Jan. 22, 2023, 1:11 a.m.