R/enrichment.R

In rwcrocker/SeuratAddons: Auxillary Functions for Seurat

### Functions for enrichR

#' Clean the dirty data output of enrichr()
#' @import enrichR
#' @import dplyr
#' @import magrittr
#' @import stringr
#' @param df EnrichR output dataframe
#' @return Dataframe with cleaner data and statistical cutoff
#' @export

clean_enrichR = function(df, padj_cutoff = 0.05){
  enrichr_vars = c("Term", "Overlap", "P.value", "Adjusted.P.value", "Odds.Ratio", "Combined.Score", "Genes")
  is_enrichr = all(enrichr_vars %in% colnames(df))
  if(!is_enrichr){stop("Data.frame doesn't look like an enrichR output (Column Names)")}

  df = df %>%
    arrange(Adjusted.P.value) %>%
    filter(Adjusted.P.value < padj_cutoff)
  df$Term = factor(df$Term, levels = df$Term)

  # Correct gene symbols and reconcatonate
  df$Genes = df$Genes %>%
    str_split(pattern = ";") %>%
    lapply(FUN = function(i){
      i = stringr::str_to_title(i)
      i = paste(i, collapse = ",")
      return(i)
    })

  # Parse/Evaluate the fraction to a numeric
  df$Overlap = sapply(df$Overlap, function(i){eval(parse(text = i))})

  df = df %>%
    select(Term, Overlap, P.value, Adjusted.P.value, Odds.Ratio, Combined.Score, Genes)

  return(df)
}

#' Plot the output of the enrichR dataframe
#' @import ggplot2 dplyr
#' @import dplyr
#' @import magrittr
#' @import ggprism
#' @param df EnrichR output dataframe cleaned with clean_enrichR
#' @param ntop Number of top terms to plot
#' @param fill_by Variable to fill with
#' @param title Plot title
#' @param term_char Max number of characters before truncating term
#' @return A ggplot2 plot of top enrichement hits
#' @export
plot_enrichr = function(df, ntop, fill_by = "Overlap", title = NULL, term_char=100){
  fill_by_legit = fill_by %in% colnames(df)
  if(!fill_by_legit){stop("Provided 'fill_by' is not a column of provided data.frame.")}

  df = df %>%
    arrange(desc(Combined.Score)) %>%
    slice(1:ntop)

  df$Term = substr(df$Term, start = 1, stop = term_char)
  df$Term = factor(df$Term, levels = rev(df$Term))

  plot = ggplot(df)+
    geom_col(aes(x=Combined.Score, y=Term, fill = get(fill_by)))+
    labs(title = title, fill = fill_by)+
    theme_prism()+
    theme(legend.title = element_text(),
          plot.title = element_text())

  return(plot)
}