R/gene_stat.R

In imsig: Immune Cell Gene Signatures for Profiling the Microenvironment of Solid Tumours

#' @title General stastitics of ImSig analysis
#' @description [Total genes in ImSig]: The total number of genes in ImSig list. [No. of ImSig genes in user dataset]: The number of ImSig genes found in user's dataset. Like all signatures, ImSig works best when this overlap is high, preferably over 75%. [Median correlation of ImSig genes]: This is the number of remaining ImSig genes after \code{\link{feature_select}}. If this number drops drastically or if the median correlation is low, it may indicate the absence of the particular cell type in users dataset. [Median correlation of feature selected ImSig genes]: These values again can be used gauge the presence or absence of a cell type. As ImSig genes were designed to be co-expressed when the cell type is present, poor correlations may indicate absence of the cell type in the dataset. A network graph (\code{\link{plot_network }}) may be generated to visually confirm the observation.
#' @param exp Dataframe of transcriptomic data (natural scale) containing genes as rows and samples as columns. Note: Gene names should be set as row names and duplicates are not allowed. Missing values are not allowed within the expression matrix. Check example- head(example_data): \code{\link{example_data}}.
#' @param r Use a value between 0 and 1. Default is 0.6. This is a user defined correlation cut-off to perform feature selection (\code{\link{feature_select}}). Feature selection aids to enrich the prediction of relative abundance of immune cells by filtering off poorly correlated ImSig genes. To get an idea of what cut-off to use check the results of (\code{\link{gene_stat}}) and choose a cut-off that displays high median correlation and maintains a high proportion of genes after feature selection.
#' @return Dataframe of general statistics of ImSig analysis.
#' @import stats
#' @examples
#' gene_stat (exp = example_data, r = 0.7)
#' @seealso \code{\link{feature_select}}
#' @export

gene_stat <- function(exp, r = 0.6){
  sig_total <- sig
  exp <- pp_exp(exp)
  sig <- pp_sig(exp)
  fg <- feature_select(exp, r)
  e_cor_data <- fastCor(t(exp))
  diag(e_cor_data) = NA
  genestat <- data.frame(matrix(nrow = 5))
  genestat <- genestat[,-1]
  g <- row.names(exp)
  for (i in levels(sig$cell)){
    sig_total_sub <- sig_total[sig_total$cell %in% i,]
    sig_len <- length(sig_total_sub$gene)
    sig_sub <- sig[sig$cell %in% i,]
    user_len <- length(sig_sub$gene[sig_sub$gene %in% g])
    feature_sub <- sig_sub[as.character(sig_sub$gene) %in% fg,]
    feature_len <- length(feature_sub$gene)
    sig_cor <- e_cor_data[as.character(sig_sub$gene), as.character(sig_sub$gene)]
    s_cor <- round(median(apply(sig_cor, 1, median, na.rm = T)), 2)
    feature_cor <- e_cor_data[as.character(feature_sub$gene), as.character(feature_sub$gene)]
    f_cor <- round(median(apply(feature_cor, 1, median, na.rm = T)), 2)
    stat <- data.frame(c(sig_len, user_len, feature_len, s_cor, f_cor))
    genestat <- cbind(genestat,stat)
  }
  row.names(genestat) <- c("Total genes in ImSig", "No. of ImSig genes in user dataset", "No. of ImSig genes after feature selection", "Median correlation of ImSig genes", "Median correlation of feature selected ImSig genes")
  names(genestat) <- levels(sig$cell)
  return(genestat)
}