R/System_Expression.R

In hyginn/BCB420.2019.ESA: Exploratory Systems Analysis Tools

# System_Expression.R

#' System level expression values.
#'
#' \code{exProf} Returns the log-normalized expression values for gene provided.
#'
#' \code{System_Expression} Provides the mean expression level for each experiment
#' present in the dataset. As well, displays which conditions are up and down
#' regulted significantly.
#'
#' For further information, see details in the \href{../doc/BCB420.2019.ESA-Vignette_McNeil.html}{ System Expression vignette} (or load the vignette with \code{vignette("BCB420.2019.ESA-Vignette_McNeil", package = "BCB420.2019.ESA")}).
#'
#' @param sysname (character) 5-letter code of system being investigated
#' @param myQNXP (matrix) Matrix of expressions. Columns represent experiments and
#' rows individual genes
#'
#' @return (list) A list containing 3 elements. The first the mean log normalized
#' gene expression by condition. Second and third the conditions where the system
#' is significantly up and down regulated respectively.
#'
#' @author \href{https://orcid.org/0000-0002-2821-6204}{Matthew McNei;} (aut)
#'
#' @export

System_Expression<- function(sysname, myQNXP){
  sysDB <- fetchData("SysDB")
  # Get a list of genes present in the given system
  components <- SyDBgetSysSymbols(sysDB, sysname)[[1]]

  # Normalized expression values for each gene in the system
  expressions <- sapply(components, exProf, myQNXP)

  # Find the means and variance of the expression changes caused by each condition
  meanExpression <- rowMeans(expressions, na.rm = T)
  varExpression <- apply(expressions, 1, stats::var, na.rm = T)
  num_observations <- apply(expressions, 1, function(x) sum(!is.na(x)))

  # calculate the one-sided p-values for each experiment
  p_values <- pnorm(meanExpression, sd = sqrt(varExpression/num_observations))
  p_values[p_values > 0.5] <- 1 - p_values[p_values > 0.5]

  # Find the experiments where the system was significantly up
  # or down regulated at the 0.05 level
  UpConditions <- which(meanExpression > 0 & p_values < 0.05)
  DownConditions <- which(meanExpression < 0 & p_values < 0.05)

  experiments <- rownames(expressions)
  # Package results into list for export
  SysExpression <- list(`Mean Expression` = meanExpression,
                        `Upregulated Conditions` = experiments[UpConditions],
                        `Downregulated Conditions` = experiments[DownConditions])

  return(SysExpression)
}

# [END]