R/ss2n.R

In ahopki14/immunoSeqR: Tools for analyzing data from Adaptive's ImmunoSeq platform

#' ss2n
#' 
#' Signal to noise score for two groups with replcates. This is defined as the difference in
#' means between groups divided by the sum of their standard deviations.
#'
#' @param x,y Matricies of the same length and width >1 
#' @return A vector the same length as x and y with the signal to noise scores
#' @author Alexander Hopkins
#' @export
ss2n <- function(x,y){
    mean_x <- apply(x,1,mean)
    sd_x <- apply(x,1,sd)
    mean_y <- apply(y,1,mean)
    sd_y <- apply(y,1,sd)
    s2n <- (mean_x - mean_y) / (sd_x + sd_y)
    s2n[is.na(s2n)==TRUE] <- 0 # enforce sd > 0
    s2n[is.infinite(s2n)==TRUE] <- 0
#
  s2n
}