R/CheckPossibleValuesAreValid.r

In CausalR: Causal network analysis methods

#' @title  check possible values are valid
#' @description
#' Checks if the a given set of possible values for n++, n+-, n-+ and n-- are agree with the predicted and experimental
#' data
#'

#' @param predictionDataStats a vector of predicted results
#' @param experimentalDataStats a vector of observed experimental results
#' @param possibleValues a vector of possible values  n++, n+-, n-+ and n--

#' @return TRUE if and only if the given vector of possible values is valid

CheckPossibleValuesAreValid <- function(predictionDataStats, experimentalDataStats, possibleValues) {
    
    numPositivePredictions <- predictionDataStats[1]
    numNegativePredictions <- predictionDataStats[2]
    numZeroPredictions <- predictionDataStats[3]
    numUpregulatedGenes <- experimentalDataStats[1]
    numDownregulatedGenes <- experimentalDataStats[2]
    numNonResponsiveGenes <- experimentalDataStats[3]
    
    # n_pp = possible value for numCorrectPositivePredictions
    n_pp <- possibleValues[1]
    # n_pm = possible value for numIncorrectPositivePredictions
    n_pm <- possibleValues[2]
    # n_mp = possible value for numIncorrectNegativePredictions
    n_mp <- possibleValues[3]
    # n_mm = possible value for numCorrectNegativePredictions
    n_mm <- possibleValues[4]
    
    # The conditions that need to be fulfilled are a series of inequalities The conditions will be number so that they can matched to the test function
    # Condition 1
    if (((n_pp + n_pm) > numPositivePredictions)) {
        return(FALSE)
    }
    # Condition 2
    if (((n_mp + n_mm) > numNegativePredictions)) {
        return(FALSE)
    }
    # Condition 3
    if (((n_pp + n_mp) > numUpregulatedGenes)) {
        return(FALSE)
    }
    # Condition 4
    if (((n_pm + n_mm) > numDownregulatedGenes)) {
        return(FALSE)
    }
    
    # n_pz (n+0) is the number of genes with positive predictions but are in the set of non-responsive genes in the results.
    n_pz <- (numPositivePredictions - n_pp - n_pm)
    # Condition 5
    if (n_pz > numNonResponsiveGenes) {
        return(FALSE)
    }
    # n_mz (n-0) is the number of genes with negative predictions but are in the set of non-responsive genes in the results.
    n_mz <- (numNegativePredictions - n_mp - n_mm)
    # Condition 6
    if (n_mz > numNonResponsiveGenes) {
        return(FALSE)
    }
    # n_zp (n0p) is the number of genes predicted to have a non-significant interaction with the hypothesis but are actually in the positively affected
    # in the results
    n_zp <- (numUpregulatedGenes - n_pp - n_mp)
    # Condition 7
    if (n_zp > numZeroPredictions) {
        return(FALSE)
    }
    # n_zm (n0m) is the number of genes predicted to have a non-significant interaction with the hypothesis but are actually in the negatively affected
    # in the results
    n_zm <- (numDownregulatedGenes - n_pm - n_mm)
    # Condition 8
    if (n_zm > numZeroPredictions) {
        return(FALSE)
    }
    # Condition 9
    if (((n_pz + n_mz) > numNonResponsiveGenes)) {
        return(FALSE)
    }
    # Condition 10
    if (((n_zp + n_zm) > numZeroPredictions)) {
        return(FALSE)
    }
    
    return(TRUE)
}