R/handle_missing_values.R

In proBatch: Tools for Diagnostics and Corrections of Batch Effects in Proteomics

handle_missing_values <- function(data_matrix, warning_message, 
                                  fill_the_missing = NULL) {
  if (any(is.na(as.vector(data_matrix)))){
    warning(warning_message)
    if(!is.null(fill_the_missing)){
      if (!is.numeric(fill_the_missing)){
        fill_the_missing = 0
      } else{
        warning(sprintf('filling missing value with %s', fill_the_missing))
        data_matrix[is.na(data_matrix)] = fill_the_missing
      }
    } else {
      complete_cases = complete.cases(data_matrix)
      if(ncol(data_matrix) == nrow(data_matrix)){
        pre_corr_dim = nrow(data_matrix)
        if (isSymmetric(data_matrix)){
          warning('removing rows and columns with missing values, as matrix is square')
          if(all(!complete_cases)){
            warning('removing rows with all values missing')
            all_missing_rows = apply(data_matrix, 2, function(x) all(is.na(x)))
            bad_rows= names(all_missing_rows[all_missing_rows])
            good_rows = setdiff(colnames(data_matrix), bad_rows)
            data_matrix = (data_matrix[good_rows, good_rows])
          }
          data_matrix = data_matrix[complete.cases(data_matrix), complete.cases(data_matrix)]
          after_corr_dim = nrow(data_matrix)
          rem_rows = pre_corr_dim - after_corr_dim
          warning(sprintf('removed %s rows of the matrix with missing values', rem_rows))
        } else {
          warning('matrix is square, but not symmetric, coincidence or error?')
          data_matrix = data_matrix[complete.cases(data_matrix),]
        }
      } else {
        warning('filling value is NULL, removing rows with missing values from data matrix')
        data_matrix = data_matrix[complete.cases(data_matrix),]
      }
    }
  }
  return(data_matrix)
}