R/subset_bigsnp.R

In plmmr: Penalized Linear Mixed Models for Correlated Data

#' A helper function to subset `big.matrix` objects
#'
#' @param X                     A filebacked `big.matrix` with the to-be-standardized design matrix
#' @param new_file              Optional user-specified new_file for the to-be-created .rds/.bk files.
#' @param complete_samples      Numeric vector with indicesmarking the rows of the original data which have a non-missing entry in the 6th column of the `.fam` file
#' @param ns                    Numeric vector with the indices of the non-singular columns
#'                              This vector is created in `handle_missingness()`
#' @param rds_dir               The path to the directory in which you want to create the new '.rds' and '.bk' files. Defaults to `data_dir`
#' @param outfile               Optional: the name (character string) of the new_file of the logfile to be written. Defaults to 'process_plink', i.e. you will get 'process_plink.log' as the outfile.
#' @param quiet                 Logical: should messages be printed to the console? Defaults to FALSE (which leaves the print messages on...)
#'
#' @return A list with two components. First, a `big.matrix` object, 'subset_X', representing a design matrix wherein:
#' *  rows are subset according to user's specification in `handle_missing_phen`
#' *  columns are subset so that no constant features remain -- this is important for standardization downstream
#' The list also includes the integer vector 'ns' which marks which columns of the original matrix were 'non-singular' (i.e. *not* constant features).
#' The 'ns' index plays an important role in `plmm_format()` and `untransform()` (both helper functions in model fitting)
#' @keywords internal
#'
subset_filebacked <- function(X, new_file, complete_samples, ns, rds_dir, outfile, quiet){
  # goal here is to subset the features so that constant features (monomorphic SNPs) are not
  # included in analysis
  # NB: this is also where we remove observations with missing phenotypes, if that was requested
  if (!quiet){
    cat("Subsetting data to exclude constant features (e.g., monomorphic SNPs)\n")
  }
  cat("Subsetting data to exclude constant features (e.g., monomorphic SNPs)\n",
      file = outfile, append = TRUE)

  subset_X <- bigmemory::deepcopy(x = X,
                                  row = complete_samples, # filters out samples not represented in both feature data and outcome/predictor data
                                  col = ns, # filters out singular (constant) features, including any constant predictors
                                  type = "double", # note object storage type -- this is key...
                                  backingfile = paste0(new_file, ".bk"),
                                  backingpath = rds_dir,
                                  descriptorfile = paste0(new_file, '.desc'))

  # save ns indices as part of our object
  return(list(subset_X = subset_X, ns = ns))

}