NKI_data_import.R

In tulliapadellini/energytree: Conditional Inference trees for mixed type data

# Function to import the data ---------------------------------------------

generate_dataset <- function(data_folder = 'NKI_Rockland/',
                             y_filename = 'NKI_clinical_information.txt',
                             y_column = 'WASI_FULL_4',
                             output_filename = 'NKIdata.RData',
                             output_folder = ".",
                             ext_save = FALSE){

  # Response variable
  y_df <- read.csv(y_filename, colClasses = c('Subject' = 'character'))[c('Subject', y_column)]

  # Exclude obs with no response
  y_df <- na.exclude(y_df)

  # Ids for the patients who have the response
  id_y <- y_df$Subject

  # Function to retrieve the ids for the patients who have a certain 'pattern'
  id_pattern <- function(pattern){
    files <- list.files(data_folder, pattern = pattern)
    sapply(files, FUN = function(file){
      id <- strsplit(file, split = '_', fixed = TRUE)[[1]][1]
    })
  }

  # Ids for the patients who have DTI
  id_dti <- id_pattern('_DTI_connectivity_matrix_file.txt')

  # Ids for the patients who have noGSR fcMRI
  id_fcmri_nogsr <- id_pattern('_fcMRI_noGSR_connectivity_matrix_file.txt')

  # Ids for the patients who have response, DTI and noGSR fcMRI
  id_final <- Reduce(intersect, list(id_y, id_dti, id_fcmri_nogsr))

  # Sort ids
  id_final <- sort(id_final)

  # Sort y_df
  y_df <- y_df[order(y_df$Subject),]

  # Dataset initialization (response only for id_final)
  data_out <- list(y = y_df[[y_column]][y_df$Subject %in% id_final])

  # Retrieve all the files' names for the DTI matrices
  dti_filenames <- sapply(id_final, function(id){
    paste0(data_folder, id, '_DTI_connectivity_matrix_file.txt')
  })

  # Retrieve all the files' names for the noGSR fcMRI matrices
  fcmri_filenames <- sapply(id_final, function(id){
    paste0(data_folder, id, '_fcMRI_noGSR_connectivity_matrix_file.txt')
  })

  # Import the DTI matrices
  data_out$structural <- lapply(dti_filenames, function(d) as.matrix(read.table(d)))

  # Import the noGSR fcMRI matrices
  data_out$functional <- lapply(fcmri_filenames, function(f) as.matrix(read.table(f)))

  # Optional saving
  if(ext_save){
    save(data_out, file = paste0(output_folder, output_filename))
  }

  return(data_out)

}



# Example -------------------------------------------------------------

# nki <- generate_dataset(data_folder = 'NKI_Rockland/',
#                         y_filename = 'NKI_clinical_information.txt',
#                         y_column = 'WASI_FULL_4',
#                         output_filename = 'NKIdata.RData',
#                         output_folder = ".",
#                         ext_save = FALSE)



# Label comparison --------------------------------------------------------

### DTI ###

# # DTI_region_names_full_file comparison
# dti_full_files <- list.files(pattern='_DTI_region_names_full_file.txt', full.names = FALSE)
# comp_full_files <- sapply(dti_full_files, FUN = function(dti) {
#   all.equal(readLines('1013090_DTI_region_names_full_file.txt'), readLines(dti))
# }
# )
# comp_full_files
# #many differences
#
# # DTI_region_names_abbrev_file comparison
# dti_abbrev_files <- list.files(pattern='_DTI_region_names_abbrev_file.txt', full.names = FALSE)
# comp_abbrev_files <- sapply(dti_abbrev_files, FUN = function(dti) {
#   all.equal(readLines('1013090_DTI_region_names_abbrev_file.txt'), readLines(dti))
# }
# )
# all(comp_abbrev_files)
# #all equal
#
# # DTI_region_xyz_centers_file comparison
# dti_xyz_files <- list.files(pattern='_DTI_region_xyz_centers_file.txt', full.names = FALSE)
# comp_xyz_files <- sapply(dti_xyz_files, FUN = function(dti) {
#   all.equal(readLines('1013090_DTI_region_xyz_centers_file.txt'), readLines(dti))
# }
# )
# all(comp_xyz_files)
# #all equal


### fRMI (noGSR) ###

# # fcRMI_region_names_full_file comparison
# fcmri_full_files <- list.files(pattern='_fcMRI_noGSR_region_names_full_file.txt', full.names = FALSE)
# comp_full_files <- sapply(fcmri_full_files, FUN = function(fcmri) {
#   all.equal(readLines('1013090_fcMRI_noGSR_region_names_full_file.txt'), readLines(fcmri))
# }
# )
# comp_full_files
# #many differences
#
# # fcMRI_noGSR_region_names_abbrev_file comparison
# fcmri_abbrev_files <- list.files(pattern='_fcMRI_noGSR_region_names_abbrev_file.txt', full.names = FALSE)
# comp_abbrev_files <- sapply(fcmri_abbrev_files, FUN = function(fcmri) {
#   all.equal(readLines('1013090_fcMRI_noGSR_region_names_abbrev_file.txt'), readLines(fcmri))
# }
# )
# all(comp_abbrev_files)
# #all equal
#
# # fcMRI_noGSR_region_xyz_centers_file comparison
# fcmri_xyz_files <- list.files(pattern='_fcMRI_noGSR_region_xyz_centers_file.txt', full.names = FALSE)
# comp_xyz_files <- sapply(fcmri_xyz_files, FUN = function(fcmri) {
#   all.equal(readLines('1013090_fcMRI_noGSR_region_xyz_centers_file.txt'), readLines(fcmri))
# }
# )
# all(comp_xyz_files)
# #all equal