data-raw/derive-variables.R

In Syksy/curatedPCaData: Curated Prostate Cancer Data

#####
##
## -- DERIVING NEW VARIABLES --
## Set of scripts that derive new variables from the original raw data
##
#####

## IMMUNE DECONVOLUTION VARIABLES

library(immunedeconv) # Wrapper for all immune deconvolution methods
library(MultiAssayExperiment) # Generic handling of MAE-objects
library(RaggedExperiment) # Necessary, otherwise MAE objects cannot be fully opened correctly

## Assistance function for adding a new slot in case it doesn't exist, or to avoid existing slot name overlap
addSlotMAE <- function(mae, # MultiAssayExperiment object
                       ... # Added slots
) {
  # Transform to a list
  vals <- list(...)
  nams <- names(vals)
  # Iterate over ellipsis slot additions
  for (nam in nams) {
    # Subset to samples available on the sample_name of metadata
    vals[[nam]] <- vals[[nam]][, which(colnames(vals[[nam]]) %in% MultiAssayExperiment::colData(mae)[, "sample_name"])]
    # New slot
    if (!nam %in% names(mae)) {
      eval(parse(text = paste("mae <- c(mae,", nam, "=vals[[nam]])", collapse = "")))
      # Replace existing slot
    } else {
      mae[[nam]] <- vals[[nam]]
    }
  }
  # Return the MultiAssayExperiment object
  mae
}

# Notes from the immunedeconv-package:
# https://omnideconv.org/immunedeconv/articles/immunedeconv.html
#
# "Input Data
# The input data is a gene × sample gene expression matrix. In general values should be
#
# - TPM-normalized
# - not log-transformed.
# For xCell and MCP-counter this is not so important. xCell works on the ranks of the gene expression only and MCP-counter sums up the gene expression values."
#
# -> While not TPM-normalized, for most data we take the 2^x if data has since been log-transformed (i.e. most array normalization methods have done this by default)

## AGGREGATE LOAD
## Load temporary MAE-objects so the whole script can be run and output saved over all iterated
## Use latest MAE-objects from the package, coming with GEX and other slots necessary for calculating derived variables here-in
## Update: load the data from working directory '/data/mae_*.rda'
# Processed datasets
# - Abida et al.
load("data/mae_abida.rda")
# - Barbieri et al.
load("data/mae_barbieri.rda")
# - Barwick et al.
load("data/mae_barwick.rda")
# - Chandran et al.
load("data/mae_chandran.rda")
# - Friedrich et al.
load("data/mae_friedrich.rda")
# - ICGC CA
load("data/mae_icgcca.rda")
# - IGC
load("data/mae_igc.rda")
# - Kim et al.
load("data/mae_kim.rda")
# - Kunderfranco et al.
load("data/mae_kunderfranco.rda")
# - Ren et al.
load("data/mae_ren.rda")
# - Sun et al.
load("data/mae_sun.rda")
# - Taylor et al.
load("data/mae_taylor.rda")
# - TCGA
load("data/mae_tcga.rda")
# - True et al.
load("data/mae_true.rda")
# - Wallace et al.
load("data/mae_wallace.rda")
# - Wang et al.
load("data/mae_wang.rda")
# - Weiner et al.
load("data/mae_weiner.rda")

## IMMUNE DECONVOLUTION VARIABLES

#####################################################
#####################################################
##                                                 ##
##                     CIBERSORTX                  ##
##                                                 ##
#####################################################
#####################################################

# Abida et al.
cibersort_polyA_abida <- rio::import("data-raw/CIBERSORTx_abida_Results.csv")
cibersort_polyA_abida <- cibersort_polyA_abida[, -which(names(cibersort_polyA_abida) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_polyA_abida <- t(cibersort_polyA_abida)
colnames(cibersort_polyA_abida) <- cibersort_polyA_abida[1, ]
cibersort_polyA_abida <- cibersort_polyA_abida[-1, ]
cibersort_polyA_abida <- as.matrix(cibersort_polyA_abida)
class(cibersort_polyA_abida) <- "numeric"
mae_abida <- addSlotMAE(mae_abida, cibersort = cibersort_polyA_abida)

# Barbieri et al.
cibersort_barbieri <- rio::import("data-raw/CIBERSORTx_barbieri_Results.csv")
cibersort_barbieri <- cibersort_barbieri[, -which(names(cibersort_barbieri) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_barbieri <- t(cibersort_barbieri)
colnames(cibersort_barbieri) <- cibersort_barbieri[1, ]
cibersort_barbieri <- cibersort_barbieri[-1, ]
cibersort_barbieri <- as.matrix(cibersort_barbieri)
class(cibersort_barbieri) <- "numeric"
mae_barbieri <- addSlotMAE(mae_barbieri, cibersort = cibersort_barbieri)

# Barwick et al.
cibersort_barwick <- rio::import("data-raw/CIBERSORTx_barwick_Results.csv")
cibersort_barwick <- cibersort_barwick[, -which(names(cibersort_barwick) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_barwick <- t(cibersort_barwick)
colnames(cibersort_barwick) <- cibersort_barwick[1, ]
cibersort_barwick <- cibersort_barwick[-1, ]
cibersort_barwick <- as.matrix(cibersort_barwick)
class(cibersort_barwick) <- "numeric"
mae_barwick <- addSlotMAE(mae_barwick, cibersort = cibersort_barwick)

# Chandran et al.
cibersort_chandran <- rio::import("data-raw/CIBERSORTx_chandran_Results.csv")
cibersort_chandran <- cibersort_chandran[, -which(names(cibersort_chandran) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_chandran <- t(cibersort_chandran)
colnames(cibersort_chandran) <- cibersort_chandran[1, ]
cibersort_chandran <- cibersort_chandran[-1, ]
cibersort_chandran <- as.matrix(cibersort_chandran)
class(cibersort_chandran) <- "numeric"
mae_chandran <- addSlotMAE(mae_chandran, cibersort = cibersort_chandran)

# Friedrich et al.
cibersort_friedrich <- rio::import("data-raw/CIBERSORTx_friedrich_Results.csv")
cibersort_friedrich <- cibersort_friedrich[, -which(names(cibersort_friedrich) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_friedrich <- t(cibersort_friedrich)
colnames(cibersort_friedrich) <- cibersort_friedrich[1, ]
cibersort_friedrich <- cibersort_friedrich[-1, ]
cibersort_friedrich <- as.matrix(cibersort_friedrich)
class(cibersort_friedrich) <- "numeric"
mae_friedrich <- addSlotMAE(mae_friedrich, cibersort = cibersort_friedrich)

# ICGC Canadian set
cibersort_icgcca <- rio::import("data-raw/CIBERSORTx_icgcca_Results.csv")
cibersort_icgcca <- cibersort_icgcca[, -which(names(cibersort_icgcca) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_icgcca <- t(cibersort_icgcca)
colnames(cibersort_icgcca) <- cibersort_icgcca[1, ]
cibersort_icgcca <- cibersort_icgcca[-1, ]
cibersort_icgcca <- as.matrix(cibersort_icgcca)
class(cibersort_icgcca) <- "numeric"
mae_icgcca <- addSlotMAE(mae_icgcca, cibersort = cibersort_icgcca)

# IGC
cibersort_igc <- rio::import("data-raw/CIBERSORTx_igc_Results.csv")
cibersort_igc <- cibersort_igc[, -which(names(cibersort_igc) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_igc <- t(cibersort_igc)
colnames(cibersort_igc) <- cibersort_igc[1, ]
cibersort_igc <- cibersort_igc[-1, ]
cibersort_igc <- as.matrix(cibersort_igc)
class(cibersort_igc) <- "numeric"
mae_igc <- addSlotMAE(mae_igc, cibersort = cibersort_igc)

# Kim et al.
cibersort_kim <- rio::import("data-raw/CIBERSORTx_kim_Results.csv")
cibersort_kim <- cibersort_kim[, -which(names(cibersort_kim) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_kim <- t(cibersort_kim)
colnames(cibersort_kim) <- cibersort_kim[1, ]
cibersort_kim <- cibersort_kim[-1, ]
cibersort_kim <- as.matrix(cibersort_kim)
class(cibersort_kim) <- "numeric"
mae_kim <- addSlotMAE(mae_kim, cibersort = cibersort_kim)

# Kunderfranco et al.
cibersort_kunderfranco <- rio::import("data-raw/CIBERSORTx_kunderfranco_Results.csv")
cibersort_kunderfranco <- cibersort_kunderfranco[, -which(names(cibersort_kunderfranco) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_kunderfranco <- t(cibersort_kunderfranco)
colnames(cibersort_kunderfranco) <- cibersort_kunderfranco[1, ]
cibersort_kunderfranco <- cibersort_kunderfranco[-1, ]
cibersort_kunderfranco <- as.matrix(cibersort_kunderfranco)
class(cibersort_kunderfranco) <- "numeric"
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, cibersort = cibersort_kunderfranco)

# Ren et al.
cibersort_ren <- rio::import("data-raw/CIBERSORTx_ren_Results.csv")
cibersort_ren <- cibersort_ren[, -which(names(cibersort_ren) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_ren <- t(cibersort_ren)
colnames(cibersort_ren) <- cibersort_ren[1, ]
cibersort_ren <- cibersort_ren[-1, ]
cibersort_ren <- as.matrix(cibersort_ren)
class(cibersort_ren) <- "numeric"
mae_ren <- addSlotMAE(mae_ren, cibersort = cibersort_ren)

# Sun et al.
cibersort_sun <- rio::import("data-raw/CIBERSORTx_sun_Results.csv")
cibersort_sun <- cibersort_sun[, -which(names(cibersort_sun) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_sun <- t(cibersort_sun)
colnames(cibersort_sun) <- cibersort_sun[1, ]
cibersort_sun <- cibersort_sun[-1, ]
cibersort_sun <- as.matrix(cibersort_sun)
class(cibersort_sun) <- "numeric"
mae_sun <- addSlotMAE(mae_sun, cibersort = cibersort_sun)

# Taylor et al.
cibersort_taylor <- rio::import("data-raw/CIBERSORTx_taylor_Results.csv")
cibersort_taylor <- cibersort_taylor[, -which(names(cibersort_taylor) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_taylor <- t(cibersort_taylor)
colnames(cibersort_taylor) <- cibersort_taylor[1, ]
cibersort_taylor <- cibersort_taylor[-1, ]
cibersort_taylor <- as.matrix(cibersort_taylor)
class(cibersort_taylor) <- "numeric"
mae_taylor <- addSlotMAE(mae_taylor, cibersort = cibersort_taylor)

# TCGA
cibersort_tcga <- rio::import("data-raw/CIBERSORTx_tcga_Results.csv")
cibersort_tcga <- cibersort_tcga[, -which(names(cibersort_tcga) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_tcga <- t(cibersort_tcga)
colnames(cibersort_tcga) <- cibersort_tcga[1, ]
cibersort_tcga <- cibersort_tcga[-1, ]
cibersort_tcga <- as.matrix(cibersort_tcga)
class(cibersort_tcga) <- "numeric"
mae_tcga <- addSlotMAE(mae_tcga, cibersort = cibersort_tcga)

# True et al.
cibersort_true <- rio::import("data-raw/CIBERSORTx_true_Results.csv")
cibersort_true <- cibersort_true[, -which(names(cibersort_true) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_true <- t(cibersort_true)
colnames(cibersort_true) <- cibersort_true[1, ]
cibersort_true <- cibersort_true[-1, ]
cibersort_true <- as.matrix(cibersort_true)
class(cibersort_true) <- "numeric"
mae_true <- addSlotMAE(mae_true, cibersort = cibersort_true)

# Wallace et al.
cibersort_wallace <- rio::import("data-raw/CIBERSORTx_wallace_Results.csv")
cibersort_wallace <- cibersort_wallace[, -which(names(cibersort_wallace) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_wallace <- t(cibersort_wallace)
colnames(cibersort_wallace) <- cibersort_wallace[1, ]
cibersort_wallace <- cibersort_wallace[-1, ]
cibersort_wallace <- as.matrix(cibersort_wallace)
class(cibersort_wallace) <- "numeric"
mae_wallace <- addSlotMAE(mae_wallace, cibersort = cibersort_wallace)

# Wang et al.
cibersort_wang <- rio::import("data-raw/CIBERSORTx_wang_Results.csv")
cibersort_wang <- cibersort_wang[, -which(names(cibersort_wang) %in% c("P-value", "Correlation", "RMSE"))]
cibersort_wang <- t(cibersort_wang)
colnames(cibersort_wang) <- cibersort_wang[1, ]
cibersort_wang <- cibersort_wang[-1, ]
cibersort_wang <- as.matrix(cibersort_wang)
class(cibersort_wang) <- "numeric"
mae_wang <- addSlotMAE(mae_wang, cibersort = cibersort_wang)

# Weiner et al.
cibersort_weiner <- rio::import("data-raw/CIBERSORTx_weiner_Results.csv")
cibersort_weiner <- cibersort_weiner[, -which(names(cibersort_weiner) %in% c("P-value", "Pearson Correlation", "RMSE", "Absolute score"))]
cibersort_weiner <- t(cibersort_weiner)
colnames(cibersort_weiner) <- cibersort_weiner[1, ]
cibersort_weiner <- cibersort_weiner[-1, ]
cibersort_weiner <- as.matrix(cibersort_weiner)
class(cibersort_weiner) <- "numeric"
mae_weiner <- addSlotMAE(mae_weiner, cibersort = cibersort_weiner)

#####################################################
#####################################################
##                                                 ##
##                     XCELL                       ##
##                                                 ##
#####################################################
#####################################################

# Abida et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_abida[["gex.relz"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_abida <- addSlotMAE(mae_abida, xcell = tmp)

# Barbieri et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_barbieri[["gex.relz"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barbieri <- addSlotMAE(mae_barbieri, xcell = tmp)

## METHOD FAILS DUE TO LACK OF GENE OVERLAP
if (FALSE) {
  # Barwick et al.
  tmp <- as.data.frame(immunedeconv::deconvolute(mae_barwick[["gex.logq"]], method = "xcell"))
  rownames(tmp) <- tmp$cell_type
  # Omit cell type column and store only data of cell type populations
  tmp <- as.matrix(tmp[, -1])
  mae_barwick <- addSlotMAE(mae_barwick, xcell = tmp)
}

# Chandran et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_chandran[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_chandran <- addSlotMAE(mae_chandran, xcell = tmp)

# Friedrich et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_friedrich[["gex.logq"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_friedrich <- addSlotMAE(mae_friedrich, xcell = tmp)

# ICGCCA
tmp <- as.data.frame(immunedeconv::deconvolute(mae_icgcca[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_icgcca <- addSlotMAE(mae_icgcca, xcell = tmp)

# Kunderfranco et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kunderfranco[["gex.logr"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, xcell = tmp)

# Ren et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_ren[["gex.relz"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_ren <- addSlotMAE(mae_ren, xcell = tmp)

# Sun et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_sun[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_sun <- addSlotMAE(mae_sun, xcell = tmp)

# Taylor et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_taylor[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_taylor <- addSlotMAE(mae_taylor, xcell = tmp)

# TCGA
tmp <- as.data.frame(immunedeconv::deconvolute(mae_tcga[["gex.rsem.log"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_tcga <- addSlotMAE(mae_tcga, xcell = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_wang[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
tmp <- as.matrix(tmp[, -1])
mae_wang <- addSlotMAE(mae_wang, xcell = tmp)

# Kim et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kim[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_kim <- addSlotMAE(mae_kim, xcell = tmp)

# Wallace et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_wallace[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wallace <- addSlotMAE(mae_wallace, xcell = tmp)

# Igc
tmp <- as.data.frame(immunedeconv::deconvolute(mae_igc[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_igc <- addSlotMAE(mae_igc, xcell = tmp)

# Weiner et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_weiner[["gex.rma"]], method = "xcell"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_weiner <- addSlotMAE(mae_weiner, xcell = tmp)

## METHOD FAILS DUE TO LACK OF GENE OVERLAP
if (FALSE) {
  # True et al.
  tmp <- as.data.frame(immunedeconv::deconvolute(mae_true[["gex.logr"]], method = "xcell"))
  rownames(tmp) <- tmp$cell_type
  # Omit cell type column and store only data of cell type populations
  tmp <- as.matrix(tmp[, -1])
  mae_true <- addSlotMAE(mae_true, xcell = tmp)
  # Save the derived new 'assay' types to the mae-object
}


#####################################################
#####################################################
##                                                 ##
##                     EPIC                        ##
##                                                 ##
#####################################################
#####################################################

# Abida et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_abida[["gex.relz"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_abida <- addSlotMAE(mae_abida, epic = tmp)

# Barbieri et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_barbieri[["gex.relz"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barbieri <- addSlotMAE(mae_barbieri, epic = tmp)

# FAILS DUE TO OVERLAP OF GENES
if (FALSE) {
  # Barwick et al.
  tmp <- as.data.frame(immunedeconv::deconvolute(mae_barwick[["gex.logq"]], method = "epic"))
  rownames(tmp) <- tmp$cell_type
  # Omit cell type column and store only data of cell type populations
  tmp <- as.matrix(tmp[, -1])
  mae_barwick <- addSlotMAE(mae_barwick, epic = tmp)
  # Save the derived new 'assay' types to the mae-object
  # usethis::use_data(mae_barwick, overwrite = TRUE)
}

# Chandran et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_chandran[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_chandran <- addSlotMAE(mae_chandran, epic = tmp)

# Friedrich et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_friedrich[["gex.logq"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_friedrich <- addSlotMAE(mae_friedrich, epic = tmp)

# ICGCCA
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_icgcca[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_icgcca <- addSlotMAE(mae_icgcca, epic = tmp)

# Kunderfranco et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kunderfranco[["gex.logr"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, epic = tmp)

# Ren et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_ren[["gex.relz"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_ren <- addSlotMAE(mae_ren, epic = tmp)

# Sun et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_sun[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_sun <- addSlotMAE(mae_sun, epic = tmp)

# Taylor et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_taylor[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_taylor <- addSlotMAE(mae_taylor, epic = tmp)

# TCGA
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_tcga[["gex.rsem.log"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_tcga <- addSlotMAE(mae_tcga, epic = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wang[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_wang <- addSlotMAE(mae_wang, epic = tmp)

# Kim et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_kim[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_kim <- addSlotMAE(mae_kim, epic = tmp)

# Wallace et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wallace[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wallace <- addSlotMAE(mae_wallace, epic = tmp)

# IGC
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_igc[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_igc <- addSlotMAE(mae_igc, epic = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wang[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wang <- addSlotMAE(mae_wang, epic = tmp)

# Weiner et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_weiner[["gex.rma"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_weiner <- addSlotMAE(mae_weiner, epic = tmp)

# True et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_true[["gex.logr"]], method = "epic"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_true <- addSlotMAE(mae_true, epic = tmp)


#####################################################
#####################################################
##                                                 ##
##                     QUANTISEQ                   ##
##                                                 ##
#####################################################
#####################################################

# Abida et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_abida[["gex.relz"]], method = "quantiseq"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_abida <- addSlotMAE(mae_abida, quantiseq = tmp)

# Barbieri et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_barbieri[["gex.relz"]], method = "quantiseq"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barbieri <- addSlotMAE(mae_barbieri, quantiseq = tmp)

# Barwick et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_barwick[["gex.logq"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barwick <- addSlotMAE(mae_barwick, quantiseq = tmp)

# Chandran et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_chandran[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_chandran <- addSlotMAE(mae_chandran, quantiseq = tmp)

# Friedrich et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_friedrich[["gex.logq"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_friedrich <- addSlotMAE(mae_friedrich, quantiseq = tmp)

# ICGCCA
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_icgcca[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_icgcca <- addSlotMAE(mae_icgcca, quantiseq = tmp)

# Kunderfranco et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kunderfranco[["gex.logr"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, quantiseq = tmp)

# Ren et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_ren[["gex.relz"]], method = "quantiseq"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_ren <- addSlotMAE(mae_ren, quantiseq = tmp)

# Sun et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_sun[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_sun <- addSlotMAE(mae_sun, quantiseq = tmp)

# Taylor et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_taylor[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_taylor <- addSlotMAE(mae_taylor, quantiseq = tmp)

# TCGA
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_tcga[["gex.rsem.log"]], method = "quantiseq"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_tcga <- addSlotMAE(mae_tcga, quantiseq = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wang[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_wang <- addSlotMAE(mae_wang, quantiseq = tmp)

# Kim et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_kim[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_kim <- addSlotMAE(mae_kim, quantiseq = tmp)

# Wallace et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wallace[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wallace <- addSlotMAE(mae_wallace, quantiseq = tmp)

# IGC
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_igc[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_igc <- addSlotMAE(mae_igc, quantiseq = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_wang[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wang <- addSlotMAE(mae_wang, quantiseq = tmp)

# Weiner et al.
tmp <- as.data.frame(immunedeconv::deconvolute(2^mae_weiner[["gex.rma"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_weiner <- addSlotMAE(mae_weiner, quantiseq = tmp)

# True et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_true[["gex.logr"]], method = "quantiseq", arrays = TRUE))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_true <- addSlotMAE(mae_true, quantiseq = tmp)



#####################################################
#####################################################
##                                                 ##
##                     MCP COUNTER                 ##
##                                                 ##
#####################################################
#####################################################

# Abida et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_abida[["gex.relz"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_abida <- addSlotMAE(mae_abida, mcp = tmp)

# Barbieri et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_barbieri[["gex.relz"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barbieri <- addSlotMAE(mae_barbieri, mcp = tmp)

# Barwick et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_barwick[["gex.logq"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_barwick <- addSlotMAE(mae_barwick, mcp = tmp)

# Chandran et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_chandran[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_chandran <- addSlotMAE(mae_chandran, mcp = tmp)

# Friedrich et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_friedrich[["gex.logq"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_friedrich <- addSlotMAE(mae_friedrich, mcp = tmp)

# ICGCCA
tmp <- as.data.frame(immunedeconv::deconvolute(mae_icgcca[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_icgcca <- addSlotMAE(mae_icgcca, mcp = tmp)

# IGC
tmp <- as.data.frame(immunedeconv::deconvolute(mae_igc[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_igc <- addSlotMAE(mae_igc, mcp = tmp)

# Kim et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kim[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_kim <- addSlotMAE(mae_kim, mcp = tmp)

# Kunderfranco et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_kunderfranco[["gex.logr"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, mcp = tmp)

# Ren et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_ren[["gex.relz"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_ren <- addSlotMAE(mae_ren, mcp = tmp)

# Sun et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_sun[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_sun <- addSlotMAE(mae_sun, mcp = tmp)

# Taylor et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_taylor[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_taylor <- addSlotMAE(mae_taylor, mcp = tmp)

# TCGA
tmp <- as.data.frame(immunedeconv::deconvolute(mae_tcga[["gex.rsem.log"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_tcga <- addSlotMAE(mae_tcga, mcp = tmp)

# True et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_true[["gex.logr"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_true <- addSlotMAE(mae_true, mcp = tmp)

# Wang et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_wang[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
# Concatenate the new results to the MAE-object
mae_wang <- addSlotMAE(mae_wang, mcp = tmp)

# Wallace et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_wallace[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_wallace <- addSlotMAE(mae_wallace, mcp = tmp)

# Weiner et al.
tmp <- as.data.frame(immunedeconv::deconvolute(mae_weiner[["gex.rma"]], method = "mcp_counter"))
rownames(tmp) <- tmp$cell_type
# Omit cell type column and store only data of cell type populations
tmp <- as.matrix(tmp[, -1])
mae_weiner <- addSlotMAE(mae_weiner, mcp = tmp)

#####
##
## ESTIMATE
## (Available from immunedeconv >= 2.1.0, built from GitHub
##
#####

# Abida et al.
tmp <- immunedeconv::deconvolute_estimate(mae_abida[["gex.relz"]])
mae_abida <- addSlotMAE(mae_abida, estimate = as.matrix(tmp))

# Barbieri et al.
tmp <- immunedeconv::deconvolute_estimate(mae_barbieri[["gex.relz"]])
mae_barbieri <- addSlotMAE(mae_barbieri, estimate = as.matrix(tmp))

# Barwick et al.
tmp <- immunedeconv::deconvolute_estimate(mae_barwick[["gex.logq"]])
mae_barwick <- addSlotMAE(mae_barwick, estimate = as.matrix(tmp))

# Chandran et al.
tmp <- immunedeconv::deconvolute_estimate(mae_chandran[["gex.rma"]])
mae_chandran <- addSlotMAE(mae_chandran, estimate = as.matrix(tmp))

# Friedrich et al.
tmp <- immunedeconv::deconvolute_estimate(mae_friedrich[["gex.logq"]])
mae_friedrich <- addSlotMAE(mae_friedrich, estimate = as.matrix(tmp))

# ICGCCA
tmp <- immunedeconv::deconvolute_estimate(mae_icgcca[["gex.rma"]])
mae_icgcca <- addSlotMAE(mae_icgcca, estimate = as.matrix(tmp))

# IGC
tmp <- immunedeconv::deconvolute_estimate(mae_igc[["gex.rma"]])
mae_igc <- addSlotMAE(mae_igc, estimate = as.matrix(tmp))

# Kim et al.
tmp <- immunedeconv::deconvolute_estimate(mae_kim[["gex.rma"]])
mae_kim <- addSlotMAE(mae_kim, estimate = as.matrix(tmp))

# Kunderfranco et al.
tmp <- immunedeconv::deconvolute_estimate(mae_kunderfranco[["gex.logr"]])
mae_kunderfranco <- addSlotMAE(mae_kunderfranco, estimate = as.matrix(tmp))

# Ren et al.
tmp <- immunedeconv::deconvolute_estimate(mae_ren[["gex.relz"]])
mae_ren <- addSlotMAE(mae_ren, estimate = as.matrix(tmp))

# Sun et al.
tmp <- immunedeconv::deconvolute_estimate(mae_sun[["gex.rma"]])
mae_sun <- addSlotMAE(mae_sun, estimate = as.matrix(tmp))

# Taylor et al.
tmp <- immunedeconv::deconvolute_estimate(mae_taylor[["gex.rma"]])
mae_taylor <- addSlotMAE(mae_taylor, estimate = as.matrix(tmp))

# TCGA
tmp <- immunedeconv::deconvolute_estimate(mae_tcga[["gex.rsem.log"]])
mae_tcga <- addSlotMAE(mae_tcga, estimate = as.matrix(tmp))

# True et al.
tmp <- immunedeconv::deconvolute_estimate(mae_true[["gex.logr"]])
mae_true <- addSlotMAE(mae_true, estimate = as.matrix(tmp))

# Wang et al.
tmp <- immunedeconv::deconvolute_estimate(mae_wang[["gex.rma"]])
mae_wang <- addSlotMAE(mae_wang, estimate = as.matrix(tmp))

# Wallace et al.
tmp <- immunedeconv::deconvolute_estimate(mae_wallace[["gex.rma"]])
mae_wallace <- addSlotMAE(mae_wallace, estimate = as.matrix(tmp))

# Weiner et al.
tmp <- immunedeconv::deconvolute_estimate(mae_weiner[["gex.rma"]])
mae_weiner <- addSlotMAE(mae_weiner, estimate = as.matrix(tmp))


#####
##
## Genomic risk scores: Prolaris, OncotypeDX & Decipher
## AR scores as used by TCGA, originally presented in Hieronymus et al. 2006
##
#####


## Abida
mae_abida <- addSlotMAE(mae_abida,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_abida, slot = "gex.relz", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_abida, slot = "gex.relz", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_abida, slot = "gex.relz", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_abida, slot = "gex.relz", test = "AR")
  )
)

## Barbieri
mae_barbieri <- addSlotMAE(mae_barbieri,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_barbieri, slot = "gex.relz", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_barbieri, slot = "gex.relz", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_barbieri, slot = "gex.relz", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_barbieri, slot = "gex.relz", test = "AR")
  )
)

## Barwick
mae_barwick <- addSlotMAE(mae_barwick,
  scores = rbind(
    # Lack of gene overlap
    # decipher = curatedPCaData:::genomic_risk(mae_barwick, slot = "gex.logq", test = "decipher", log=FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_barwick, slot = "gex.logq", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_barwick, slot = "gex.logq", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_barwick, slot = "gex.logq", test = "AR")
  )
)

## Chandran
mae_chandran <- addSlotMAE(mae_chandran,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_chandran, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_chandran, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_chandran, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_chandran, slot = "gex.rma", test = "AR")
  )
)

## Friedrich
mae_friedrich <- addSlotMAE(mae_friedrich,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_friedrich, slot = "gex.logq", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_friedrich, slot = "gex.logq", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_friedrich, slot = "gex.logq", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_friedrich, slot = "gex.logq", test = "AR")
  )
)

## ICGC-CA
mae_icgcca <- addSlotMAE(mae_icgcca,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_icgcca, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_icgcca, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_icgcca, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_icgcca, slot = "gex.rma", test = "AR")
  )
)

## IGC
mae_igc <- addSlotMAE(mae_igc,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_igc, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_igc, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_igc, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_igc, slot = "gex.rma", test = "AR")
  )
)

## Kim
mae_kim <- addSlotMAE(mae_kim,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_kim, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_kim, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_kim, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_kim, slot = "gex.rma", test = "AR")
  )
)

## Kunderfranco
mae_kunderfranco <- addSlotMAE(mae_kunderfranco,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_kunderfranco, slot = "gex.logr", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_kunderfranco, slot = "gex.logr", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_kunderfranco, slot = "gex.logr", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_kunderfranco, slot = "gex.logr", test = "AR")
  )
)

## Ren
mae_ren <- addSlotMAE(mae_ren,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_ren, slot = "gex.relz", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_ren, slot = "gex.relz", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_ren, slot = "gex.relz", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_ren, slot = "gex.relz", test = "AR")
  )
)

## Sun
mae_sun <- addSlotMAE(mae_sun,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_sun, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_sun, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_sun, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_sun, slot = "gex.rma", test = "AR")
  )
)

## Taylor
mae_taylor <- addSlotMAE(mae_taylor,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_taylor, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_taylor, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_taylor, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_taylor, slot = "gex.rma", test = "AR")
  )
)

## TCGA
mae_tcga <- addSlotMAE(mae_tcga,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_tcga, slot = "gex.rsem.log", test = "decipher", log = TRUE),
    oncotype = curatedPCaData:::genomic_risk(mae_tcga, slot = "gex.rsem.log", test = "oncotype", log = TRUE),
    prolaris = curatedPCaData:::genomic_risk(mae_tcga, slot = "gex.rsem.log", test = "prolaris", log = TRUE),
    ar_score = curatedPCaData:::genomic_score(mae_tcga, slot = "gex.rsem.log", test = "AR")
  )
)

## True
mae_true <- addSlotMAE(mae_true,
  scores = rbind(
    # Lack of gene overlap
    # decipher = curatedPCaData:::genomic_risk(mae_true, slot = "gex.logr", test = "decipher", log=FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_true, slot = "gex.logr", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_true, slot = "gex.logr", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_true, slot = "gex.logr", test = "AR")
  )
)

## Wallace
mae_wallace <- addSlotMAE(mae_wallace,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_wallace, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_wallace, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_wallace, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_wallace, slot = "gex.rma", test = "AR")
  )
)

## Wang
mae_wang <- addSlotMAE(mae_wang,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_wang, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_wang, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_wang, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_wang, slot = "gex.rma", test = "AR")
  )
)

## Weiner
mae_weiner <- addSlotMAE(mae_weiner,
  scores = rbind(
    decipher = curatedPCaData:::genomic_risk(mae_weiner, slot = "gex.rma", test = "decipher", log = FALSE),
    oncotype = curatedPCaData:::genomic_risk(mae_weiner, slot = "gex.rma", test = "oncotype", log = FALSE),
    prolaris = curatedPCaData:::genomic_risk(mae_weiner, slot = "gex.rma", test = "prolaris", log = FALSE),
    ar_score = curatedPCaData:::genomic_score(mae_weiner, slot = "gex.rma", test = "AR")
  )
)



#####
##
## Purity estimate scores
## Potentially methods such as: DeMixT (tumor vs. ctrl), ISOpureR (tumor vs. ctrl), ESTIMATE (tumor-only)
##
#####

##
## For start, DeMixT for tumor-control design
##

# Remove DeMixT from current release
if (FALSE) {
  ## Chandran et al.
  chandran_demixt <- t(read.csv("data-raw/DeMixT_Chandran.txt", row.names = 1))
  rownames(chandran_demixt) <- "demixt"
  mae_chandran <- addSlotMAE(mae_chandran, purity = chandran_demixt)

  ## Friedrich et al.
  friedrich_demixt <- t(read.csv("data-raw/DeMixT_Friedrich.txt", row.names = 1))
  rownames(friedrich_demixt) <- "demixt"
  mae_friedrich <- addSlotMAE(mae_friedrich, purity = friedrich_demixt)

  ## Kunderfranco et al.
  kunderfranco_demixt <- t(read.csv("data-raw/DeMixT_Kunderfranco.txt", row.names = 1))
  rownames(kunderfranco_demixt) <- "demixt"
  mae_kunderfranco <- addSlotMAE(mae_kunderfranco, purity = kunderfranco_demixt)

  ## Taylor et al.
  taylor_demixt <- t(read.csv("data-raw/DeMixT_Taylor.txt", row.names = 1))
  rownames(taylor_demixt) <- "demixt"
  mae_taylor <- addSlotMAE(mae_taylor, purity = taylor_demixt)

  ## TCGA
  tcga_demixt <- t(read.csv("data-raw/DeMixT_TCGA.txt", row.names = 1))
  rownames(tcga_demixt) <- "demixt"
  mae_tcga <- addSlotMAE(mae_tcga, purity = tcga_demixt)

  ## Wallace et al.
  wallace_demixt <- t(read.csv("data-raw/DeMixT_Wallace.txt", row.names = 1))
  rownames(wallace_demixt) <- "demixt"
  mae_wallace <- addSlotMAE(mae_wallace, purity = wallace_demixt)
}

## ----

## AGGREGATE SAVE
## One by one save back all the MAE objects back so all updated slots are correctly stored in the new MAE objects
##

# Processed datasets
# - Abida et al.
usethis::use_data(mae_abida, overwrite = TRUE)
# - Barbieri et al.
usethis::use_data(mae_barbieri, overwrite = TRUE)
# - Barwick et al.
usethis::use_data(mae_barwick, overwrite = TRUE)
# - Chandran et al.
usethis::use_data(mae_chandran, overwrite = TRUE)
# - Friedrich et al.
usethis::use_data(mae_friedrich, overwrite = TRUE)
# - ICGC CA
usethis::use_data(mae_icgcca, overwrite = TRUE)
# - IGC
usethis::use_data(mae_igc, overwrite = TRUE)
# - Kim et al.
usethis::use_data(mae_kim, overwrite = TRUE)
# - Kunderfranco et al.
usethis::use_data(mae_kunderfranco, overwrite = TRUE)
# - Ren et al.
usethis::use_data(mae_ren, overwrite = TRUE)
# - Sun et al.
usethis::use_data(mae_sun, overwrite = TRUE)
# - Taylor et al.
usethis::use_data(mae_taylor, overwrite = TRUE)
# - TCGA
usethis::use_data(mae_tcga, overwrite = TRUE)
# - True et al.
usethis::use_data(mae_true, overwrite = TRUE)
# - Wallace et al.
usethis::use_data(mae_wallace, overwrite = TRUE)
# - Wang et al.
usethis::use_data(mae_wang, overwrite = TRUE)
# - Weiner et al.
usethis::use_data(mae_weiner, overwrite = TRUE)

# Re-save the processed datasets using the compression indicated at LazyDataCompression-field in DESCRIPTION
tools::resaveRdaFiles("data/mae_abida.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_barbieri.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_barwick.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_chandran.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_friedrich.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_icgcca.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_igc.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_kim.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_kunderfranco.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_ren.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_sun.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_taylor.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_tcga.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_true.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_wallace.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_wang.rda", compress = "xz")
tools::resaveRdaFiles("data/mae_weiner.rda", compress = "xz")