Gdr 2516

DESCRIPTION

@@ -2,8 +2,8 @@ Type: Package
 Package: gDRcore
 Title: Processing functions and interface to process and analyze drug
     dose-response data
-Version: 1.1.20
-Date: 2024-05-08
+Version: 1.1.21
+Date: 2024-05-13
 Authors@R: c(
     person("Bartosz", "Czech", , "[email protected]", role = "aut",
            comment = c(ORCID = "0000-0002-9908-3007")),

NAMESPACE

@@ -24,7 +24,9 @@ export(create_and_normalize_SE)
 export(data_model)
 export(fit_SE)
 export(fit_SE.combinations)
+export(get_cellline_annotation_from_dt)
 export(get_default_nested_identifiers)
+export(get_drug_annotation_from_dt)
 export(grr_matches)
 export(identify_data_type)
 export(identify_keys)

NEWS.md

@@ -1,3 +1,6 @@
+## gDRcore 1.1.21 - 2024-05-13
+* add functions for retrieving annotation data from input dt
+
 ## gDRcore 1.1.20 - 2024-05-08
 * fix typo
 

R/add_annotation.R

@@ -53,18 +53,23 @@ add_CellLine_annotation <- function(
     # Assertions:
     checkmate::assert_data_table(dt_metadata)
     checkmate::assert_string(fill, null.ok = TRUE)
-    checkmate::assert_string(externalSource, null.ok = TRUE)
+    checkmate::assert(checkmate::check_string(externalSource, null.ok = TRUE),
+                      checkmate::check_data_table(externalSource,
+                                                  col.names = "named"))
+
 
     cellline <- gDRutils::get_env_identifiers("cellline")
     cellline_name <- gDRutils::get_env_identifiers("cellline_name")
     add_clid <- gDRutils::get_header("add_clid")
 
     if (all(c(cellline, cellline_name, add_clid) %in% names(dt_metadata))) {
-      return(dt_metadata)
+      dt_metadata[, (unlist(add_clid)) := NULL]
     }
 
-    CLs_info <- if (nchar(externalSource) && file.exists(externalSource)) {
+    CLs_info <- if (is.character(externalSource) && nchar(externalSource) && file.exists(externalSource)) {
       data.table::fread(externalSource)
+    } else if (data.table::is.data.table(externalSource)) {
+      externalSource
     } else if (annotationPackage == "gDRtestData") {
       data.table::fread(
         system.file("annotation_data", fname, package = annotationPackage), header = TRUE
@@ -154,8 +159,9 @@ add_Drug_annotation <- function(
   # Assertions:
   checkmate::assert_data_table(dt_metadata)
   checkmate::assert_string(fill, null.ok = TRUE)
-  checkmate::assert_string(externalSource, null.ok = TRUE)
-
+  checkmate::assert(checkmate::check_string(externalSource, null.ok = TRUE),
+                    checkmate::check_data_table(externalSource,
+                                                col.names = "named"))  
   nrows_df <- nrow(dt_metadata)
 
   drug <- unlist(gDRutils::get_env_identifiers(c(
@@ -171,86 +177,88 @@ add_Drug_annotation <- function(
   if (all(c(drug[["drug"]],
             drug_name[["drug_name"]],
             drug_moa[["drug_moa"]]) %in% names(dt_metadata))) {
-    return(dt_metadata)
-  } else {
-    drug_full_identifiers <- c(
-      drug[drug_idx], 
-      drug_name[drug_idx], 
-      drug_moa[drug_idx]
+    dt_metadata[, (unlist(intersect(c(drug_name, drug_moa), names(dt_metadata)))) := NULL]
+  }
+  drug_full_identifiers <- c(
+    drug[drug_idx], 
+    drug_name[drug_idx], 
+    drug_moa[drug_idx]
+  )
+  drug_ids <- stringr::str_extract(names(drug_full_identifiers), "[0-9]") 
+  drug_ids <- ifelse(is.na(drug_ids), 1, drug_ids)
+  drug_identifiers_list <- split(drug_full_identifiers, drug_ids)
+  names(drug_identifiers_list) <- drug[drug_idx]
+
+  # Read local drugs annotations
+
+  Drug_info <- if (is.character(externalSource) && nchar(externalSource) && file.exists(externalSource)) {
+    data.table::fread(externalSource)
+  } else if (data.table::is.data.table(externalSource)) {
+    externalSource
+  } else if (annotationPackage == "gDRtestData") {
+    data.table::fread(
+      system.file("annotation_data", fname, package = annotationPackage), header = TRUE
     )
-    drug_ids <- stringr::str_extract(names(drug_full_identifiers), "[0-9]") 
-    drug_ids <- ifelse(is.na(drug_ids), 1, drug_ids)
-    drug_identifiers_list <- split(drug_full_identifiers, drug_ids)
-    names(drug_identifiers_list) <- drug[drug_idx]
-
-    # Read local drugs annotations
-    Drug_info <- if (nchar(externalSource) && file.exists(externalSource)) {
-      data.table::fread(externalSource)
-    } else if (annotationPackage == "gDRtestData") {
-      data.table::fread(
-        system.file("annotation_data", fname, package = annotationPackage), header = TRUE
-      )
-    } else {
-      eval(parse(text = paste0(annotationPackage,
-                               "::",
-                               "get_drug_annotations")))()
-    }
-
-
-    Drug_info <- Drug_info[, c("gnumber", "drug_name", "drug_moa"), with = FALSE]
-    data.table::setnames(Drug_info, c("gnumber", "drug_name", "drug_moa"),
-                         c("drug", "drug_name", "drug_moa"))
-    drugsTreated <- drugsTreated[!drugsTreated %in% untreated_tag]
-    validatedDrugs <- 
-      remove_drug_batch(drugsTreated) %in% remove_drug_batch(Drug_info[["drug"]])
-    #### function should be parallelizeized
-    missingTblDrugs <- NULL
-    if (!is.null(fill) && any(!validatedDrugs)) {
-      missingTblDrugs <- data.table::data.table(
-        drug = remove_drug_batch(drugsTreated[!validatedDrugs]),
-        drug_name = drugsTreated[!validatedDrugs],
-        drug_moa = fill
-      )
-    }
-
-    if (nrow(Drug_info) == 0) {
-      drug_name <- intersect(drug_name, colnames(dt_metadata))
-      drug <- intersect(drug, colnames(dt_metadata))
-      dt_metadata[, (drug_name) := get(drug)]
-      return(dt_metadata)
-    }
+  } else {
+    eval(parse(text = paste0(annotationPackage,
+                             "::",
+                             "get_drug_annotations")))()
+  }
 
-    Drug_info <- rbind(Drug_info, missingTblDrugs)
 
-    Drug_info$drug <- remove_drug_batch(Drug_info$drug)
-    Drug_info <-
-      rbind(data.table::data.table(
-        drug = untreated_tag,
-        drug_name = untreated_tag,
-        drug_moa = untreated_tag
-      ),
-      Drug_info)
-    Drug_info <- Drug_info[!duplicated(Drug_info[["drug"]]), ]
-    if (any(!remove_drug_batch(drugsTreated) %in% Drug_info$drug) && 
-        !is.null(missingTblDrugs)) {
-      Drug_info <- rbind(Drug_info, stats::setNames(
-        missingTblDrugs[!(remove_drug_batch(missingTblDrugs$drug) %in% 
-                            Drug_info$drug), ],
-        names(Drug_info)
-      ))
-    }
-    for (drug_idf in names(drug_identifiers_list)) {
-      colnames(Drug_info) <- drug_identifiers_list[[drug_idf]]
-      dt_metadata$batch <- dt_metadata[[drug_idf]]
-      dt_metadata[[drug_idf]] <- remove_drug_batch(dt_metadata[[drug_idf]])
-      req_col <- c(drug_idf, setdiff(colnames(dt_metadata), colnames(Drug_info)))
-      dt_metadata <- Drug_info[dt_metadata[, req_col, with = FALSE],  on = drug_idf]
-      dt_metadata[[drug_idf]] <- dt_metadata$batch
-      dt_metadata$batch <- NULL
-    }
-    stopifnot(nrows_df == nrow(dt_metadata))
-    dt_metadata
+  Drug_info <- Drug_info[, c("gnumber", "drug_name", "drug_moa"), with = FALSE]
+  data.table::setnames(Drug_info, c("gnumber", "drug_name", "drug_moa"),
+                       c("drug", "drug_name", "drug_moa"))
+  drugsTreated <- drugsTreated[!drugsTreated %in% untreated_tag]
+  validatedDrugs <- 
+    remove_drug_batch(drugsTreated) %in% remove_drug_batch(Drug_info[["drug"]])
+  #### function should be parallelizeized
+  missingTblDrugs <- NULL
+  if (!is.null(fill) && any(!validatedDrugs)) {
+    missingTblDrugs <- data.table::data.table(
+      drug = remove_drug_batch(drugsTreated[!validatedDrugs]),
+      drug_name = drugsTreated[!validatedDrugs],
+      drug_moa = fill
+    )
   }
+
+  if (nrow(Drug_info) == 0) {
+    drug_name <- intersect(drug_name, colnames(dt_metadata))
+    drug <- intersect(drug, colnames(dt_metadata))
+    dt_metadata[, (drug_name) := get(drug)]
+    return(dt_metadata)
+  }
+
+  Drug_info <- rbind(Drug_info, missingTblDrugs)
+
+  Drug_info$drug <- remove_drug_batch(Drug_info$drug)
+  Drug_info <-
+    rbind(data.table::data.table(
+      drug = untreated_tag,
+      drug_name = untreated_tag,
+      drug_moa = untreated_tag
+    ),
+    Drug_info)
+  Drug_info <- Drug_info[!duplicated(Drug_info[["drug"]]), ]
+  if (any(!remove_drug_batch(drugsTreated) %in% Drug_info$drug) && 
+      !is.null(missingTblDrugs)) {
+    Drug_info <- rbind(Drug_info, stats::setNames(
+      missingTblDrugs[!(remove_drug_batch(missingTblDrugs$drug) %in% 
+                          Drug_info$drug), ],
+      names(Drug_info)
+    ))
+  }
+  for (drug_idf in names(drug_identifiers_list)) {
+    colnames(Drug_info) <- drug_identifiers_list[[drug_idf]]
+    dt_metadata$batch <- dt_metadata[[drug_idf]]
+    dt_metadata[[drug_idf]] <- remove_drug_batch(dt_metadata[[drug_idf]])
+    req_col <- c(drug_idf, setdiff(colnames(dt_metadata), colnames(Drug_info)))
+    dt_metadata <- Drug_info[dt_metadata[, req_col, with = FALSE],  on = drug_idf]
+    dt_metadata[[drug_idf]] <- dt_metadata$batch
+    dt_metadata$batch <- NULL
+  }
+  stopifnot(nrows_df == nrow(dt_metadata))
+  dt_metadata
 }
 
 #' Remove batch from Gnumber
@@ -266,3 +274,74 @@ add_Drug_annotation <- function(
 remove_drug_batch <- function(drug) {
   gsub("\\.[0-9]+.*", "", drug)
 }
+
+
+#' Retrieve the drug annotation from the annotated dt input
+#'
+#' @param dt annotated data.table
+#'
+#' @return data.table with drug annotation
+#' @export
+#'
+#' @examples
+#' dt <- data.table::data.table(Gnumber = "A",
+#' DrugName = "drugA",
+#' drug_moa = "drug_moa_A")
+#' get_drug_annotation_from_dt(dt)
+get_drug_annotation_from_dt <- function(dt) {
+  checkmate::assert_data_table(dt)
+  require(data.table)
+  drug_cols <- intersect(gDRutils::get_env_identifiers()
+                         [grep("drug", names(gDRutils::get_env_identifiers()))],
+                         names(dt))
+  dt_drug <- dt[, unlist(drug_cols), with = FALSE]
+  dt_long <- data.table::melt(dt_drug,
+                              measure.vars = patterns(paste0("^",
+                                                             unlist(drug_cols[c("drug",
+                                                                                             "drug_name",
+                                                                                             "drug_moa")]))),
+                              value.name = unlist(drug_cols[c("drug",
+                                                              "drug_name",
+                                                              "drug_moa")]))
+  dt_long[, "variable" := NULL]
+  data.table::setnames(dt_long,
+                       unlist(drug_cols[c("drug",
+                                          "drug_name",
+                                          "drug_moa")]),
+                       c("gnumber", "drug_name", "drug_moa"))
+  unique_dt <- unique(dt_long)
+  unique_dt[!unique_dt$gnumber %in% gDRutils::get_env_identifiers("untreated_tag"), ]
+}
+
+
+#' Retrieve the cell line annotation from the annotated dt input
+#'
+#' @param dt annotated data.table
+#'
+#' @return data.table with cell line annotation
+#' @export
+#'
+#' @examples
+#' dt <- data.table::data.table(Gnumber = "A",
+#' clid = "CL123",
+#' CellLineName = "cl name",
+#' Tissue = "Bone",
+#' parental_identifier = "some cl",
+#' subtype = "cortical",
+#' ReferenceDivisionTime = 5)
+#' get_cellline_annotation_from_dt(dt)
+get_cellline_annotation_from_dt <- function(dt) {
+  checkmate::assert_data_table(dt)
+  cell_cols <- c(gDRutils::get_env_identifiers("cellline"),
+                 gDRutils::get_header("add_clid"))
+  cell_dt <- dt[, unlist(cell_cols), with = FALSE]
+  data.table::setnames(cell_dt,
+                       unlist(cell_cols),
+                       c("cell_line_identifier",
+                         "cell_line_name",
+                         "primary_tissue",
+                         "parental_identifier",
+                         "subtype",
+                         "doubling_time"))
+  unique(cell_dt)
+}

tests/testthat/test-add_annotation.R

@@ -115,22 +115,28 @@ test_that("remove_drug_batch works", {
   expect_equal(gnumber_without_batch, "DRUG")
 })
 
-test_that("add_CellLine_annotation works with custom annotations", {
-  dt_unknown <- data.table::data.table(ReadoutValue = runif(5),
+
+test_that("get_drug_annotation_from_dt works as expected", {
+  dt_example <- data.table::data.table(Gnumber = "drug_id",
+                                       DrugName = "DrugName",
+                                       drug_moa = "drug_moa",
+                                       some_col = "value")
+  annotation <- get_drug_annotation_from_dt(dt_example)
+  testthat::expect_true(data.table::is.data.table(annotation))
+  testthat::expect_equal(dim(annotation), c(1, 3))
+})
+
+test_that("get_cellline_annotation_from_dt worksas expected", {
+  dt_example <- data.table::data.table(ReadoutValue = runif(5),
                                        clid = paste0("CL", 1:5),
                                        CellLineName = paste0("RandomName", 1:5),
                                        Tissue =  paste0("Tissue", 1:5),
                                        ReferenceDivisionTime = 1:5,
                                        parental_identifier = 1:5,
-                                       subtype = "subtype")
-  dt_unknown_annotated <- add_CellLine_annotation(dt_unknown)
-  expect_identical(dt_unknown, dt_unknown_annotated)
+                                       subtype = "subtype",
+                                       some_col = "value")
+  annotation <- get_cellline_annotation_from_dt(dt_example)
+  testthat::expect_true(data.table::is.data.table(annotation))
+  testthat::expect_equal(dim(annotation), c(5, 6))
 })
 
-test_that("add_Drug_annotation works with custom annotations", {
-  dt_unknown <- data.table::data.table(Gnumber = "drug_id",
-                                       DrugName = "DrugName",
-                                       drug_moa = "drug_moa")
-  dt_unknown_annotated <- add_Drug_annotation(dt_unknown)
-  expect_identical(dt_unknown, dt_unknown_annotated)
-})