Merge pull request #133 from gdrplatform/GDR-2591

Gdr 2591

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: gDRutils
 Type: Package
 Title: A package with helper functions for processing drug response data
-Version: 1.3.13
-Date: 2024-09-16
+Version: 1.3.14
+Date: 2024-10-03
 Authors@R: c(person("Bartosz", "Czech", role=c("aut"),
                    comment = c(ORCID = "0000-0002-9908-3007")),
              person("Arkadiusz", "Gladki", role=c("cre", "aut"), email="[email protected]",

NEWS.md

@@ -1,3 +1,6 @@
+## gDRutils 1.3.14 - 2024-10-03
+* fixed issue in average_biological_replicated (fit_type)
+
 ## gDRutils 1.3.13 - 2024-09-16
 * add functions set_unique_cl_names_dt and set_unique_drug_names_dt
 

R/headers_list.R

@@ -191,8 +191,20 @@
       "GR50",
       "GEC50",
       "IC50",
-      "EC50"
-    )
+      "EC50",
+      "GR_xc50",
+      "RV_xc50",
+      "GR_ec50",
+      "RV_ec50"
+    ),
+    fit_type = c(
+      "fit_type",
+      "Fit Type",
+      "Fit Type RV",
+      "Fit Type GR",
+      "RV_fit_type",
+      "GR_fit_type"
+   )
   )
 }
 

R/utils.R

@@ -433,6 +433,8 @@ geometric_mean <- function(x, fixed = TRUE, maxlog10Concentration = 1) {
 #' @param fixed Flag indicating whether to add a fix for -Inf in the geometric mean.
 #' @param geometric_average_fields Character vector of column names in \code{dt} 
 #' to take the geometric average of.
+#' @param fit_type_average_fields Character vector of column names in \code{dt} 
+#' that should be treated as a column with fit type data
 #' @param add_sd Flag indicating whether to add standard deviation and count columns.
 #' 
 #' @examples
@@ -449,8 +451,16 @@ average_biological_replicates_dt <- function(
     prettified = FALSE,
     fixed = TRUE,
     geometric_average_fields = get_header("metric_average_fields")$geometric_mean,
+    fit_type_average_fields = get_header("metric_average_fields")$fit_type,
     add_sd = FALSE) {
 
+  checkmate::assert_data_table(dt)
+  checkmate::assert_string(var)
+  checkmate::assert_flag(prettified)
+  checkmate::assert_character(geometric_average_fields)
+  checkmate::assert_character(fit_type_average_fields)
+  checkmate::assert_flag(add_sd)
+
   data <- data.table::copy(dt)
 
   if (prettified) {
@@ -465,7 +475,8 @@ average_biological_replicates_dt <- function(
 
   average_fields <- setdiff(names(Filter(is.numeric, data)), c(unlist(pidfs), var, iso_cols))
   geometric_average_fields <- intersect(geometric_average_fields, names(dt))
-  group_by <- setdiff(names(data), c(average_fields, var, id_cols, "fit_type", "Fit Type"))
+  fit_type_average_fields <- intersect(fit_type_average_fields, names(dt))
+  group_by <- setdiff(names(data), c(average_fields, var, id_cols, fit_type_average_fields))
 
   if (add_sd) {
     # Calculate standard deviation for both average_fields and geometric_average_fields

tests/testthat/test-utils.R

@@ -223,6 +223,36 @@ test_that("average_biological_replicates_dt works as expected", {
   expect_equal(dim(avg_metrics_data2), c(40, 44))
   expect_equal(sum(grepl("_sd", names(avg_metrics_data2))), 15)
   expect_true("count" %in% names(avg_metrics_data2))
+
+  # protection against regression
+  # fit_type correctly recognized in wide and long format
+  sdata <- get_synthetic_data("finalMAE_small")
+  smetrics_data <- convert_se_assay_to_dt(sdata[[1]], "Metrics")
+  tdata <- smetrics_data[1:8, ]
+  tdata$Gnumber <- tdata$Gnumber[1]
+  tdata$DrugName <- tdata$DrugName[1]
+  tdata$source_id <- paste0("DS", rep(1:4, each = 2))
+  tdata$fit_type <- letters[1:8]
+
+  av1b <- average_biological_replicates_dt(tdata, var = "source_id")
+  av1f <- gDRutils::flatten(
+    av1b,
+    groups = c("normalization_type", "fit_source"),
+    wide_cols = gDRutils::get_header("response_metrics")
+  )
+
+  av2f <- gDRutils::flatten(
+    tdata,
+    groups = c("normalization_type", "fit_source"),
+    wide_cols = gDRutils::get_header("response_metrics")
+  )
+  av2b <- average_biological_replicates_dt(av2f, var = "source_id")
+  expect_true(all.equal(av1f, av2b))
+  expect_true(nrow(av1f) == 1)
+  av1i <- average_biological_replicates_dt(tdata, var = "source_id", fit_type_average_fields = "bad_value")
+  expect_true(nrow(av1i) == 8)
+
+
 })
 
 test_that("get_duplicated_rows works as expected", {