Bug multiple analyte metabolite

R/format_data.R

@@ -104,7 +104,8 @@ format_pkncadose_data <- function(pkncaconc_data,
 #'
 #' This function creates a dataset with dose intervals and specified pharmacokinetic parameters.
 #'
-#' @param df_dose A PKNCAdose object
+#' @param pknca_conc A PKNCAdose object containing the concentration data.
+#' @param pknca_dose A PKNCAdose object containing the dose data.
 #' @param params A character vector specifying the pharmacokinetic parameters to include.
 #' @param start_from_last_dose Logical defining if start is at time of last dose or C1.
 #'
@@ -119,21 +120,26 @@ format_pkncadose_data <- function(pkncaconc_data,
 #' @examples
 #' \dontrun{
 #'   # Example usage:
-#'   dose_intervals <- format_pkncadata_intervals(df_dose, params)
+#'   dose_intervals <- format_pkncadata_intervals(pknca_conc, pknca_dose, params)
 #' }
 #'
 #' @import dplyr
 #' @export
-format_pkncadata_intervals <- function(pknca_dose,
+format_pkncadata_intervals <- function(pknca_conc,
+                                       pknca_dose,
                                        params =  c("aucinf.obs", "aucint.last", "auclast",
                                                    "cmax", "half.life", "tmax", "lambda.z",
                                                    "lambda.z.n.points", "r.squared",
                                                    "adj.r.squared", "lambda.z.time.first",
                                                    "aucpext.obs", "aucpext.pred", "clast.obs",
                                                    "cl.obs"),
                                        start_from_last_dose = TRUE) {
+  if (!inherits(pknca_conc, "PKNCAconc")) {
+    stop("Input pknca_conc must be a PKNCAconc object from the PKNCA package.")
+  }
+
   if (!inherits(pknca_dose, "PKNCAdose")) {
-    stop("Input must be a PKNCAdose object from the PKNCA package.")
+    stop("Input pknca_dose must be a PKNCAdose object from the PKNCA package.")
   }
 
   required_columns <- c(unname(unlist(pknca_dose$columns$groups)), pknca_dose$columns$time)
@@ -142,15 +148,44 @@ format_pkncadata_intervals <- function(pknca_dose,
     stop(paste("Missing required columns:", paste(missing_columns, collapse = ", ")))
   }
 
+  # Select relevant group columns
+  conc_groups <- unname(unlist(pknca_conc$columns$groups))
+  dose_groups <- unname(unlist(pknca_dose$columns$groups))
+  time_dose_seg <- if ("DOSNO" %in% names(pknca_dose$data)) "DOSNO" else "time_dose"
+
+  # Select conc data and for time column give priority to non-predose samples
+  sub_pknca_conc <- pknca_conc$data %>%
+    select(any_of(c(conc_groups, "AFRLT", "ARRLT", "DOSNO"))) %>%
+    arrange(!!!syms(conc_groups), ARRLT < 0, AFRLT) %>%
+    ungroup()
+
+  # Select dose data and use its time column as a time of last dose reference
+  sub_pknca_dose <- pknca_dose$data %>%
+    select(any_of(c(unname(unlist(pknca_dose$columns$groups)),
+                    pknca_dose$columns$time, "DOSNO"))) %>%
+    rename_with(~ "time_dose", pknca_dose$columns$time)
+
   # Based on dose times create a data frame with start and end times
-  dose_intervals <- pknca_dose$data %>%
-    mutate(start = if (start_from_last_dose) !!sym(pknca_dose$columns$time)
-           else !!sym(pknca_dose$columns$time) + !!sym("ARRLT")) %>%
-    group_by(!!!syms(unname(unlist(pknca_dose$columns$groups)))) %>%
-    arrange(!!sym(pknca_dose$columns$time)) %>%
-    mutate(end = lead(as.numeric(!!sym(pknca_dose$columns$time)), default = Inf)) %>%
+  dose_intervals <- left_join(sub_pknca_conc,
+                              sub_pknca_dose,
+                              relationship = "many-to-many") %>%
+
+    # Pick 1 per concentration group and dose number
+    arrange(!!!syms(conc_groups), ARRLT < 0, AFRLT) %>%
+    group_by(!!!syms(c(conc_groups, time_dose_seg))) %>%
+    slice(1) %>%
+    ungroup() %>%
+
+    # Make start from last dose (pknca_dose) or first concentration (pknca_conc)
+    mutate(start = if (start_from_last_dose) time_dose
+           else time_dose + !!sym("ARRLT")) %>%
+    group_by(!!!syms(c(conc_groups))) %>%
+    arrange(time_dose) %>%
+
+    # Make end based on next dose time (if no more, Inf)
+    mutate(end = lead(as.numeric(time_dose), default = Inf)) %>%
     ungroup() %>%
-    select(any_of(c("start", "end", unname(unlist(pknca_dose$columns$groups)), "DOSNO"))) %>%
+    select(any_of(c("start", "end", unname(unlist(pknca_conc$columns$groups)), "DOSNO"))) %>%
 
     # Create logical columns with the TRUE and as names params argument
     mutate(!!!setNames(rep(TRUE, length(params)), params)) %>%

inst/shiny/tabs/nca.R

@@ -233,7 +233,7 @@ observeEvent(input$submit_analyte, priority = 2, {
     duration = "ADOSEDUR"
   )
 
-  myintervals <- format_pkncadata_intervals(mydose) %>%
+  myintervals <- format_pkncadata_intervals(pknca_conc = myconc, pknca_dose = mydose) %>%
     # Filter only the doses requested by the user
     dplyr::filter(
       !!sym(dosno_column) %in% input$select_dosno
@@ -484,6 +484,7 @@ observeEvent(input$nca, {
 res_nca <- reactiveVal(NULL)
 observeEvent(pk_nca_trigger(), {
   req(mydata())
+
   withProgress(message = "Calculating NCA...", value = 0, {
     myres <- PKNCA::pk.nca(data = mydata(), verbose = FALSE)
 
@@ -492,7 +493,8 @@ observeEvent(pk_nca_trigger(), {
 
     # Make the starts and ends of results relative to last dose using the dose data
     myres$result <- myres$result %>%
-      inner_join(select(mydata()$dose$data, -exclude)) %>%
+      left_join(select(mydata()$dose$data, any_of(c(unname(unlist(mydata()$dose$columns$groups)),
+                                                    mydata()$dose$columns$time)))) %>%
       mutate(start = start - !!sym(mydata()$dose$columns$time),
              end = end - !!sym(mydata()$dose$columns$time)) %>%
       select(names(myres$result))
@@ -520,6 +522,7 @@ final_res_nca <- reactiveVal(NULL)
 # creative final_res_nca, aiming to present the results in a more comprehensive way
 observeEvent(res_nca(), {
   req(res_nca())
+
   # Create a reshaped object that will be used to display the results in the UI
   final_res_nca <- pivot_wider_pknca_results(res_nca())
 

tests/testthat/test-format_data.R

@@ -198,11 +198,39 @@ test_that("format_pkncadose_data handles empty input", {
 })
 
 test_that("format_pkncadata_intervals handles incorrect input type", {
-  mydose <- list(data = data.frame(STUDYID = 1, USUBJID = 1, TIME = 0),
-                 columns = list(groups = c("STUDYID", "USUBJID"),
-                                time = "TIME"))
-  expect_error(format_pkncadata_intervals(mydose),
-               regexp = "Input must be a PKNCAdose object from the PKNCA package.")
+  ADNCA <- data.frame(
+    STUDYID = rep(1, 20),
+    USUBJID = rep(1:2, each = 10),
+    PCSPEC = rep("Plasma", 20),
+    DRUG = rep("DrugA", 20),
+    ANALYTE = rep("Analyte1", 20),
+    AFRLT = rep(seq(0, 9), 2),
+    ARRLT = rep(seq(0, 4), 4),
+    NFRLT = rep(seq(0, 9), 2),
+    ROUTE = "extravascular",
+    DOSEA = 10,
+    ADOSEDUR = 0,
+    AVAL = runif(20)
+  )
+
+  # Call format_pkncaconc_data
+  df_conc <- format_pkncaconc_data(ADNCA,
+                                   group_columns = c("STUDYID", "USUBJID", "PCSPEC",
+                                                     "DRUG", "ANALYTE"),
+                                   time_column = "AFRLT")
+
+  # Generate PKNCAconc and PKNCAdose objects
+  pknca_conc <- PKNCA::PKNCAconc(
+    df_conc,
+    formula = AVAL ~ TIME | STUDYID + PCSPEC + DRUG + USUBJID / ANALYTE,
+    exclude_half.life = "exclude_half.life",
+    time.nominal = "NFRLT"
+  )
+  expect_error(format_pkncadata_intervals(pknca_conc = df_conc, data.frame()),
+               regexp = "Input pknca_conc must be a PKNCAconc object from the PKNCA package.")
+
+  expect_error(format_pkncadata_intervals(pknca_conc = pknca_conc, data.frame()),
+               regexp = "Input pknca_dose must be a PKNCAdose object from the PKNCA package.")
 })
 
 test_that("format_pkncadose_data handles negative time values", {
@@ -257,7 +285,7 @@ test_that("format_pkncadata_intervals generates correct dataset", {
     DRUG = rep("DrugA", 20),
     ANALYTE = rep("Analyte1", 20),
     AFRLT = rep(seq(0, 9), 2),
-    ARRLT = c(rep(seq(0, 4), 2), rep(seq(5, 9), 2)),
+    ARRLT = rep(seq(0, 4), 4),
     NFRLT = rep(seq(0, 9), 2),
     ROUTE = "extravascular",
     DOSEA = 10,
@@ -279,14 +307,14 @@ test_that("format_pkncadata_intervals generates correct dataset", {
                                    since_lastdose_time_column = "ARRLT")
 
   # Generate PKNCAconc and PKNCAdose objects
-  myconc <- PKNCA::PKNCAconc(
+  pknca_conc <- PKNCA::PKNCAconc(
     df_conc,
     formula = AVAL ~ TIME | STUDYID + PCSPEC + DRUG + USUBJID / ANALYTE,
     exclude_half.life = "exclude_half.life",
     time.nominal = "NFRLT"
   )
 
-  mydose <- PKNCA::PKNCAdose(
+  pknca_dose <- PKNCA::PKNCAdose(
     data = df_dose,
     formula = DOSEA ~ TIME | STUDYID + PCSPEC + DRUG + USUBJID,
     route = "ROUTE",
@@ -298,7 +326,10 @@ test_that("format_pkncadata_intervals generates correct dataset", {
   params <- c("cmax", "tmax", "half.life", "cl.obs")
 
   # Call format_pkncadata_intervals
-  myintervals <- format_pkncadata_intervals(mydose, params = params, start_from_last_dose = TRUE)
+  myintervals <- format_pkncadata_intervals(pknca_conc,
+                                            pknca_dose,
+                                            params = params,
+                                            start_from_last_dose = TRUE)
 
   # Test if myintervals is a data frame
   expect_s3_class(myintervals, "data.frame")
@@ -313,15 +344,98 @@ test_that("format_pkncadata_intervals generates correct dataset", {
   # Test if myintervals can be used with PKNCAdata by testing its output
   expect_no_error(
     PKNCA::PKNCAdata(
-      data.conc = myconc,
-      data.dose = mydose,
+      data.conc = pknca_conc,
+      data.dose = pknca_dose,
       intervals = myintervals,
       units = PKNCA::pknca_units_table(
-        concu = myconc$data$PCSTRESU[1],
-        doseu = myconc$data$DOSEU[1],
-        amountu = myconc$data$PCSTRESU[1],
-        timeu = myconc$data$RRLTU[1]
+        concu = pknca_conc$data$PCSTRESU[1],
+        doseu = pknca_conc$data$DOSEU[1],
+        amountu = pknca_conc$data$PCSTRESU[1],
+        timeu = pknca_conc$data$RRLTU[1]
+      )
+    )
+  )
+})
+
+test_that("format_pkncadata_intervals handles multiple analytes with metabolites", {
+
+  # Create mock data with multiple analytes and metabolites
+  ADNCA <- data.frame(
+    STUDYID = rep(1, 20),
+    USUBJID = rep(1, 20),
+    PCSPEC = rep("Plasma", 20),
+    DRUG = rep("DrugA", 20),
+    DOSEA = 10,
+    ANALYTE = rep(c("Analyte1", "Metabolite1"), each = 10),
+    AFRLT = rep(seq(0, 9), 2),
+    ARRLT = rep(seq(0, 4), 4),
+    AVAL = c(rep(seq(0, 8, 2), 2),
+             rep(seq(0, 4), 2))
+  )
+
+  # Call format_pkncaconc_data
+  df_conc <- format_pkncaconc_data(ADNCA,
+                                   group_columns = c("STUDYID", "USUBJID", "PCSPEC",
+                                                     "DRUG", "ANALYTE"),
+                                   time_column = "AFRLT")
+
+  # Call format_pkncadose_data
+  df_dose <- format_pkncadose_data(df_conc,
+                                   group_columns = c("STUDYID", "USUBJID", "PCSPEC",
+                                                     "DRUG"),
+                                   time_column = "AFRLT",
+                                   since_lastdose_time_column = "ARRLT")
+
+  # Generate PKNCAconc and PKNCAdose objects
+  pknca_conc <- PKNCA::PKNCAconc(
+    df_conc,
+    formula = AVAL ~ TIME | STUDYID + PCSPEC + DRUG + USUBJID / ANALYTE,
+    exclude_half.life = "exclude_half.life",
+    time.nominal = "NFRLT"
+  )
+
+  pknca_dose <- PKNCA::PKNCAdose(
+    data = df_dose,
+    formula = DOSEA ~ TIME | STUDYID + PCSPEC + DRUG + USUBJID
+  )
+
+  # Define the parameters for the dose intervals
+  params <- c("cmax", "tmax", "half.life", "cl.obs")
+
+  # Test function
+  result <- format_pkncadata_intervals(pknca_conc, pknca_dose, params = params)
+
+  expect_equal(
+    as.data.frame(result),
+    data.frame(
+      start = c(0, 0, 5, 5),
+      end = c(5, 5, Inf, Inf),
+      STUDYID = c(1, 1, 1, 1),
+      PCSPEC = c("Plasma", "Plasma", "Plasma", "Plasma"),
+      DRUG = c("DrugA", "DrugA", "DrugA", "DrugA"),
+      USUBJID = c(1, 1, 1, 1),
+      ANALYTE = c("Analyte1", "Metabolite1", "Analyte1", "Metabolite1"),
+      cmax = c(TRUE, TRUE, TRUE, TRUE),
+      tmax = c(TRUE, TRUE, TRUE, TRUE),
+      half.life = c(TRUE, TRUE, TRUE, TRUE),
+      cl.obs = c(TRUE, TRUE, TRUE, TRUE),
+      type_interval = c("main", "main", "main", "main")
+    )
+  )
+
+  # Test if result can be used with PKNCAdata by testing its output
+  expect_no_error(
+    PKNCA::PKNCAdata(
+      data.conc = pknca_conc,
+      data.dose = pknca_dose,
+      intervals = result,
+      units = PKNCA::pknca_units_table(
+        concu = "ng/mL",
+        doseu = "mg",
+        amountu = "ng",
+        timeu = "h"
       )
     )
   )
 })
+