Merge pull request #166 from pharmaverse/enhancement/multiple-analytes

Enhancement: NCA for multiple analytes

R/format_data.R

@@ -47,7 +47,11 @@ format_pkncaconc_data <- function(ADNCA, group_columns, time_column = "AFRLT") {
 #'
 #' This function creates a pharmacokinetic dose dataset from the provided concentration data.
 #'
-#' @param ADNCA_conc A data frame containing the concentration data.
+#' @param pkncaconc_data A data frame containing the concentration data.
+#' @param group_columns A character vector specifying the columns to group by.
+#' @param dosno_column A character string specifying the dose number column.
+#' @param time_column A character string specifying the time column.
+#' @param since_lastdose_time_column A character string specifying the time since last dose column.
 #'
 #' @returns A data frame containing the dose data.
 #'
@@ -93,6 +97,7 @@ format_pkncadose_data <- function(pkncaconc_data,
     slice(1) %>%
     ungroup() %>%
     arrange(!!!syms(group_columns))
+
 }
 
 #' Create Dose Intervals Dataset

R/general_lineplot.R

@@ -6,6 +6,7 @@
 #'
 #' @param data              A data frame containing the ADNCA dataset.
 #' @param selected_analytes A character vector of selected analytes to be included in the plot.
+#' @param selected_pcspec   A character vector of selected matrix to be included in the plot.
 #' @param selected_usubjids A character vector of selected unique subject identifiers (USUBJIDs)
 #'                          to be included in the plot.
 #' @param colorby_var       A character string specifying the variable by which to color
@@ -22,7 +23,7 @@
 #' @details
 #' The function performs the following steps:a
 #' \itemize{
-#'   \item Filters the data based on the selected analytes and subjects.
+#'   \item Filters the data based on the selected analytes, matrices, and subjects.
 #'   \item Selects relevant columns and removes rows with missing concentration values.
 #'   \item Converts 'USUBJID', 'DOSNO', and 'DOSEA' to factors.
 #'   \item Filters the data by cycle if `time_scale` is "By Cycle".
@@ -36,6 +37,7 @@
 #'   # Example usage:
 #'   plot <- general_lineplot(data = adnca_data,
 #'                            selected_analytes = c("Analyte1", "Analyte2"),
+#'                            selected_pcspec = c("Spec1", "Spec2"),
 #'                            selected_usubjids = c("Subject1", "Subject2"),
 #'                            colorby_var = "DOSNO",
 #'                            time_scale = "By Cycle",
@@ -50,7 +52,8 @@
 #' @importFrom tern g_ipp
 #' @export
 general_lineplot <- function(
-  data, selected_analytes, selected_usubjids, colorby_var, time_scale, yaxis_scale, cycle = NULL
+  data, selected_analytes, selected_pcspec, selected_usubjids,
+  colorby_var, time_scale, yaxis_scale, cycle = NULL
 ) {
 
   # Check if the data is empty
@@ -63,6 +66,7 @@ general_lineplot <- function(
     filter(
       USUBJID %in% selected_usubjids,
       ANALYTE %in% selected_analytes,
+      PCSPEC %in% selected_pcspec,
       if ("EVID" %in% names(data)) EVID == 0 else TRUE
     ) %>%
     filter(!is.na(AVAL)) %>%
@@ -72,6 +76,10 @@ general_lineplot <- function(
       DOSEA = factor(DOSEA),
       id_var = interaction(!!!syms(colorby_var), sep = ", ")
     )
+  # Check if the data is empty
+  if (nrow(preprocessed_data) == 0) {
+    return(ggplot() + ggtitle("No data available for selected parameters"))
+  }
 
   # If there are predose records duplicate them in the previous line so they are considered
   if ("ARRLT" %in% names(preprocessed_data) &&
@@ -125,8 +133,8 @@ general_lineplot <- function(
     subtitle = paste0(
       "Subjects: ",
       paste(unique(preprocessed_data$USUBJID), collapse = ", "),
-      "\nAnalyte: ",
-      unique(preprocessed_data$ANALYTE)
+      "\nAnalyte(s): ",
+      paste(unique(preprocessed_data$ANALYTE), collapse = ", ")
     ),
     caption = NULL,
     add_baseline_hline = FALSE,

R/general_meanplot.R

@@ -7,6 +7,7 @@
 #' @param data              A data frame containing the ADNCA dataset.
 #' @param selected_studyids A character vector of selected study IDs to be included in the plot.
 #' @param selected_analytes A character vector of selected analytes to be included in the plot.
+#' @param selected_pcspecs  A character vector of selected matrices to be included in the plot.
 #' @param selected_cycles   A character vector or numeric vector of selected cycles to be
 #'                          included in the plot.
 #' @param id_variable       A character string specifying the variable by which to color the lines
@@ -28,6 +29,7 @@
 general_meanplot <- function(data,
                              selected_studyids,
                              selected_analytes,
+                             selected_pcspecs,
                              selected_cycles,
                              id_variable = "DOSEA",
                              plot_ylog = FALSE,
@@ -39,6 +41,7 @@ general_meanplot <- function(data,
     filter(
       STUDYID %in% selected_studyids,
       ANALYTE %in% selected_analytes,
+      PCSPEC %in% selected_pcspecs,
       DOSNO %in% selected_cycles,
       if ("EVID" %in% names(data)) EVID == 0 else TRUE,
       NRRLT > 0
@@ -50,7 +53,7 @@ general_meanplot <- function(data,
   summarised_data <- preprocessed_data %>%
     mutate(id_variable = as.factor(!!sym(id_variable))) %>%
     # Create a groups variables for the labels
-    mutate(groups = paste(STUDYID, ANALYTE, DOSNO, sep = ", ")) %>%
+    mutate(groups = paste(STUDYID, ANALYTE, PCSPEC, DOSNO, sep = ", ")) %>%
     group_by(id_variable, NRRLT, groups) %>%
     summarise(
               Mean = round(mean(AVAL, na.rm = TRUE), 3),

R/lambda_slope_plot.R

@@ -9,6 +9,8 @@
 #' @param conc_pknca_df  Data frame containing the concentration data
 #'                      (default is `mydata$conc$data`).
 #' @param dosno         Numeric value representing the dose number (default is `profile`).
+#' @param analyte       Character value representing the analyte (default is `analyte`).
+#' @param pcspec        Character value representing the matrix (default is `pcspec`).
 #' @param usubjid       Character value representing the unique subject identifier
 #'                      (default is `patient`).
 #' @param R2ADJTHRESHOL Numeric value representing the R-squared adjusted threshold for determining
@@ -37,6 +39,8 @@
 #'   # Example usage:
 #'   plot <- lambda_slope_plot(res_pknca_df = myres$result,
 #'                             conc_pknca_df = mydata$conc$data,
+#'                             analyte = "analyte_1",
+#'                             pcspec = "pcspec_1",
 #'                             dosno = 1,
 #'                             usubjid = "subject_1",
 #'                             R2ADJTHRESHOL = 0.7)
@@ -52,14 +56,17 @@ lambda_slope_plot <- function(
   conc_pknca_df = mydata$conc$data,
   dosno = profile,
   usubjid = patient,
+  analyte = analyte,
+  pcspec = pcspec,
   R2ADJTHRESHOL = 0.7
 ) {
 
   # Obtain all information relevant regarding lambda calculation
   lambda_res <- res_pknca_df %>%
-    filter(DOSNO == dosno, USUBJID == usubjid, type_interval == "main")  %>%
-    arrange(USUBJID, DOSNO, start, desc(end)) %>%
-    filter(!duplicated(paste0(USUBJID, DOSNO, PPTESTCD)))
+    filter(DOSNO == dosno, USUBJID == usubjid, ANALYTE == analyte,
+           PCSPEC == pcspec, type_interval == "main")  %>%
+    arrange(USUBJID, DOSNO, ANALYTE, PCSPEC, start, desc(end)) %>%
+    filter(!duplicated(paste0(USUBJID, DOSNO, PCSPEC, ANALYTE, PPTESTCD)))
 
   # Obtain the number of data points used to calculate lambda
   lambda_z_n_points <- as.numeric(lambda_res$PPORRES[lambda_res$PPTESTCD == "lambda.z.n.points"])
@@ -141,7 +148,7 @@ lambda_slope_plot <- function(
 
   # Include in the data the aesthetics for the plot
   plot_data <- conc_pknca_df %>%
-    filter(DOSNO == dosno, USUBJID == usubjid) %>%
+    filter(DOSNO == dosno, USUBJID == usubjid, ANALYTE == analyte, PCSPEC == pcspec) %>%
     arrange(IX) %>%
     mutate(
       IX_shape = ifelse(is.excluded.hl, "excluded", "included"),
@@ -174,7 +181,8 @@ lambda_slope_plot <- function(
       size = 5
     ) +
     labs(
-      title = paste0("USUBJID: ", usubjid, ", DOSNO: ", dosno),
+      title = paste0("USUBJID: ", usubjid, ", DOSNO: ", dosno,
+                     ", ANALYTE: ", analyte, ", PCSPEC: ", pcspec),
       y = paste0("Log10 Concentration (", conc_pknca_df $PCSTRESU[1], ")"),
       x = paste0("Actual time post dose (", conc_pknca_df $RRLTU[1], ")")
     ) +
@@ -229,7 +237,8 @@ lambda_slope_plot <- function(
   pl <- pl %>%
     # Make this trace the only one
     add_trace(
-      data = plot_data %>% filter(DOSNO == dosno, USUBJID == usubjid),
+      data = plot_data %>% filter(DOSNO == dosno, USUBJID == usubjid,
+                                  ANALYTE == analyte, PCSPEC == pcspec),
       x = ~TIME, y = ~log10(AVAL),
       customdata = ~paste0(USUBJID, "_", DOSNO, "_", IX),
       text = ~paste0("Data Point: ", IX, "\n", "(", signif(TIME, 2), " , ", signif(AVAL, 2), ")"),

R/utils-slope_selector.R

@@ -28,12 +28,12 @@
 
   # Eliminate all rows with conflicting or blank values
   slopes <- slopes %>%
-    dplyr::filter(
-      TYPE %in% c("Selection", "Exclusion"),
-      PATIENT %in% names(profiles),
-      PROFILE %in% unname(unlist(profiles[PATIENT])),
-      all(!is.na(sapply(IXrange, function(x) .eval_range(x))))
-    )
+    filter(TYPE %in% c("Selection", "Exclusion")) %>%
+    semi_join(
+      select(profiles, USUBJID, ANALYTE, PCSPEC, DOSNO),
+      by = c("PATIENT" = "USUBJID", "ANALYTE", "PCSPEC", "PROFILE" = "DOSNO")
+    ) %>%
+    filter(all(!is.na(sapply(IXrange, .eval_range))))
 
   if (nrow(slopes) != 0) {
     # Go over all rules and check if there is no overlap - if there is, edit accordingly
@@ -47,6 +47,8 @@
   for (i in seq_len(nrow(slopes))) {
     selection_index <- which(
       data$conc$data$USUBJID == slopes$PATIENT[i] &
+        data$conc$data$ANALYTE == slopes$ANALYTE[i] &
+        data$conc$data$PCSPEC == slopes$PCSPEC[i] &
         data$conc$data$DOSNO == slopes$PROFILE[i] &
         data$conc$data$IX %in% .eval_range(slopes$IXrange[i])
     )
@@ -61,7 +63,7 @@
   }
 
   data$conc$data <- data$conc$data %>%
-    dplyr::group_by(STUDYID, USUBJID, PCSPEC, DOSNO) %>%
+    dplyr::group_by(STUDYID, USUBJID, ANALYTE, PCSPEC, DOSNO) %>%
     dplyr::mutate(exclude_half.life = {
       if (any(is.included.hl)) {
         is.excluded.hl | !is.included.hl
@@ -89,6 +91,8 @@
   existing_index <- which(
     existing$TYPE == new$TYPE &
       existing$PATIENT == new$PATIENT &
+      existing$ANALYTE == new$ANALYTE &
+      existing$PCSPEC == new$PCSPEC &
       existing$PROFILE == new$PROFILE
   )