Update create_protein_boxplot() to make it external

DESCRIPTION

@@ -26,7 +26,6 @@ Imports:
     limma,
     missForest,
     parsnip,
-    purrr,
     readxl,
     recipes,
     rlang,

NAMESPACE

@@ -5,6 +5,7 @@ export(clean_metadata)
 export(cluster_data)
 export(create_corr_heatmap)
 export(create_dir)
+export(create_protein_boxplot)
 export(do_elnet)
 export(do_limma)
 export(do_pca)

R/classification_models.R

@@ -1,9 +1,5 @@
 utils::globalVariables(c("roc_auc", ".config", ".pred_class", ".pred_0", "Scaled_Importance",
                          "Importance", "Variable", "std_err"))
-utils::globalVariables(c(
-  ".config", "test_data", "test_set", ".pred_class", ".pred_0",
-  "Scaled_Importance", "Importance", "Variable", "std_err"
-))
 #' Split data into training and test sets
 #'
 #' This function splits the data into training and test sets based on user defined ratio.
@@ -556,6 +552,8 @@ plot_var_imp <- function (finalfit_res,
 #' @param palette (character or vector). The color palette for the plot. If it is a character, it should be one of the palettes from get_hpa_palettes(). Default is NULL.
 #' @param vline (logical). Whether to add a vertical line at 50% importance. Default is TRUE.
 #' @param subtitle (vector). Vector of subtitles to include in the plot. Default is a list with all.
+#' @param nfeatures (numeric). Number of top features to include in the boxplot. Default is 9.
+#' @param points (logical). Whether to add points to the boxplot. Default is TRUE.
 #' @param seed (numeric). Seed for reproducibility. Default is 123.
 #'
 #' @return A list with results for each disease. The list contains:
@@ -662,12 +660,18 @@ do_elnet <- function(olink_data,
                                 vline = vline,
                                 subtitle)
 
+    top_features <- var_imp_res$features |>
+      dplyr::arrange(dplyr::desc(Scaled_Importance)) |>
+      dplyr::select(Variable) |>
+      dplyr::mutate(dplyr::across(dplyr::everything(), as.character)) |>
+      utils::head(nfeatures)
+    proteins <- top_features[['Variable']]
+
     boxplot_res <- create_protein_boxplot(join_data,
+                                          proteins,
                                           disease,
-                                          features = var_imp_res$features,
-                                          nfeatures = nfeatures,
-                                          palette = palette,
-                                          points = points)
+                                          points,
+                                          palette)
 
     return(list("hypopt_res" = hypopt_res,
                 "finalfit_res" = finalfit_res,

R/visualize.R

@@ -1,7 +1,27 @@
+utils::globalVariables(c("Value"))
+#' Create boxplots for proteins
+#'
+#' This function creates a boxplot for the top n proteins in the dataset.
+#' It annotates the boxplot with color for the selected disease.
+#' It is also possible to add points to the boxplot.
+#'
+#' @param join_data (tibble). The dataset with the wide Olink data joined with the metadata.
+#' @param proteins (vector). The proteins to include in the boxplot.
+#' @param disease (character). The disease to annotate.
+#' @param points (logical). Whether to add points to the boxplot.
+#' @param palette (character). The color palette to use. Default is red3.
+#'
+#' @return boxplot_panel (plot). The boxplot panel with the selected proteins.
+#' @export
+#'
+#' @examples
+#' wide_data <- widen_data(example_data, FALSE)
+#' join_data <- wide_data |>
+#'   dplyr::left_join(example_metadata |> dplyr::select(DAid, Disease, Sex))
+#' create_protein_boxplot(join_data, c("A1BG", "A2M"), "AML", palette = "cancers12")
 create_protein_boxplot <- function(join_data,
-                                   features,
+                                   proteins,
                                    disease,
-                                   nfeatures = 9,
                                    points = T,
                                    palette = NULL) {
 
@@ -16,18 +36,11 @@ create_protein_boxplot <- function(join_data,
     pal <- "red3"
   }
 
-  top_features <- features |>
-    dplyr::arrange(desc(Scaled_Importance)) |>
-    dplyr::select(Variable) |>
-    dplyr::mutate(dplyr::across(dplyr::everything(), as.character)) |>
-    head(nfeatures)
-  top_features <- top_features[['Variable']]
-
   long_data <- join_data |>
-    dplyr::select(Disease, dplyr::all_of(top_features)) |>
+    dplyr::select(Disease, dplyr::all_of(proteins)) |>
     tidyr::pivot_longer(cols = !Disease, names_to = "Protein", values_to = "Value")
 
-  long_data$Protein <- factor(long_data$Protein, levels = top_features, labels = top_features)
+  long_data$Protein <- factor(long_data$Protein, levels = proteins, labels = proteins)
 
   # Create boxplot
   boxplot <- long_data |>