Hardcode variables for rank plots

qhtcp-workflow/apps/r/calculate_interaction_zscores.R

@@ -672,20 +672,18 @@ generate_interaction_plot_configs <- function(df, variables) {
   return(configs)
 }
 
-generate_rank_plot_configs <- function(df, variables, is_lm = FALSE, adjust = FALSE) {
-  
-  df_filtered <- filter_data(df, variables, missing = TRUE)
+generate_rank_plot_configs <- function(df_filtered, is_lm = FALSE, adjust = FALSE) {
   
   # Define SD bands
   sd_bands <- c(1, 2, 3)
   
   # Define variables for Avg ZScore and Rank Avg ZScore plots
-  avg_zscore_vars <- c("r", "L", "K", "AUC")
+  variables <- c("r", "L", "K", "AUC")
   
   # Initialize list to store plot configurations
   configs <- list()
   
-  #### 1. SD-Based Plots for L and K ####
+  # SD-based plots for L and K
   for (var in c("L", "K")) {
     for (sd_band in sd_bands) {
       
@@ -741,8 +739,8 @@ generate_rank_plot_configs <- function(df, variables, is_lm = FALSE, adjust = FA
     }
   }
   
-  #### 2. Avg ZScore and Rank Avg ZScore Plots for r, L, K, and AUC ####
-  for (var in avg_zscore_vars) {
+  # Average ZScore and Rank Avg ZScore Plots for r, L, K, and AUC
+  for (var in variables) {
     for (plot_type in c("Avg_Zscore_vs_lm", "Rank_Avg_Zscore_vs_lm")) {
       
       # Define x and y variables based on plot type
@@ -821,9 +819,22 @@ generate_correlation_plot_configs <- function(df, variables) {
   return(configs)
 }
 
-filter_data <- function(df, variables, nf = FALSE, missing = FALSE, limits_map = NULL, verbose = TRUE) {
+filter_data <- function(df, variables, nf = FALSE, missing = FALSE, adjust = FALSE,
+  limits_map = NULL, verbose = TRUE) {
   
   for (variable in variables) {
+
+    avg_zscore_col <- paste0("Avg_Zscore_", var)
+    z_lm_col <- paste0("Z_lm_", var)
+    rank_col <- paste0("Rank_", var)
+    rank_lm_col <- paste0("Rank_lm_", var)
+
+    if (adjust) {
+      message("Replacing NA with 0.001 for interaction variables")
+      df[[avg_zscore_col]] <- if_else(is.na(df[[avg_zscore_col]]), 0.001, df[[avg_zscore_col]])
+      df[[z_lm_col]] <- if_else(is.na(df[[z_lm_col]]), 0.001, df[[z_lm_col]])
+    }
+
     if (nf) {
       non_finite <- df %>% filter(!is.finite(.data[[variable]]))
       if (verbose && nrow(non_finite) > 0) {
@@ -1242,33 +1253,27 @@ main <- function() {
         file = file.path(out_dir, "ZScores_Interaction_Deletion_Suppressors_K_lm.csv"), row.names = FALSE)
 
       message("Generating rank plots")
-      # Generate rank plots for L and K using standard ranks
+      zscores_interactions_joined_filtered <- filter_data(zscores_interactions_joined, variables, missing = TRUE, adjust = TRUE)
       rank_plot_configs <- generate_rank_plot_configs(
-        df = zscores_interactions_joined,
+        df = zscores_interactions_joined_filtered,
         variables = interaction_vars,
         is_lm = FALSE,
         adjust = TRUE
       )
-
-      # Save the generated rank plots for L and K
       generate_and_save_plots(output_dir = out_dir, file_name = "RankPlots",
         plot_configs = rank_plot_configs, grid_layout = list(ncol = 3, nrow = 2))
 
       message("Generating ranked linear model plots")
-      # Generate rank plots for L and K using linear model (`lm`) ranks
       rank_lm_plot_configs <- generate_rank_plot_configs(
-        df = zscores_interactions_joined,
+        df = zscores_interactions_joined_filtered,
         variables = interaction_vars,
         is_lm = TRUE,
         adjust = TRUE
       )
-
-      # Save the linear model based rank plots for L and K
       generate_and_save_plots(output_dir = out_dir, file_name = "RankPlots_lm",
         plot_configs = rank_lm_plot_configs, grid_layout = list(ncol = 3, nrow = 2))
 
       message("Filtering and reranking plots")
-      # Filter rows where either Z_lm_L or Avg_Zscore_L is not NA
       # Formerly X_NArm
       zscores_interactions_filtered <- zscores_interactions %>%
         group_by(across(all_of(orf_group_vars))) %>%
@@ -1296,7 +1301,6 @@ main <- function() {
       message("Generating filtered ranked plots")
       rank_plot_filtered_configs <- generate_rank_plot_configs(
         df = zscores_interactions_filtered,
-        variables = interaction_vars,
         is_lm = FALSE,
         adjust = FALSE
       )
@@ -1307,7 +1311,6 @@ main <- function() {
       message("Generating filtered ranked linear model plots")
       rank_plot_lm_filtered_configs <- generate_rank_plot_configs(
         df = zscores_interactions_filtered,
-        variables = interaction_vars,
         is_lm = TRUE,
         adjust = FALSE
       )