Use cartesian limits for plots

qhtcp-workflow/apps/r/calculate_interaction_zscores.R

@@ -281,9 +281,12 @@ calculate_interaction_scores <- function(df, max_conc, variables, group_vars = c
   lm_AUC <- lm(Delta_AUC ~ conc_num_factor, data = stats)
 
   interactions <- stats %>%
-    transmute(
+    summarise(
       OrfRep = first(OrfRep),
       Gene = first(Gene),
+      num = first(num),
+      conc_num = first(conc_num),
+      conc_num_factor = first(conc_num_factor),
       Raw_Shift_L = first(Raw_Shift_L),
       Raw_Shift_K = first(Raw_Shift_K),
       Raw_Shift_r = first(Raw_Shift_r),
@@ -352,9 +355,26 @@ calculate_interaction_scores <- function(df, max_conc, variables, group_vars = c
     
   calculations_joined <- df %>% select(-any_of(setdiff(names(calculations), c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor"))))
   calculations_joined <- left_join(calculations_joined, calculations, by = c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor"))
+
+
+
+  # # TODO for debug
+  # df_duplicates <- df %>%
+  # group_by(OrfRep, Gene, num) %>%
+  # filter(n() > 1)
+
+  # interactions_duplicates <- interactions %>%
+  #   group_by(OrfRep, Gene, num) %>%
+  #   filter(n() > 1)
+
+  # print(df_duplicates)
+  # print(interactions_duplicates)
+
   
-  interactions_joined <- df %>% select(-any_of(setdiff(names(interactions), c("OrfRep", "Gene", "num"))))
-  interactions_joined <- left_join(interactions_joined, interactions, by = c("OrfRep", "Gene", "num"))
+
+
+  interactions_joined <- df %>% select(-any_of(setdiff(names(interactions), c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor"))))
+  interactions_joined <- left_join(interactions_joined, interactions, by = c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor"))
 
   return(list(calculations = calculations, interactions = interactions, interactions_joined = interactions_joined,
     calculations_joined = calculations_joined))
@@ -420,7 +440,20 @@ generate_and_save_plots <- function(output_dir, file_name, plot_configs, grid_la
 }
 
 generate_scatter_plot <- function(plot, config, interactive = FALSE) {
-  
+  # Check for missing or out-of-range data
+  missing_data <- config$df %>%
+    filter(
+      is.na(!!sym(config$x_var)) | is.na(!!sym(config$y_var)) |
+      !!sym(config$y_var) < min(config$ylim_vals, na.rm = TRUE) |
+      !!sym(config$y_var) > max(config$ylim_vals, na.rm = TRUE)
+    )
+
+  # Print the rows with missing or out-of-range data if any
+  if (nrow(missing_data) > 0) {
+    message("Missing or out-of-range data for ", config$title, ":")
+    print(missing_data %>% select(any_of(c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor", config$x_var, config$y_var))), n = 100)
+  }
+
   # Add the interactive `text` aesthetic if `interactive` is TRUE
   if (interactive) {
     plot <- if (!is.null(config$delta_bg_point) && config$delta_bg_point) {
@@ -475,32 +508,14 @@ generate_scatter_plot <- function(plot, config, interactive = FALSE) {
       labels = config$x_labels)
   }
 
-  # Add y-axis limits if specified
-  if (!is.null(config$ylim_vals)) {
-    plot <- plot + scale_y_continuous(limits = config$ylim_vals)
-  }
-
-  # Add Cartesian coordinates customization if specified
+  # Use coord_cartesian for zooming in without removing data outside the range
   if (!is.null(config$coord_cartesian)) {
     plot <- plot + coord_cartesian(ylim = config$coord_cartesian)
   }
 
-  return(plot)
-}
-
-generate_box_plot <- function(plot, config) {
-  plot <- plot + geom_boxplot()
-  
-  if (!is.null(config$x_breaks) && !is.null(config$x_labels) && !is.null(config$x_label)) {
-    plot <- plot + scale_x_discrete(
-      name = config$x_label,
-      breaks = config$x_breaks,
-      labels = config$x_labels
-    )
-  }
-
-  if (!is.null(config$coord_cartesian)) {
-    plot <- plot + coord_cartesian(ylim = config$coord_cartesian)
+  # Use scale_y_continuous for setting the y-axis limits
+  if (!is.null(config$ylim_vals)) {
+    plot <- plot + scale_y_continuous(limits = config$ylim_vals)
   }
 
   return(plot)
@@ -562,14 +577,13 @@ generate_interaction_plot_configs <- function(df, variables) {
     # Dynamically generate the names of the columns
     var_info <- list(
       ylim = limits_map[[variable]],
-      lm_model = df[[paste0("lm_", variable)]][[1]],
-      sd_col = paste0("WT_sd_", variable),
+      sd_col = paste0("WT_sd_", variable)
     )
 
     # Extract the precomputed linear model coefficients
     lm_line <- list(
-      intercept = coef(var_info$lm_model)[1],
-      slope = coef(var_info$lm_model)[2]
+      intercept = df[[paste0("lm_intercept_", variable)]],
+      slope = df[[paste0("lm_slope_", variable)]]
     )
 
     annotations <- lapply(names(annotation_positions[[variable]]), function(annotation_name) {
@@ -669,7 +683,6 @@ generate_rank_plot_configs <- function(df, rank_var, zscore_var, var, is_lm = FA
   return(configs)
 }
 
-
 generate_correlation_plot_configs <- function(df, variables) {
   configs <- list()
 
@@ -960,16 +973,16 @@ main <- function() {
       )
     )
 
-    message("Generating quality control plots")
-    generate_and_save_plots(out_dir_qc, "L_vs_K_before_quality_control", l_vs_k_plots)
-    generate_and_save_plots(out_dir_qc, "frequency_delta_background", frequency_delta_bg_plots)
-    generate_and_save_plots(out_dir_qc, "L_vs_K_above_threshold", above_threshold_plots)
-    generate_and_save_plots(out_dir_qc, "plate_analysis", plate_analysis_plots)
-    generate_and_save_plots(out_dir_qc, "plate_analysis_boxplots", plate_analysis_boxplots)
-    generate_and_save_plots(out_dir_qc, "plate_analysis_no_zeros", plate_analysis_no_zeros_plots)
-    generate_and_save_plots(out_dir_qc, "plate_analysis_no_zeros_boxplots", plate_analysis_no_zeros_boxplots)
-    generate_and_save_plots(out_dir_qc, "L_vs_K_for_strains_2SD_outside_mean_K", l_outside_2sd_k_plots)
-    generate_and_save_plots(out_dir_qc, "delta_background_vs_K_for_strains_2sd_outside_mean_K", delta_bg_outside_2sd_k_plots)
+    # message("Generating quality control plots")
+    # generate_and_save_plots(out_dir_qc, "L_vs_K_before_quality_control", l_vs_k_plots)
+    # generate_and_save_plots(out_dir_qc, "frequency_delta_background", frequency_delta_bg_plots)
+    # generate_and_save_plots(out_dir_qc, "L_vs_K_above_threshold", above_threshold_plots)
+    # generate_and_save_plots(out_dir_qc, "plate_analysis", plate_analysis_plots)
+    # generate_and_save_plots(out_dir_qc, "plate_analysis_boxplots", plate_analysis_boxplots)
+    # generate_and_save_plots(out_dir_qc, "plate_analysis_no_zeros", plate_analysis_no_zeros_plots)
+    # generate_and_save_plots(out_dir_qc, "plate_analysis_no_zeros_boxplots", plate_analysis_no_zeros_boxplots)
+    # generate_and_save_plots(out_dir_qc, "L_vs_K_for_strains_2SD_outside_mean_K", l_outside_2sd_k_plots)
+    # generate_and_save_plots(out_dir_qc, "delta_background_vs_K_for_strains_2sd_outside_mean_K", delta_bg_outside_2sd_k_plots)
 
     # Clean up
     rm(df, df_above_tolerance, df_no_zeros, df_no_zeros_stats, df_no_zeros_filtered_stats, ss)