From d5141d04c4a550e8fe353cd059f8f59db288f8f7 Mon Sep 17 00:00:00 2001
From: Bryan Roessler <bryanroessler@gmail.com>
Date: Mon, 16 Sep 2024 21:10:22 -0400
Subject: [PATCH] Refactor data filtering

---
 .../apps/r/calculate_interaction_zscores.R    | 43 ++++---------------
 1 file changed, 9 insertions(+), 34 deletions(-)

diff --git a/qhtcp-workflow/apps/r/calculate_interaction_zscores.R b/qhtcp-workflow/apps/r/calculate_interaction_zscores.R
index 945cd74d..cc65a84e 100644
--- a/qhtcp-workflow/apps/r/calculate_interaction_zscores.R
+++ b/qhtcp-workflow/apps/r/calculate_interaction_zscores.R
@@ -155,9 +155,7 @@ update_gene_names <- function(df, sgd_gene_list) {
   return(df)
 }
 
-# Calculate summary statistics for all variables
 calculate_summary_stats <- function(df, variables, group_vars = c("OrfRep", "conc_num", "conc_num_factor")) {
-
   # Summarize the variables within the grouped data
   summary_stats <- df %>%
     group_by(across(all_of(group_vars))) %>%
@@ -166,19 +164,18 @@ calculate_summary_stats <- function(df, variables, group_vars = c("OrfRep", "con
       across(all_of(variables), list(
         mean = ~mean(., na.rm = TRUE),
         median = ~median(., na.rm = TRUE),
-        max = ~ ifelse(all(is.na(.)), NA, max(., na.rm = TRUE)),
-        min = ~ ifelse(all(is.na(.)), NA, min(., na.rm = TRUE)),
+        max = ~ifelse(all(is.na(.)), NA, max(., na.rm = TRUE)),
+        min = ~ifelse(all(is.na(.)), NA, min(., na.rm = TRUE)),
         sd = ~sd(., na.rm = TRUE),
-        se = ~ ifelse(all(is.na(.)), NA, sd(., na.rm = TRUE) / sqrt(sum(!is.na(.)) - 1))
+        se = ~ifelse(all(is.na(.)), NA, sd(., na.rm = TRUE) / sqrt(sum(!is.na(.)) - 1))
       ), .names = "{.fn}_{.col}")
     )
 
-  # print(summary_stats)
-
-  # Prevent .x and .y suffix issues by renaming columns
+  # Create a cleaned version of df that doesn't overlap with summary_stats
+  cols_to_exclude <- unique(c(variables, group_vars))
   df_cleaned <- df %>%
-    select(-any_of(setdiff(names(summary_stats), group_vars))) # Avoid duplicate columns in the final join
-  
+    select(-any_of(cols_to_exclude))
+
   # Join the stats back to the original dataframe
   df_with_stats <- left_join(df_cleaned, summary_stats, by = group_vars)
 
@@ -205,22 +202,8 @@ calculate_interaction_scores <- function(df, max_conc, variables, group_vars = c
     AUC = df %>% filter(conc_num_factor == 0) %>% pull(sd_AUC) %>% first()
   )
 
-  stats <- df %>%
-    group_by(OrfRep, Gene, num, conc_num, conc_num_factor) %>%
-    summarise(
-      N = sum(!is.na(L)),
-      NG = sum(NG, na.rm = TRUE),
-      DB = sum(DB, na.rm = TRUE),
-      SM = sum(SM, na.rm = TRUE),
-      across(all_of(variables), list(
-        mean = ~mean(., na.rm = TRUE),
-        median = ~median(., na.rm = TRUE),
-        max = ~ifelse(all(is.na(.)), NA, max(., na.rm = TRUE)),
-        min = ~ifelse(all(is.na(.)), NA, min(., na.rm = TRUE)),
-        sd = ~sd(., na.rm = TRUE),
-        se = ~ifelse(sum(!is.na(.)) > 1, sd(., na.rm = TRUE) / sqrt(sum(!is.na(.)) - 1), NA)
-      ), .names = "{.fn}_{.col}")
-    )
+  stats <- calculate_summary_stats(df, variables,
+    group_vars = c("OrfRep", "Gene", "num", "conc_num", "conc_num_factor"))$summary_stats
 
   stats <- df %>%
     group_by(OrfRep, Gene, num) %>%
@@ -873,14 +856,6 @@ filter_data <- function(df, variables, nf = FALSE, missing = FALSE, adjust = FAL
   avg_zscore_cols <- paste0("Avg_Zscore_", variables)
   z_lm_cols <- paste0("Z_lm_", variables)
   
-  # # Optional: Validate that the expected columns exist in the dataframe
-  # expected_cols <- c(avg_zscore_cols, z_lm_cols, variables)
-  # missing_cols <- setdiff(expected_cols, names(df))
-  # if (length(missing_cols) > 0) {
-  #   stop("The following expected columns are missing from the dataframe: ",
-  #        paste(missing_cols, collapse = ", "))
-  # }
-  
   # Adjust NAs if 'adjust' is TRUE
   if (adjust) {
     if (verbose) message("Replacing NA with 0.001 for Avg_Zscore_ and Z_lm_ columns.")