bryan/hartman-server
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Use a single N value

Browse Source
This commit is contained in:
Bryan Roessler
2024-09-05 20:18:36 -04:00
parent 6d678e661b
commit e82b6c5e59
1 changed files with 12 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
workflow/apps/r/calculate_interaction_zscores5.R
View File

@@ -133,7 +133,6 @@ load_and_process_data <- function(easy_results_file, sd = 3) {
return(df) return(df)
} }
# Update Gene names using the SGD gene list # Update Gene names using the SGD gene list
update_gene_names <- function(df, sgd_gene_list) { update_gene_names <- function(df, sgd_gene_list) {
# Load SGD gene list # Load SGD gene list
@@ -184,21 +183,20 @@ process_strains <- function(df) {
# Calculate summary statistics for all variables # Calculate summary statistics for all variables
calculate_summary_stats <- function(df, variables, group_vars = c("conc_num", "conc_num_factor")) { calculate_summary_stats <- function(df, variables, group_vars = c("conc_num", "conc_num_factor")) {
# Generate summary statistics
summary_stats <- df %>% summary_stats <- df %>%
group_by(across(all_of(group_vars))) %>% group_by(across(all_of(group_vars))) %>%
reframe(across(all_of(variables), list( reframe(across(all_of(variables), list(
N = ~sum(!is.na(.)), # Count of non-NA values mean = ~mean(., na.rm = TRUE),
mean = ~mean(., na.rm = TRUE), # Mean ignoring NAs median = ~median(., na.rm = TRUE),
median = ~median(., na.rm = TRUE), # Median ignoring NAs max = ~ifelse(all(is.na(.)), NA, max(., na.rm = TRUE)),
max = ~ifelse(all(is.na(.)), NA, max(., na.rm = TRUE)), # Return NA if all values are NA min = ~ifelse(all(is.na(.)), NA, min(., na.rm = TRUE)),
min = ~ifelse(all(is.na(.)), NA, min(., na.rm = TRUE)), # Return NA if all values are NA sd = ~sd(., na.rm = TRUE),
sd = ~sd(., na.rm = TRUE), # Standard deviation ignoring NAs se = ~ifelse(sum(!is.na(.)) > 1, sd(., na.rm = TRUE) / sqrt(sum(!is.na(.)) - 1), NA)
se = ~ifelse(N > 1, sd(., na.rm = TRUE) / sqrt(N - 1), NA) # Standard Error using precomputed N
# TODO: not in original stats but better to do here than in calculate_interactions? # TODO: not in original stats but better to do here than in calculate_interactions?
# z_max = ~ifelse(sd(., na.rm = TRUE) == 0 | all(is.na(.)), NA, # z_max = ~ifelse(sd(., na.rm = TRUE) == 0 | all(is.na(.)), NA,
# (max(., na.rm = TRUE) - mean(., na.rm = TRUE)) / sd(., na.rm = TRUE)) # Z-score # (max(., na.rm = TRUE) - mean(., na.rm = TRUE)) / sd(., na.rm = TRUE)) # Z-score
), .names = "{.fn}_{.col}")) ), .names = "{.fn}_{.col}")) %>%
mutate(N = ~sum(!is.na(L))) # count of non-NA L values
# Join the summary stats back to the original dataframe # Join the summary stats back to the original dataframe
df_with_stats <- left_join(df, summary_stats, by = group_vars) df_with_stats <- left_join(df, summary_stats, by = group_vars)
@@ -207,7 +205,7 @@ calculate_summary_stats <- function(df, variables, group_vars = c("conc_num", "c
return(list(summary_stats = summary_stats, df_with_stats = df_with_stats)) return(list(summary_stats = summary_stats, df_with_stats = df_with_stats))
} }
# Calculate interaction scores
calculate_interaction_scores <- function(df, max_conc, variables, group_vars = c("OrfRep", "Gene", "num")) { calculate_interaction_scores <- function(df, max_conc, variables, group_vars = c("OrfRep", "Gene", "num")) {
# Calculate total concentration variables # Calculate total concentration variables
@@ -691,6 +689,9 @@ main <- function() {
file = file.path(out_dir, paste0("SummaryStats_BackgroundStrains_", strain, ".csv")), file = file.path(out_dir, paste0("SummaryStats_BackgroundStrains_", strain, ".csv")),
row.names = FALSE) row.names = FALSE)
print("Background summary stats:")
print(head(summary_stats_bg))
# Filter reference and deletion strains # Filter reference and deletion strains
# Formerly X2_RF (reference strain) # Formerly X2_RF (reference strain)
df_reference <- df_bg_stats %>% df_reference <- df_bg_stats %>%