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

Move correlation modeling to calculate_interction_scores

Browse Source
This commit is contained in:
Bryan Roessler
2024-10-04 02:30:13 -04:00
parent b23c6dafef
commit c4f398be82
1 changed files with 187 additions and 194 deletions
Download Patch File Download Diff File Expand all files Collapse all files

381
qhtcp-workflow/apps/r/calculate_interaction_zscores.R
View File

@@ -214,7 +214,7 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
# Calculate WT statistics from df_bg
wt_stats <- df_bg %>%
filter(conc_num == 0) %>% # use the zero drug concentration background
filter(conc_num == 0) %>%
summarise(
WT_L = mean(mean_L, na.rm = TRUE),
WT_sd_L = mean(sd_L, na.rm = TRUE),
@@ -294,7 +294,7 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
Delta_r = if_else(NG == 1, mean_r - WT_r, Delta_r),
Delta_AUC = if_else(NG == 1, mean_AUC - WT_AUC, Delta_AUC),
Delta_L = if_else(SM == 1, mean_L - WT_L, Delta_L),
# Calculate Z-scores
Zscore_L = Delta_L / WT_sd_L,
Zscore_K = Delta_K / WT_sd_K,
@@ -302,31 +302,48 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
Zscore_AUC = Delta_AUC / WT_sd_AUC
) %>%
group_modify(~ {
# Perform linear models
lm_L <- lm(Delta_L ~ conc_num_factor, data = .x)
lm_K <- lm(Delta_K ~ conc_num_factor, data = .x)
lm_r <- lm(Delta_r ~ conc_num_factor, data = .x)
lm_AUC <- lm(Delta_AUC ~ conc_num_factor, data = .x)
.x %>%
mutate(
lm_intercept_L = coef(lm_L)[1],
lm_slope_L = coef(lm_L)[2],
R_Squared_L = summary(lm_L)$r.squared,
lm_Score_L = max_conc * lm_slope_L + lm_intercept_L,
lm_intercept_K = coef(lm_K)[1],
lm_slope_K = coef(lm_K)[2],
R_Squared_K = summary(lm_K)$r.squared,
lm_Score_K = max_conc * lm_slope_K + lm_intercept_K,
lm_intercept_r = coef(lm_r)[1],
lm_slope_r = coef(lm_r)[2],
R_Squared_r = summary(lm_r)$r.squared,
lm_Score_r = max_conc * lm_slope_r + lm_intercept_r,
lm_intercept_AUC = coef(lm_AUC)[1],
lm_slope_AUC = coef(lm_AUC)[2],
R_Squared_AUC = summary(lm_AUC)$r.squared,
lm_Score_AUC = max_conc * lm_slope_AUC + lm_intercept_AUC
)
# Perform linear models only if there are enough unique conc_num_factor levels
if (length(unique(.x$conc_num_factor)) > 1) {
# Filter and calculate each lm() separately with individual checks for NAs
lm_L <- if (!all(is.na(.x$Delta_L))) tryCatch(lm(Delta_L ~ conc_num_factor, data = .x), error = function(e) NULL) else NULL
lm_K <- if (!all(is.na(.x$Delta_K))) tryCatch(lm(Delta_K ~ conc_num_factor, data = .x), error = function(e) NULL) else NULL
lm_r <- if (!all(is.na(.x$Delta_r))) tryCatch(lm(Delta_r ~ conc_num_factor, data = .x), error = function(e) NULL) else NULL
lm_AUC <- if (!all(is.na(.x$Delta_AUC))) tryCatch(lm(Delta_AUC ~ conc_num_factor, data = .x), error = function(e) NULL) else NULL
# Mutate results for each lm if it was successfully calculated, suppress warnings for perfect fits
.x %>%
mutate(
lm_intercept_L = if (!is.null(lm_L)) coef(lm_L)[1] else NA,
lm_slope_L = if (!is.null(lm_L)) coef(lm_L)[2] else NA,
R_Squared_L = if (!is.null(lm_L)) suppressWarnings(summary(lm_L)$r.squared) else NA,
lm_Score_L = if (!is.null(lm_L)) max_conc * coef(lm_L)[2] + coef(lm_L)[1] else NA,
lm_intercept_K = if (!is.null(lm_K)) coef(lm_K)[1] else NA,
lm_slope_K = if (!is.null(lm_K)) coef(lm_K)[2] else NA,
R_Squared_K = if (!is.null(lm_K)) suppressWarnings(summary(lm_K)$r.squared) else NA,
lm_Score_K = if (!is.null(lm_K)) max_conc * coef(lm_K)[2] + coef(lm_K)[1] else NA,
lm_intercept_r = if (!is.null(lm_r)) coef(lm_r)[1] else NA,
lm_slope_r = if (!is.null(lm_r)) coef(lm_r)[2] else NA,
R_Squared_r = if (!is.null(lm_r)) suppressWarnings(summary(lm_r)$r.squared) else NA,
lm_Score_r = if (!is.null(lm_r)) max_conc * coef(lm_r)[2] + coef(lm_r)[1] else NA,
lm_intercept_AUC = if (!is.null(lm_AUC)) coef(lm_AUC)[1] else NA,
lm_slope_AUC = if (!is.null(lm_AUC)) coef(lm_AUC)[2] else NA,
R_Squared_AUC = if (!is.null(lm_AUC)) suppressWarnings(summary(lm_AUC)$r.squared) else NA,
lm_Score_AUC = if (!is.null(lm_AUC)) max_conc * coef(lm_AUC)[2] + coef(lm_AUC)[1] else NA
)
} else {
# If not enough conc_num_factor levels, set lm-related values to NA
.x %>%
mutate(
lm_intercept_L = NA, lm_slope_L = NA, R_Squared_L = NA, lm_Score_L = NA,
lm_intercept_K = NA, lm_slope_K = NA, R_Squared_K = NA, lm_Score_K = NA,
lm_intercept_r = NA, lm_slope_r = NA, R_Squared_r = NA, lm_Score_r = NA,
lm_intercept_AUC = NA, lm_slope_AUC = NA, R_Squared_AUC = NA, lm_Score_AUC = NA
)
}
}) %>%
ungroup()
@@ -344,6 +361,7 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
.groups = "drop"
)
# Add lm score means and standard deviations to calculations
calculations <- calculations %>%
mutate(
lm_mean_L = lm_means_sds$lm_mean_L,
@@ -356,7 +374,7 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
lm_sd_AUC = lm_means_sds$lm_sd_AUC
)
# Continue with gene Z-scores and interactions
# Calculate Z-lm scores
calculations <- calculations %>%
mutate(
Z_lm_L = (lm_Score_L - lm_mean_L) / lm_sd_L,
@@ -397,7 +415,7 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
R_Squared_K = first(R_Squared_K),
R_Squared_r = first(R_Squared_r),
R_Squared_AUC = first(R_Squared_AUC),
# NG, DB, SM values
NG = first(NG),
DB = first(DB),
@@ -406,12 +424,34 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
.groups = "drop"
)
# Add overlap threshold categories based on Z-lm and Avg-Z scores
interactions <- interactions %>%
filter(!is.na(Z_lm_L)) %>%
mutate(
Overlap = case_when(
Z_lm_L >= overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Enhancer Both",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Suppressor Both",
Z_lm_L >= overlap_threshold & Avg_Zscore_L <= overlap_threshold ~ "Deletion Enhancer lm only",
Z_lm_L <= overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Enhancer Avg Zscore only",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L >= -overlap_threshold ~ "Deletion Suppressor lm only",
Z_lm_L >= -overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Suppressor Avg Zscore only",
Z_lm_L >= overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Enhancer lm, Deletion Suppressor Avg Zscore",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Suppressor lm, Deletion Enhancer Avg Zscore",
TRUE ~ "No Effect"
),
# For correlations
lm_R_squared_L = if (!all(is.na(Z_lm_L)) && !all(is.na(Avg_Zscore_L))) summary(lm(Z_lm_L ~ Avg_Zscore_L))$r.squared else NA,
lm_R_squared_K = if (!all(is.na(Z_lm_K)) && !all(is.na(Avg_Zscore_K))) summary(lm(Z_lm_K ~ Avg_Zscore_K))$r.squared else NA,
lm_R_squared_r = if (!all(is.na(Z_lm_r)) && !all(is.na(Avg_Zscore_r))) summary(lm(Z_lm_r ~ Avg_Zscore_r))$r.squared else NA,
lm_R_squared_AUC = if (!all(is.na(Z_lm_AUC)) && !all(is.na(Avg_Zscore_AUC))) summary(lm(Z_lm_AUC ~ Avg_Zscore_AUC))$r.squared else NA
)
# Creating the final calculations and interactions dataframes with only required columns for csv output
calculations_df <- calculations %>%
select(
all_of(group_vars),
conc_num, conc_num_factor, conc_num_factor_factor,
N, NG, DB, SM,
conc_num, conc_num_factor, conc_num_factor_factor, N,
mean_L, median_L, sd_L, se_L,
mean_K, median_K, sd_K, se_K,
mean_r, median_r, sd_r, se_r,
@@ -432,23 +472,22 @@ calculate_interaction_scores <- function(df, df_bg, group_vars, overlap_threshol
Avg_Zscore_L, Avg_Zscore_K, Avg_Zscore_r, Avg_Zscore_AUC,
Z_lm_L, Z_lm_K, Z_lm_r, Z_lm_AUC,
Raw_Shift_L, Raw_Shift_K, Raw_Shift_r, Raw_Shift_AUC,
Z_Shift_L, Z_Shift_K, Z_Shift_r, Z_Shift_AUC
Z_Shift_L, Z_Shift_K, Z_Shift_r, Z_Shift_AUC,
lm_R_squared_L, lm_R_squared_K, lm_R_squared_r, lm_R_squared_AUC,
Overlap
)
# Join calculations and interactions to avoid dimension mismatch
calculations_no_overlap <- calculations %>%
# DB, NG, SM are same as in interactions, the rest may be different and need to be checked
select(-any_of(c(
"DB", "NG", "SM",
select(-any_of(c("DB", "NG", "SM",
"Raw_Shift_L", "Raw_Shift_K", "Raw_Shift_r", "Raw_Shift_AUC",
"Z_Shift_L", "Z_Shift_K", "Z_Shift_r", "Z_Shift_AUC",
"Z_lm_L", "Z_lm_K", "Z_lm_r", "Z_lm_AUC"
)))
"Z_lm_L", "Z_lm_K", "Z_lm_r", "Z_lm_AUC")))
# Use left_join to avoid dimension mismatch issues
full_data <- calculations_no_overlap %>%
left_join(interactions, by = group_vars)
left_join(interactions_df, by = group_vars)
# Return full_data and the two required dataframes (calculations and interactions)
# Return final dataframes
return(list(
calculations = calculations_df,
interactions = interactions_df,
@@ -781,7 +820,7 @@ generate_plate_analysis_plot_configs <- function(variables, df_before = NULL, df
}
generate_interaction_plot_configs <- function(df, type) {
# Define the y-limits for the plots
limits_map <- list(
L = c(0, 130),
@@ -793,16 +832,16 @@ generate_interaction_plot_configs <- function(df, type) {
stats_plot_configs <- list()
stats_boxplot_configs <- list()
delta_plot_configs <- list()
# Overall statistics plots (use df)
# Overall statistics plots
OrfRep <- first(df$OrfRep) # this should correspond to the reference strain
for (plot_type in c("scatter", "box")) {
for (var in names(limits_map)) {
y_limits <- limits_map[[var]]
y_span <- y_limits[2] - y_limits[1]
# Common plot configuration
plot_config <- list(
df = df,
@@ -814,7 +853,7 @@ generate_interaction_plot_configs <- function(df, type) {
x_breaks = unique(df$conc_num_factor_factor),
x_labels = as.character(unique(df$conc_num))
)
# Add specific configurations for scatter and box plots
if (plot_type == "scatter") {
plot_config$plot_type <- "scatter"
@@ -827,96 +866,91 @@ generate_interaction_plot_configs <- function(df, type) {
center_point = TRUE
)
plot_config$position <- "jitter"
# Annotation labels
annotations <- list(
list(x = 0, y = y_limits[1] + 0.1 * y_span, label = "NG ="),
list(x = 0, y = y_limits[1] + 0.05 * y_span, label = "DB ="),
list(x = 0, y = y_limits[1], label = "SM =")
)
# Annotation values
for (x_val in unique(df$conc_num_factor_factor)) {
current_df <- df %>% filter(.data[[plot_config$x_var]] == x_val)
annotations <- append(annotations, list(
list(x = x_val, y = y_limits[1] + 0.1 * y_span, label = sum(current_df$NG, na.rm = TRUE)),
list(x = x_val, y = y_limits[1] + 0.05 * y_span, label = sum(current_df$DB, na.rm = TRUE)),
list(x = x_val, y = y_limits[1], label = sum(current_df$SM, na.rm = TRUE))
))
}
plot_config$annotations <- annotations
annotations <- list(
list(x = 0.25, y = y_limits[1] + 0.1 * y_span, label = " NG ="), # Slightly above y-min
list(x = 0.25, y = y_limits[1] + 0.05 * y_span, label = " DB ="),
list(x = 0.25, y = y_limits[1], label = " SM =")
)
# Loop over unique x values and add NG, DB, SM values at calculated y positions
for (x_val in unique(df$conc_num_factor_factor)) {
current_df <- df %>% filter(.data[[plot_config$x_var]] == x_val)
annotations <- append(annotations, list(
list(x = x_val, y = y_limits[1] + 0.1 * y_span, label = first(current_df$NG, default = 0)),
list(x = x_val, y = y_limits[1] + 0.05 * y_span, label = first(current_df$DB, default = 0)),
list(x = x_val, y = y_limits[1], label = first(current_df$SM, default = 0))
))
}
plot_config$annotations <- annotations
# Append to scatter plot configurations
stats_plot_configs <- append(stats_plot_configs, list(plot_config))
} else if (plot_type == "box") {
plot_config$plot_type <- "box"
plot_config$title <- sprintf("%s Boxplot RF for %s with SD", OrfRep, var)
plot_config$position <- "dodge" # Boxplots don't need jitter, use dodge instead
# Append to boxplot configurations
stats_boxplot_configs <- append(stats_boxplot_configs, list(plot_config))
}
}
}
# Delta interaction plots
if (type == "reference") {
group_vars <- c("OrfRep", "Gene", "num")
} else if (type == "deletion") {
group_vars <- c("OrfRep", "Gene")
}
delta_limits_map <- list(
L = c(-60, 60),
K = c(-60, 60),
r = c(-0.6, 0.6),
AUC = c(-6000, 6000)
)
grouped_data <- df_calculations %>%
grouped_data <- df %>%
group_by(across(all_of(group_vars))) %>%
group_split()
for (group_data in grouped_data) {
OrfRep <- first(group_data$OrfRep)
Gene <- first(group_data$Gene)
num <- if ("num" %in% names(group_data)) first(group_data$num) else ""
if (type == "reference") {
OrfRepTitle <- paste(OrfRep, Gene, num, sep = "_")
} else if (type == "deletion") {
OrfRepTitle <- OrfRep
}
# Get corresponding interaction row
interaction_row <- df_interactions %>%
filter(if_all(all_of(group_vars), ~ . == first(.))) %>%
slice(1)
for (var in names(delta_limits_map)) {
y_limits <- delta_limits_map[[var]]
y_span <- y_limits[2] - y_limits[1]
# Error bars
WT_sd_value <- first(group_data[[paste0("WT_sd_", var)]], default = 0)
# Z_Shift and lm values from interaction_row
Z_Shift_value <- round(first(interaction_row[[paste0("Z_Shift_", var)]], default = 0), 2)
Z_lm_value <- round(first(interaction_row[[paste0("Z_lm_", var)]], default = 0), 2)
R_squared_value <- round(first(interaction_row[[paste0("R_Squared_", var)]], default = 0), 2)
# NG, DB, SM values from interaction_row
NG_value <- first(interaction_row$NG, default = 0)
DB_value <- first(interaction_row$DB, default = 0)
SM_value <- first(interaction_row$SM, default = 0)
# Use the pre-calculated lm intercept and slope from group_data
# Z_Shift and lm values
Z_Shift_value <- round(first(group_data[[paste0("Z_Shift_", var)]], default = 0), 2)
Z_lm_value <- round(first(group_data[[paste0("Z_lm_", var)]], default = 0), 2)
R_squared_value <- round(first(group_data[[paste0("R_Squared_", var)]], default = 0), 2)
# NG, DB, SM values
NG_value <- first(group_data$NG, default = 0)
DB_value <- first(group_data$DB, default = 0)
SM_value <- first(group_data$SM, default = 0)
# Use the pre-calculated lm intercept and slope from the dataframe
lm_intercept_col <- paste0("lm_intercept_", var)
lm_slope_col <- paste0("lm_slope_", var)
lm_intercept_value <- first(group_data[[lm_intercept_col]], default = 0)
lm_slope_value <- first(group_data[[lm_slope_col]], default = 0)
plot_config <- list(
df = group_data,
plot_type = "scatter",
@@ -935,9 +969,8 @@ generate_interaction_plot_configs <- function(df, type) {
),
error_bar = TRUE,
error_bar_params = list(
# Passing constants directly
ymin = -2 * WT_sd_value,
ymax = 2 * WT_sd_value,
ymin = 0 - (2 * WT_sd_value),
ymax = 0 + (2 * WT_sd_value),
color = "black"
),
smooth = TRUE,
@@ -952,45 +985,43 @@ generate_interaction_plot_configs <- function(df, type) {
delta_plot_configs <- append(delta_plot_configs, list(plot_config))
}
}
# Calculate dynamic grid layout
grid_ncol <- 4
num_plots <- length(delta_plot_configs)
grid_nrow <- ceiling(num_plots / grid_ncol)
return(list(
list(grid_layout = list(ncol = 2, nrow = 2), plots = stats_plot_configs),
list(grid_layout = list(ncol = 2, nrow = 2), plots = stats_boxplot_configs),
list(grid_layout = list(ncol = grid_ncol, nrow = grid_nrow), plots = delta_plot_configs)
list(grid_layout = list(ncol = 4, nrow = grid_nrow), plots = delta_plot_configs)
))
}
generate_rank_plot_configs <- function(df_interactions, is_lm = FALSE, adjust = FALSE, overlap_color = FALSE) {
plot_configs <- list()
generate_rank_plot_configs <- function(df, is_lm = FALSE, adjust = FALSE, overlap_color = FALSE) {
sd_bands <- c(1, 2, 3)
plot_configs <- list()
# Adjust (if necessary) and rank columns
variables <- c("L", "K")
# Adjust (if necessary) and rank columns
for (variable in variables) {
if (adjust) {
df_interactions[[paste0("Avg_Zscore_", variable)]] <-
ifelse(is.na(df_interactions[[paste0("Avg_Zscore_", variable)]]), 0.001, df_interactions[[paste0("Avg_Zscore_", variable)]])
df_interactions[[paste0("Z_lm_", variable)]] <-
ifelse(is.na(df_interactions[[paste0("Z_lm_", variable)]]), 0.001, df_interactions[[paste0("Z_lm_", variable)]])
df[[paste0("Avg_Zscore_", variable)]] <- ifelse(is.na(df[[paste0("Avg_Zscore_", variable)]]), 0.001, df[[paste0("Avg_Zscore_", variable)]])
df[[paste0("Z_lm_", variable)]] <- ifelse(is.na(df[[paste0("Z_lm_", variable)]]), 0.001, df[[paste0("Z_lm_", variable)]])
}
df_interactions[[paste0("Rank_", variable)]] <- rank(df_interactions[[paste0("Avg_Zscore_", variable)]], na.last = "keep")
df_interactions[[paste0("Rank_lm_", variable)]] <- rank(df_interactions[[paste0("Z_lm_", variable)]], na.last = "keep")
df[[paste0("Rank_", variable)]] <- rank(df[[paste0("Avg_Zscore_", variable)]], na.last = "keep")
df[[paste0("Rank_lm_", variable)]] <- rank(df[[paste0("Z_lm_", variable)]], na.last = "keep")
}
# Helper function to create a plot configuration
create_plot_config <- function(variable, rank_var, zscore_var, y_label, sd_band, with_annotations = TRUE) {
num_enhancers <- sum(df_interactions[[zscore_var]] >= sd_band, na.rm = TRUE)
num_suppressors <- sum(df_interactions[[zscore_var]] <= -sd_band, na.rm = TRUE)
num_enhancers <- sum(df[[zscore_var]] >= sd_band, na.rm = TRUE)
num_suppressors <- sum(df[[zscore_var]] <= -sd_band, na.rm = TRUE)
# Default plot config
plot_config <- list(
df = df_interactions,
df = df,
x_var = rank_var,
y_var = zscore_var,
plot_type = "scatter",
@@ -1007,18 +1038,18 @@ generate_rank_plot_configs <- function(df_interactions, is_lm = FALSE, adjust =
x_label = "Rank",
legend_position = "none"
)
if (with_annotations) {
# Add specific annotations for plots with annotations
plot_config$annotations <- list(
list(
x = median(df_interactions[[rank_var]], na.rm = TRUE),
y = max(df_interactions[[zscore_var]], na.rm = TRUE) * 0.9,
x = median(df[[rank_var]], na.rm = TRUE),
y = max(df[[zscore_var]], na.rm = TRUE) * 0.9,
label = paste("Deletion Enhancers =", num_enhancers)
),
list(
x = median(df_interactions[[rank_var]], na.rm = TRUE),
y = min(df_interactions[[zscore_var]], na.rm = TRUE) * 0.9,
x = median(df[[rank_var]], na.rm = TRUE),
y = min(df[[zscore_var]], na.rm = TRUE) * 0.9,
label = paste("Deletion Suppressors =", num_suppressors)
)
)
@@ -1032,12 +1063,12 @@ generate_rank_plot_configs <- function(df_interactions, is_lm = FALSE, adjust =
rank_var <- if (is_lm) paste0("Rank_lm_", variable) else paste0("Rank_", variable)
zscore_var <- if (is_lm) paste0("Z_lm_", variable) else paste0("Avg_Zscore_", variable)
y_label <- if (is_lm) paste("Int Z score", variable) else paste("Avg Z score", variable)
# Loop through SD bands
for (sd_band in sd_bands) {
# Create plot with annotations
plot_configs[[length(plot_configs) + 1]] <- create_plot_config(variable, rank_var, zscore_var, y_label, sd_band, with_annotations = TRUE)
# Create plot without annotations
plot_configs[[length(plot_configs) + 1]] <- create_plot_config(variable, rank_var, zscore_var, y_label, sd_band, with_annotations = FALSE)
}
@@ -1051,7 +1082,7 @@ generate_rank_plot_configs <- function(df_interactions, is_lm = FALSE, adjust =
return(list(grid_layout = list(ncol = grid_ncol, nrow = grid_nrow), plots = plot_configs))
}
generate_correlation_plot_configs <- function(df_interactions) {
generate_correlation_plot_configs <- function(df, correlation_stats) {
# Define relationships for different-variable correlations
relationships <- list(
list(x = "L", y = "K"),
@@ -1062,23 +1093,6 @@ generate_correlation_plot_configs <- function(df_interactions) {
list(x = "r", y = "AUC")
)
correlation_stats <- list()
for (rel in relationships) {
x_var <- paste0("Z_lm_", rel$x)
y_var <- paste0("Z_lm_", rel$y)
lm_fit <- lm(df_interactions[[y_var]] ~ df_interactions[[x_var]])
intercept <- coef(lm_fit)[1]
slope <- coef(lm_fit)[2]
r_squared <- summary(lm_fit)$r.squared
relationship_name <- paste0(rel$x, "_vs_", rel$y)
correlation_stats[[relationship_name]] <- list(
intercept = intercept,
slope = slope,
r_squared = r_squared
)
}
plot_configs <- list()
# Iterate over the option to highlight cyan points (TRUE/FALSE)
@@ -1102,15 +1116,15 @@ generate_correlation_plot_configs <- function(df_interactions) {
# Construct plot config
plot_config <- list(
df = df_interactions,
df = df,
x_var = x_var,
y_var = y_var,
plot_type = "scatter",
title = plot_label,
annotations = list(
list(
x = mean(df_interactions[[x_var]], na.rm = TRUE),
y = mean(df_interactions[[y_var]], na.rm = TRUE),
x = mean(df[[x_var]], na.rm = TRUE),
y = mean(df[[y_var]], na.rm = TRUE),
label = paste("R-squared =", round(r_squared, 3))
)
),
@@ -1426,39 +1440,39 @@ main <- function() {
write.csv(df_calculations_reference, file = file.path(out_dir, "zscore_calculations_reference.csv"), row.names = FALSE)
write.csv(df_interactions_reference, file = file.path(out_dir, "zscore_interactions_reference.csv"), row.names = FALSE)
message("Setting missing deletion values to the highest theoretical value at each drug conc for L")
df_deletion <- df_na_stats %>% # formerly X2
filter(OrfRep != strain) %>%
filter(!is.na(L)) %>%
group_by(OrfRep, Gene, conc_num) %>%
mutate(
max_l_theoretical = max(max_L, na.rm = TRUE),
L = ifelse(L == 0 & !is.na(L) & conc_num > 0, max_l_theoretical, L),
SM = ifelse(L >= max_l_theoretical & !is.na(L) & conc_num > 0, 1, SM),
L = ifelse(L >= max_l_theoretical & !is.na(L) & conc_num > 0, max_l_theoretical, L)) %>%
ungroup()
# message("Setting missing deletion values to the highest theoretical value at each drug conc for L")
# df_deletion <- df_na_stats %>% # formerly X2
# filter(OrfRep != strain) %>%
# filter(!is.na(L)) %>%
# group_by(OrfRep, Gene, conc_num) %>%
# mutate(
# max_l_theoretical = max(max_L, na.rm = TRUE),
# L = ifelse(L == 0 & !is.na(L) & conc_num > 0, max_l_theoretical, L),
# SM = ifelse(L >= max_l_theoretical & !is.na(L) & conc_num > 0, 1, SM),
# L = ifelse(L >= max_l_theoretical & !is.na(L) & conc_num > 0, max_l_theoretical, L)) %>%
# ungroup()
message("Calculating deletion strain(s) summary statistics")
df_deletion_stats <- calculate_summary_stats(
df = df_deletion,
variables = c("L", "K", "r", "AUC"),
group_vars = c("OrfRep", "Gene", "Drug", "conc_num", "conc_num_factor_factor")
)$df_with_stats
message("Calculating deletion strain(s) interactions scores")
results <- calculate_interaction_scores(df_deletion_stats, df_bg_stats, group_vars = c("OrfRep", "Gene", "Drug"))
df_calculations <- results$calculations
df_interactions <- results$interactions
df_interactions_joined <- results$full_data
write.csv(df_calculations, file = file.path(out_dir, "zscore_calculations.csv"), row.names = FALSE)
write.csv(df_interactions, file = file.path(out_dir, "zscore_interactions.csv"), row.names = FALSE)
# message("Calculating deletion strain(s) summary statistics")
# df_deletion_stats <- calculate_summary_stats(
# df = df_deletion,
# variables = c("L", "K", "r", "AUC"),
# group_vars = c("OrfRep", "Gene", "Drug", "conc_num", "conc_num_factor_factor")
# )$df_with_stats
# message("Calculating deletion strain(s) interactions scores")
# results <- calculate_interaction_scores(df_deletion_stats, df_bg_stats, group_vars = c("OrfRep", "Gene", "Drug"))
# df_calculations <- results$calculations
# df_interactions <- results$interactions
# df_interactions_joined <- results$full_data
# write.csv(df_calculations, file = file.path(out_dir, "zscore_calculations.csv"), row.names = FALSE)
# write.csv(df_interactions, file = file.path(out_dir, "zscore_interactions.csv"), row.names = FALSE)
# Create interaction plots
message("Generating reference interaction plots")
reference_plot_configs <- generate_interaction_plot_configs(df_interactions_reference_joined, df_bg_stats, "reference")
reference_plot_configs <- generate_interaction_plot_configs(df_interactions_reference_joined, "reference")
generate_and_save_plots(out_dir, "interaction_plots_reference", reference_plot_configs)
message("Generating deletion interaction plots")
deletion_plot_configs <- generate_interaction_plot_configs(df_interactions_joined, df_bg_stats, "deletion")
deletion_plot_configs <- generate_interaction_plot_configs(df_interactions_joined, "deletion")
generate_and_save_plots(out_dir, "interaction_plots", deletion_plot_configs)
message("Writing enhancer/suppressor csv files")
@@ -1495,7 +1509,7 @@ main <- function() {
message("Generating rank plots")
rank_plot_configs <- generate_rank_plot_configs(
df_interactions,
df_interactions_joined,
is_lm = FALSE,
adjust = TRUE
)
@@ -1504,37 +1518,16 @@ main <- function() {
message("Generating ranked linear model plots")
rank_lm_plot_configs <- generate_rank_plot_configs(
df_interactions,
df_interactions_joined,
is_lm = TRUE,
adjust = TRUE
)
generate_and_save_plots(out_dir = out_dir, filename = "rank_plots_lm",
plot_configs = rank_lm_plot_configs)
overlap_threshold <- 2 # TODO add to study config?
df_interactions_filtered <- df_interactions %>%
filter(!is.na(Z_lm_L) & !is.na(Avg_Zscore_L)) %>%
mutate(
Overlap = case_when(
Z_lm_L >= overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Enhancer Both",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Suppressor Both",
Z_lm_L >= overlap_threshold & Avg_Zscore_L <= overlap_threshold ~ "Deletion Enhancer lm only",
Z_lm_L <= overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Enhancer Avg Zscore only",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L >= -overlap_threshold ~ "Deletion Suppressor lm only",
Z_lm_L >= -overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Suppressor Avg Zscore only",
Z_lm_L >= overlap_threshold & Avg_Zscore_L <= -overlap_threshold ~ "Deletion Enhancer lm, Deletion Suppressor Avg Z score",
Z_lm_L <= -overlap_threshold & Avg_Zscore_L >= overlap_threshold ~ "Deletion Suppressor lm, Deletion Enhancer Avg Z score",
TRUE ~ "No Effect"
),
lm_R_squared_L = summary(lm(Z_lm_L ~ Avg_Zscore_L))$r.squared,
lm_R_squared_K = summary(lm(Z_lm_K ~ Avg_Zscore_K))$r.squared,
lm_R_squared_r = summary(lm(Z_lm_r ~ Avg_Zscore_r))$r.squared,
lm_R_squared_AUC = summary(lm(Z_lm_AUC ~ Avg_Zscore_AUC))$r.squared
)
message("Generating filtered ranked plots")
rank_plot_filtered_configs <- generate_rank_plot_configs(
df_interactions_filtered,
df_interactions_joined,
is_lm = FALSE,
adjust = FALSE,
overlap_color = TRUE
@@ -1546,7 +1539,7 @@ main <- function() {
message("Generating filtered ranked linear model plots")
rank_plot_lm_filtered_configs <- generate_rank_plot_configs(
df_interactions_filtered,
df_interactions_joined,
is_lm = TRUE,
adjust = FALSE,
overlap_color = TRUE
@@ -1558,7 +1551,7 @@ main <- function() {
message("Generating correlation curve parameter pair plots")
correlation_plot_configs <- generate_correlation_plot_configs(
df_interactions_filtered
df_interactions_joined
)
generate_and_save_plots(
out_dir = out_dir,