made plot size configurable (per insert correlation) and made correla…

…tion visible (per barcode correlation)

workflow/rules/statistic/correlation.smk

@@ -311,6 +311,9 @@ rule statistic_correlation_calculate:
             if "label_file" in config["experiments"][wc.project]
             else ""
         ),
+        plot_width=29,
+        plot_height=17,
+        legend_nrow=6,
     log:
         temp(
             "results/logs/statistic/correlation/calculate.{project}.{condition}.{config}.{assignment}.log"
@@ -323,6 +326,9 @@ rule statistic_correlation_calculate:
         --files {params.files} \
         --replicates {params.replicates} \
         --threshold {params.thresh} \
+        --plot_width {params.plot_width} \
+        --plot_height {params.plot_height} \
+        --legend_nrow {params.legend_nrow} \
         --outdir {params.outdir} &> {log}
         """
 

workflow/scripts/count/plot_perBCCounts_correlation.R

@@ -85,21 +85,21 @@ plot_correlations_dna <- function(data, plot_data, condition, r1, r2, name) {
     geom_point() +
     xlab(sprintf(paste("log2 Normalized DNA count per barcode,\n replicate", r1))) +
     ylab(sprintf(paste("log2 Normalized DNA count per barcode,\n replicate", r2))) +
-    geom_text(x = min + 0.5, y = max - 0.5, label = sprintf("   r = %.2f", cor(data$DNA_normalized_log2.x, data$DNA_normalized_log2.y, method = "pearson")), size = 10) +
-    geom_text(x = min + 0.5, y = max - 1.0, label = sprintf("rho = %.2f", cor(data$DNA_normalized.x, data$DNA_normalized.y, method = "spearman")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=1, label = sprintf("   r = %.2f", cor(data$DNA_normalized_log2.x, data$DNA_normalized_log2.y, method = "pearson")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=2.1, label = sprintf("rho = %.2f", cor(data$DNA_normalized.x, data$DNA_normalized.y, method = "spearman")), size = 10) +
     geom_abline(intercept = 0, slope = 1) +
     theme_classic(base_size = 30)
   return(dna_p)
 }
 plot_correlations_rna <- function(data, plot_data, condition, r1, r2, name) {
-  max <- max(data$`RNA_normalized.y_log2`)
-  min <- min(data$`RNA_normalized.x_log2`)
+  max <- max(data$`RNA_normalized_log2.y`)
+  min <- min(data$`RNA_normalized_log2.x`)
   rna_p <- ggplot(plot_data, aes(RNA_normalized_log2.x, RNA_normalized_log2.y)) +
     geom_point() +
     xlab(sprintf(paste("log2 Normalized RNA count per barcode,\n replicate", r1))) +
     ylab(sprintf(paste("log2 Normalized RNA count per barcode,\n replicate", r2))) +
-    geom_text(x = min + 0.5, y = max - 0.5, label = sprintf("   r = %.2f", cor(data$RNA_normalized_log2.x, data$RNA_normalized_log2.y, method = "pearson")), size = 10) +
-    geom_text(x = min + 0.5, y = max - 1.0, label = sprintf("rho = %.2f", cor(data$RNA_normalized.x, data$RNA_normalized.y, method = "spearman")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=1, label = sprintf("   r = %.2f", cor(data$RNA_normalized_log2.x, data$RNA_normalized_log2.y, method = "pearson")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=2.1, label = sprintf("rho = %.2f", cor(data$RNA_normalized.x, data$RNA_normalized.y, method = "spearman")), size = 10) +
     geom_abline(intercept = 0, slope = 1) +
     theme_classic(base_size = 30)
   return(rna_p)
@@ -111,8 +111,8 @@ plot_correlations_ratio <- function(data, plot_data, condition, r1, r2, name) {
     geom_point() +
     xlab(sprintf(paste("log2 RNA/DNA per barcode,\n replicate", r1))) +
     ylab(sprintf(paste("log2 RNA/DNA per barcode,\n replicate", r2))) +
-    geom_text(x = min + 0.5, y = max - 0.5, label = sprintf("   r = %.2f", cor(data$Ratio_log2.x, res$Ratio_log2.y, method = "pearson")), size = 10) +
-    geom_text(x = min + 0.5, y = max - 1.0, label = sprintf("rho = %.2f", cor(data$Ratio.x, data$Ratio.y, method = "spearman")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=1, label = sprintf("   r = %.2f", cor(data$Ratio_log2.x, res$Ratio_log2.y, method = "pearson")), size = 10) +
+    geom_text(x = -Inf, y = Inf, hjust=0, vjust=2.1, label = sprintf("rho = %.2f", cor(data$Ratio.x, data$Ratio.y, method = "spearman")), size = 10) +
     geom_abline(intercept = 0, slope = 1) +
     theme_classic(base_size = 30)
   return(ratio_p)
@@ -225,4 +225,5 @@ if (data %>% nrow() > 1) {
   writeCorrelationPlots(plots_correlations_rna, sprintf("%s_barcode_RNA_pairwise.png", outdir))
   writeCorrelationPlots(plots_correlations_ratio, sprintf("%s_barcode_Ratio_pairwise.png", outdir))
 
-}
+}
+

workflow/scripts/count/plot_perInsertCounts_correlation.R

@@ -31,6 +31,24 @@ option_list <- list(
         type = "integer",
         default = 10,
         help = "Number of required barcodes (default 10)"
+    ),
+        make_option(
+        c("-pw", "--plot_width"),
+        type = "integer",
+        default = 29,
+        help = "Width of the plots created by this script (default 29)"
+    ),
+    make_option(
+        c("-ph", "--plot_height"),
+        type = "integer",
+        default = 17,
+        help = "Height of the plots created by this script (default 17)"
+    ),
+    make_option(
+        c("-n", "--legend_nrow"),
+        type = "integer",
+        default = 6,
+        help = "Number of rows in the legend of the plots created by this script (default 6)"
     ),
     make_option(c("-o", "--outdir"),
                 type = "character",
@@ -72,7 +90,6 @@ if ("label" %in% names(opt)) {
     use_labels <- FALSE
 }
 
-
 # replicates and count files
 files <- strsplit(opt$files, ",")[[1]]
 replicates <- strsplit(opt$replicates, ",")[[1]]
@@ -85,13 +102,12 @@ data["Condition"] <- cond
 
 print(data)
 
-# pairwise comparison only if more than one replicate
 thresh <- opt$threshold
 
 plot_correlations_dna <- function(data, condition, r1, r2, name) {
     dna_p <-
         ggplot(data, aes(dna_normalized_log2.x, dna_normalized_log2.y)) +
-        geom_point(aes(colour = label.x), show.legend = TRUE) +
+        geom_point(aes(colour = label.x)) +
         xlim(-5, 5) +
         ylim(-5, 5) +
         xlab(sprintf(
@@ -126,13 +142,16 @@ plot_correlations_dna <- function(data, condition, r1, r2, name) {
             size = 10
         ) +
         geom_abline(intercept = 0, slope = 1) +
-        theme_classic(base_size = 30)
+        theme_classic(base_size = 30) + 
+        theme(legend.position="bottom") + # show legend below the plot
+        guides(fill=guide_legend(nrow=legend_nrow, byrow=TRUE)) + # show labels in rows
+        labs(color = "label\n") # legend name
     return(dna_p)
 }
 plot_correlations_rna <- function(data, condition, r1, r2, name) {
     rna_p <-
         ggplot(data, aes(rna_normalized_log2.x, rna_normalized_log2.y)) +
-        geom_point(aes(colour = label.x), show.legend = TRUE) +
+        geom_point(aes(colour = label.x)) +
         xlim(-5, 5) +
         ylim(-5, 5) +
         xlab(sprintf(
@@ -165,12 +184,15 @@ plot_correlations_rna <- function(data, condition, r1, r2, name) {
             size = 10
         ) +
         geom_abline(intercept = 0, slope = 1) +
-        theme_classic(base_size = 30)
+        theme_classic(base_size = 30) + 
+        theme(legend.position="bottom") + # show legend below the plot
+        guides(fill=guide_legend(nrow=legend_nrow, byrow=TRUE)) + # show labels in rows
+        labs(color = "label\n") # legend name
     return(rna_p)
 }
 plot_correlations_ratio <- function(data, condition, r1, r2, name) {
     ratio_p <- ggplot(data, aes(ratio_log2.x, ratio_log2.y)) +
-        geom_point(aes(colour = label.x), show.legend = TRUE) +
+        geom_point(aes(colour = label.x)) +
         xlim(-5, 5) +
         ylim(-5, 5) +
         xlab(sprintf(paste(
@@ -198,7 +220,10 @@ plot_correlations_ratio <- function(data, condition, r1, r2, name) {
             size = 10
         ) +
         geom_abline(intercept = 0, slope = 1) +
-        theme_classic(base_size = 30)
+        theme_classic(base_size = 30) + 
+        theme(legend.position="bottom") + # show legend below the plot
+        guides(fill=guide_legend(nrow=legend_nrow, byrow=TRUE)) + # show labels in rows
+        labs(color = "label\n") # legend name
     return(ratio_p)
 }
 
@@ -268,8 +293,8 @@ write_correlation_plots <- function(plots, name) {
 
     ggplot2::ggsave(name,
                     correlation_plots,
-                    width = 15,
-                    height = 10 * length(plots))
+                    width = plot_width,
+                    height = plot_height * length(plots))
 }
 
 write_correlation <- function(correlations, name) {
@@ -324,7 +349,7 @@ if (use_labels) {
         all$label <- "NA"
     }
 }
-
+# pairwise comparison only if more than one replicate
 if (data %>% nrow() > 1 && nrow(all) > 1) {
     print("Pairwise comparisons")
     # make pairwise combinations