manuscript revision

.gitignore

@@ -11,5 +11,6 @@ output*
 *.Rhistory
 *.out
 *.docx
+*.png
 *.sif
 *.DS_Store

Snakefile

@@ -10,19 +10,17 @@ include: "rules/final_imputation_report.rule"
 
 
 
-rule create_manuscript_v1_1:
+rule manuscript:
     input:
         pop_report = rules.prepare_data_v1_1.output,
         posteriors = rules.fit_models_validation_v1_1.input,
         markdown   = "manuscript/manuscript.Rmd"
     output:
-        pdf     = "output/v1.1/manuscript.docx",
-        figures = "output/v1.1/manuscript_figures.zip"
+        "output/v1.1/manuscript.docx"
     shell:
         """
         mkdir -p output/v1.1
-        Rscript -e "rmarkdown::render(\\"{input.markdown}\\", output_dir = \\"output/v1.1\\", params = list(data_dir = \\"../output/v1.1/data\\", config_file = \\"../config.yml\\"))"
-        mv manuscript/figures.zip {output.figures}
+        Rscript -e "rmarkdown::render('{input.markdown}', output_dir = 'output/v1.1', params = list(data_dir = '../output/v1.1/data', config_file = '../config.yml'))"
         """
 
 

manuscript/manuscript.Rmd

@@ -600,7 +600,7 @@ df_plot <- rbind(
    
 ggplot(df_plot, aes(x = model, y = mean_error)) + 
   geom_hline(yintercept = 0, color = "black") +
-  geom_point(aes(color = label), size = .8, position = position_dodge(.3)) +
+  geom_point(aes(color = label), size = .8, position = position_dodge(.4)) +
   geom_errorbar(
     aes(ymin = mean_error - 1.96*se_error, ymax = mean_error + 1.96*se_error, color = label),
     width = .25, position = position_dodge(.3)
@@ -616,6 +616,9 @@ ggplot(df_plot, aes(x = model, y = mean_error)) +
     axis.text.x        = element_text(angle = 66, hjust = 1),
     legend.position    = "bottom" 
   )
+
+ggsave(filename = "overall_locf.pdf", width = 6, height = 3.5)
+ggsave(filename = "overall_locf.png", width = 6, height = 3.5, dpi = 300)
 ```
 
 Firstly, LOCF is overall negatively biased, i.e., 
@@ -673,7 +676,7 @@ plot_confusion_matrices <- function(df_predictions, models, nrow = 2, legendpos,
     group_by(fold, model) %>% 
     do(
       confusion_matrix = caret::confusionMatrix(
-          data = factor(.$prediction, levels = 3:8), 
+          data      = factor(.$prediction, levels = 3:8), 
           reference = factor(.$GOSE, levels = 3:8)
         ) %>% 
         as.matrix %>% as_tibble %>% 
@@ -688,24 +691,21 @@ plot_confusion_matrices <- function(df_predictions, models, nrow = 2, legendpos,
   
   # p_cnf_mtrx_raw <- 
   df_average_confusion_matrices %>%
-    ggplot(aes(`Observed GOSE`, `Predicted GOSE`, fill = n)) +
-      geom_raster(fill = "white") +
-      geom_text(aes(
-          label = sprintf("%3.f", n) %>% 
-            ifelse(. == "0.0", "", .),
-          color = n
+    ggplot(aes(`Observed GOSE`, `Predicted GOSE`)) +
+      geom_text(
+        aes(
+          label    = sprintf("%3.f", n) %>% ifelse(. == "0.0", "", .),
+          fontface = ifelse(`Observed GOSE` == `Predicted GOSE`, 'bold', 'plain')
         ), 
-        size = scriptsize
+        vjust = .5, size = scriptsize
       ) +
       geom_hline(yintercept = c(2, 4, 6) + .5, color = "black") +
       geom_vline(xintercept = c(2, 4, 6) + .5, color = "black") +
-      scale_color_gradient(low = "#999999", high = "black") +
-      #scale_fill_gradient(low = "white", high = "#555555") +
-      coord_fixed(expand = FALSE) + 
+      coord_fixed(expand = FALSE, xlim = c(0.5, 6.5), ylim = c(0.5, 6.5)) + 
       labs(x = "observed GOSe", y = "imputed GOSe", fill = "") +
       theme_bw() +
       theme(
-        panel.grid = element_blank(),
+        panel.grid      = element_blank(),
         legend.position = "none"
       ) + 
       facet_wrap(~model, nrow = nrow) +
@@ -745,7 +745,7 @@ plot_confusion_matrices(
 )
 
 ggsave(filename = "confusion_matrices_locf.pdf", width = 6, height = 3)
-ggsave(filename = "confusion_matrices_locf.png", width = 6, height = 3)
+ggsave(filename = "confusion_matrices_locf.png", width = 6, height = 3, dpi = 300)
 ```
 
 The absolute-count confusion matrices show that most imputed values are 
@@ -800,7 +800,7 @@ plot_summary_measures_cond <- function(df_predictions, models, label) {
       geom_hline(yintercept = 0, color = "black") +
       geom_line(aes(y = mean), alpha = .5) + 
       geom_point(aes(y = mean), size = .3,
-        position = position_dodge(.2)) + 
+        position = position_dodge(.3)) + 
       geom_errorbar(aes(ymin = mean - 1.96*se, ymax = mean + 1.96*se), 
         width = .33,
         size = .5,
@@ -829,7 +829,7 @@ plot_summary_measures_cond(
 )
 
 ggsave(filename = "errors_stratified_locf.pdf", width = 6, height = 3.5)
-ggsave(filename = "errors_stratified_locf.png", width = 6, height = 3.5, scale = 1.25)
+ggsave(filename = "errors_stratified_locf.png", width = 6, height = 3.5, scale = 1.25, dpi = 300)
 ```
 
 Just as with the overall performance, differences are most pronounced in terms
@@ -885,7 +885,7 @@ plot_confusion_matrices(
 )
 
 ggsave(filename = "confusion_matrices_all.pdf", width = 6, height = 3)
-ggsave(filename = "confusion_matrices_all.png", width = 6, height = 3)
+ggsave(filename = "confusion_matrices_all.png", width = 6, height = 3, dpi = 300)
 ```
 
 
@@ -898,7 +898,7 @@ plot_summary_measures_cond(
 )
 
 ggsave(filename = "imputation_error.pdf", width = 6, height = 3.5)
-ggsave(filename = "imputation_error.png", width = 6, height = 3.5, scale = 1.25)
+ggsave(filename = "imputation_error.png", width = 6, height = 3.5, scale = 1.25, dpi = 300)
 ```
 
 
@@ -1019,7 +1019,7 @@ df_ground_truth %>%
             panel.grid.major.x = element_blank()
         )
 ggsave(filename = "gose_marignal_per_fold.pdf", width = 6, height = 3)
-ggsave(filename = "gose_marignal_per_fold.png", width = 6, height = 3)
+ggsave(filename = "gose_marignal_per_fold.png", width = 6, height = 3, dpi = 300)
 ```
 
 
@@ -1268,9 +1268,4 @@ in the README.md file of the GitLab repository.
 
 
 
-```{r zip-figures, include=FALSE}
-system("zip figures.zip *.png *.pdf")
-system("rm *.png *.pdf")
-```
-
 # References