figures: fix for mmseqs2 upsetr plots using cowplot (issue #127)

workflow_figures/scripts/fig_mmseqs2.R

@@ -8,10 +8,11 @@ sink(file=file(snakemake@log[[1]], open="wt"), type=c("output", "message"))
 ############################## LIBS
 suppressMessages(library(testit)) # assertions
 suppressMessages(library(dplyr)) # tables
-suppressMessages(library(patchwork)) # combine plots
-suppressMessages(library(ggplotify)) # for as.ggplot
-# suppressMessages(library(grid)) # to combine upsetr plots
-# suppressMessages(library(gridExtra)) # to combine upsetr plots
+# suppressMessages(library(patchwork)) # combine plots
+# suppressMessages(library(ggplotify)) # for as.ggplot
+suppressMessages(library(grid)) # to combine upsetr plots
+suppressMessages(library(gridExtra)) # to combine upsetr plots
+suppressMessages(library(cowplot))
 
 # custom
 source(snakemake@params$const)
@@ -40,36 +41,36 @@ for(sname in names(snakemake@config$samples)){
 
 ############################## PLOT
 
-# individual plots
-FIGS <- list()
-for(sname in names(snakemake@config$samples)){
-    FIGS[[sname]] <-
-        as.ggplot(plot_mmseqs2_overlap(TABS[[sname]])) +
-        labs(title=SAMPLE_NAMES[sname]) +
-        theme(plot.title=element_text(size=16, face="bold"))
-}
-
-# combine plots
-FIGS_ALL <- Reduce("+", FIGS) + plot_layout(ncol=2) #+ plot_annotation(tag_levels='A')
-
-# PDF
-pdf(snakemake@output$pdf, width=20, height=12)
-print(FIGS_ALL)
-dev.off()
-
 # # individual plots
-# # solution for how to combine upsetR plots: https://stackoverflow.com/questions/57457651/upsetr-manually-order-set-intersections-to-align-multiple-upset-plots
 # FIGS <- list()
 # for(sname in names(snakemake@config$samples)){
-#     FIGS[[sname]] <- plot_mmseqs2_overlap(TABS[[sname]])
-#     print(FIGS[[sname]])
-#     grid.text(SAMPLE_NAMES[sname], x=0.05, y=0.95, gp=gpar(fontsize=16, fontface="bold"))
-#     grid.edit('arrange', name=sname)
-#     vp <- grid.grab()
-#     FIGS[[sname]] <- vp
+#     FIGS[[sname]] <-
+#         as.ggplot(plot_mmseqs2_overlap(TABS[[sname]])) +
+#         labs(title=SAMPLE_NAMES[sname]) +
+#         theme(plot.title=element_text(size=16, face="bold"))
 # }
 
+# # combine plots
+# FIGS_ALL <- Reduce("+", FIGS) + plot_layout(ncol=2) #+ plot_annotation(tag_levels='A')
+
 # # PDF
 # pdf(snakemake@output$pdf, width=20, height=12)
-# grid.arrange(grobs = FIGS, nrow=2)
-# dev.off()
\ No newline at end of file
+# print(FIGS_ALL)
+# dev.off()
+
+# individual plots
+FIGS <- list()
+for(sname in names(snakemake@config$samples)){
+    FIGS[[sname]] <- plot_mmseqs2_overlap(TABS[[sname]])
+    # NOTE: the barplot w/ set sizes is not meaningful, therefore plot only the intersection sizes and the matrix
+    FIGS[[sname]] <- cowplot::plot_grid(
+        FIGS[[sname]]$Main_bar, FIGS[[sname]]$Matrix,
+        nrow=2, align='hv', rel_heights = c(3,1), rel_widths = c(2,3),
+        labels=SAMPLE_NAMES[sname], label_fontface = "bold"
+    )
+}
+
+# PDF
+pdf(snakemake@output$pdf, width=20, height=12)
+grid.arrange(grobs = FIGS, nrow=2)
+dev.off()
\ No newline at end of file

workflow_report/envs/r.yaml

@@ -19,4 +19,5 @@ dependencies:
   - r-upsetr=1.4.0
   - r-viridis=0.5.1
   - r-patchwork=1.1.1
-  - r-ggplotify=0.0.5
\ No newline at end of file
+  - r-ggplotify=0.0.5
+  - r-cowplot
\ No newline at end of file

workflow_report/scripts/utils.R

@@ -656,9 +656,9 @@ plot_mmseqs2_overlap <- function(df, topN=20){
         decreasing=TRUE,
         # titles
         mainbar.y.label="Total number of proteins\nin clusters",
-        sets.x.label="Total number of proteins\nper assembly",
+        sets.x.label="Ignore this statistic:\nit is not meaningful",
         # text size: intersection size title, intersection size tick labels, set size title, set size tick labels, set names, numbers above bars
-        text.scale=c(2, 1.6, 1.8, 1.3, 2, 1),
+        text.scale=c(2, 2, 1.8, 1.3, 2, 1.3),
         # ratio between matrix plot and main bar plot
         mb.ratio=c(0.6,0.4),
         # highlight those w/ only one set