R ggplot2 with stacked column instead of grouped

Question

I want to plot the data shown below as a grouped bar_plot.

I tried position = "dodge" or position = "dodge2" but it didn't work. Ι also tried position = position_dodge()

It kinda works if i use geom_bar instead of geom_col and remove the y=overlap_percent:

p3 <- ggplot(data = comp_coors, aes(x = species_code, fill = mirna_form)) + 
  geom_bar(position = "dodge2") + theme_classic()

p3

but i would like the y_axis to have the overlap_percent.

Another attempt which ends in a stacked barplot is:

p2 <- ggplot(data = comp_coors, aes(x = species_code, y = overlap_percent, fill = mirna_form)) + 
  geom_bar(stat = "identity") + theme_classic()

p2

Finally by using geom_col, it returns this which it doesn't make sense, at least to me:

p4 <- ggplot(data = comp_coors, aes(x = species_code, y = overlap_percent, fill = mirna_form)) + 
  geom_col(position = "dodge") + theme_classic()

p4

The data that i want to plot :

comp_coors <- data.table( species = c("aae","cel", "dme","hsa", "mdo"),
mirna_form = c("mature", "precursor"),
overlap_percent = c(100.0, 100.0, 88.0, 95.5, 91.7, 100.0, 96.6, 98.4),
overlapping_attribute = c("ID=MIMAT0014285;Alias=MIMAT0014285", "ID=MI0000043;Alias=MI0000043;Name=cel-mir-72", "ID=MIMAT0000401;Alias=MIMAT0000401;Name=dme-miR-", "ID=MI0000791;Alias=MI0000791;Name=hsa-mir-383", "ID=MI0005331;Alias=MI0005331;Name=mdo-let-7g")
)

Using your data example, the `geom_col` method gives a graph (after renaming `species` with `species_code`) which does make sense to me. What exactly do you expect? — Bas, May 17 '22 at 11:15

score 1 · Accepted Answer · answered May 17 '22 at 11:23

Try using species as a factor and add stat = "identity" like this:

ggplot(data = comp_coors, aes(x = factor(species), y = overlap_percent, fill = mirna_form)) + 
  geom_bar(position = "dodge", stat = "identity") + theme_classic() + labs(x = "Species", y = "Overlap percent")

Output:

A grouped barplot with overlap_percent on y-axis right.

R ggplot2 with stacked column instead of grouped

1 Answers1