Making an alluvial plot in R with ggalluvial

alluvial.R

library("tidyverse")
library("ggalluvial")

# original data
# every genomic bin of 200bp is assigned one of 8 possible chromatin states
# note that group1-3 are factors
input_data
# A tibble: 855,156 x 4
#    bin            group1 group2 group3
#    <chr>          <fct>  <fct>  <fct> 
#  1 chrX:0-200     8      8      8     
#  2 chrX:200-400   8      8      8     
#  3 chrX:400-600   8      8      8     
#  4 chrX:600-800   8      8      8     
#  5 chrX:800-1000  8      8      8     
#  6 chrX:1000-1200 8      8      8     
#  7 chrX:1200-1400 8      8      8     
#  8 chrX:1400-1600 8      8      8     
#  9 chrX:1600-1800 8      8      8     
# 10 chrX:1800-2000 8      8      8     
# … with 855,146 more rows

# group counts by experimental groups
input_data_grouped <- input_data %>%
  group_by(group1, group2, group3) %>%
  tally()
# A tibble: 265 x 4
# Groups:   group1, group2 [58]
#    group1 group2 group3     n
#    <fct>  <fct>  <fct>  <int>
#  1 1      1      1        703
#  2 1      1      2         21
#  3 1      1      3         14
#  4 1      1      4        175
#  5 1      1      5          4
#  6 1      1      6          1
#  7 1      1      8         14
#  8 1      2      1         12
#  9 1      2      2          2
# 10 1      2      3          4
# … with 255 more rows

# transform data into alluvial shape using the `to_lodes_form()` function
transition_lodes <- to_lodes_form(input_data_grouped, key = "test", axes = 1:3)
# A tibble: 795 x 4
#        n alluvium test   stratum
#    <int>    <int> <fct>  <fct>  
#  1   703        1 group1 1      
#  2    21        2 group1 1      
#  3    14        3 group1 1      
#  4   175        4 group1 1      
#  5     4        5 group1 1      
#  6     1        6 group1 1      
#  7    14        7 group1 1      
#  8    12        8 group1 1      
#  9     2        9 group1 1      
# 10     4       10 group1 1      
# … with 785 more rows

# plot
p <- ggplot(
    transition_lodes,
    aes(
      x = test,
      stratum = stratum,
      alluvium = alluvium,
      y = n
    )
  ) +
  geom_alluvium(aes(fill = stratum)) +
  geom_stratum(aes(fill = stratum)) +
  geom_label(stat = "stratum") +
  theme_bw() +
  labs(x = "", y = "") +
  theme_bw() +
  theme(
    legend.position = "none",
    axis.title = element_blank(),
    axis.text.x = element_text(face = "bold", colour = "black", size=15),
    axis.text.y = element_blank(),
    axis.ticks.y = element_blank(),
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    panel.border = element_blank(),
    axis.ticks.x = element_blank()
  )

The code above will result in this plot:

Even though it gives me a warning. I might have to revise this code in the future.

Warning message:
In f(...) :
  Some differentiation aesthetics vary within alluvia, and will be diffused by their first value.
Consider using `geom_flow()` instead.