divergence regions when rooting

[rcannood]

What to do when rooting a trajectory with divergence regions? Should we remove the region and project the cell to the closest point in the trajectory? Or do we flip the divergence region and change the edges in the trajectory? We could make this a parameter of add_root().

Example code

library(tidyverse)
library(dyno)

cell_ids <- c("a", "b", "c", "d", "e", "f")
milestone_ids <- c("W", "X", "Y", "Z", "A")

milestone_network <- tribble(
  ~from, ~to, ~length, ~directed,
  "W", "X", 2, TRUE,
  "X", "Z", 4, TRUE,
  "X", "Y", 3, TRUE,
  "Z", "A", 5, TRUE
)

divergence_regions <- tribble(
  ~divergence_id, ~milestone_id, ~is_start,
  "XYZ", "X", TRUE,
  "XYZ", "Y", FALSE,
  "XYZ", "Z", FALSE
)

milestone_percentages <- tribble(
  ~cell_id, ~milestone_id, ~percentage,
  "a", "W", .9,
  "a", "X", .1,
  "b", "W", .2,
  "b", "X", .8,
  "c", "X", .8,
  "c", "Z", .2,
  "d", "X", .2,
  "d", "Y", .7,
  "d", "Z", .1,
  "e", "X", .3,
  "e", "Y", .2,
  "e", "Z", .5,
  "f", "Z", .8,
  "f", "A", .2
)

trajectory <- wrap_data(
  id = "test",
  cell_ids = cell_ids
) %>% add_trajectory(
  milestone_ids = milestone_ids,
  milestone_network = milestone_network,
  milestone_percentages = milestone_percentages,
  divergence_regions = divergence_regions
)

g1 <- plot_graph(trajectory, label_milestones = TRUE)

trajectory2 <- trajectory %>% add_root(root_milestone_id = "Y")
g2 <- plot_graph(trajectory2, label_milestones = TRUE)

patchwork::wrap_plots(g1, g2)