Plot examples from the CRANT connectome

Load packages

library(crantr)
library(ggplot2)
library(dplyr)
library(hemibrainr)

---

1 — Get CRANTb meta data

Read CRANT meta data from the seatable database (not yet public access, enquire by email for access)

crant.meta <- crant_table_query()

---

2 — Choose olfactory projection neurons

Subset the meta data to olfactory projection neurons

crant.meta.alpns <- crant.meta %>%
  dplyr::filter(cell_class == "olfactory_projection_neuron", 
                side = "right",
                proofread) %>%
  dplyr::distinct(root_id, .keep_all = TRUE)
crant.meta.alpns.ids <- na.omit(crant.meta.alpns$root_id)

---

3 — Get neuron skeletons with synapses

Now we want to read CRANTb skeletons, re-root them, add synapse and use the flow centrality algorithm to make an axon-dendrite split

# Get the L2 skeletons
# crant.skels <- crant_read_l2skel(crant.meta.alpns.ids)

# Gethigh resolution skeletons
choose_crant()
crant.skels <- skeletor(segment = crant.meta.alpns.ids, 
                   clean = TRUE,
                   method = "wavefront",
                   save.obj = NULL, 
                   mesh3d = FALSE,
                   waves = 1,
                   k.soma.search = 100,
                   radius.soma.search = 10000,
                   heal = TRUE,
                   reroot = TRUE,
                   heal.threshold = 1000000,
                   heal.k = 10L,
                   reroot_method = "density",
                   brain = bancr::banc_neuropil.surf,
                   elapsed = 10000,
                   resample = 1000)

# Add synapse information, stored at n.syn[[1]]$connectors
crant.skels.syns <- crant_add_synapses(crant.skels)

# Split neuron
crant.neurons.flow <- hemibrainr::flow_centrality(crant.skels.syns)

---

4 — Get neuron meshes and plot our neuroanatomical data

Lets see this data together with the segmentation mesh. First, let’s get those meshes.

# read meshes
crant.meshes <- crant_read_neuron_meshes(crant.meta.alpns.ids)

# High res skeletons
# crant.skels <- fafbseg::skeletor(crant.meta.alpns.ids, OmitFailures = TRUE)

And now let’s make a pretty ggplot2 plot.

# read meshes
g <- nat.ggplot::gganat +
  nat.ggplot::geom_neuron(
    crantb.surf,
    rotation_matrix = crantr:::crant_rotation_matrices[["front"]],
    cols = c("grey95", "grey85"),
    alpha = 0.3
  ) +
  nat.ggplot::geom_neuron(
    crant.meshes,
    rotation_matrix = crantr:::crant_rotation_matrices[["front"]],
    cols = c("grey60", "grey40"),
    alpha = 0.8
  ) +
  nat.ggplot::geom_neuron(
    crant.neurons.flow,
    threshold = 15000,
    root = 2,
    size = 0.1,
    rotation_matrix = crantr:::crant_rotation_matrices[["front"]]
  ) +
  ggplot2::labs(
    title = "antennal lobe projection neurons",
    subtitle = "proofread, axon-dendrite split, CRANTb data"
  )

# Show
print(g)

# Save
ggsave(g,
       filename = "inst/crantb_antennal_lobe_projection_neurons.png")