About normalized cross-connectivity (NCC) score

[makabakayu]

Hi,
I'm very interested in the way you calculate normalized cross-connectivity (NCC) score. Is there any code for this part? If not, would I like to ask how to count the number of neighbors between the same cells in seurat?

Thank you in advance

[jleechung]

Hi @makabakayu, thanks for your interest in BANKSY!

For code to compute NCC, @chousn or @vipulsinghal02 can help.

I'm not sure what you mean by counting the number of neighbors between the same cells - could you clarify?

To obtain the spatial neighbors BANKSY uses for each cell, we use k-nearest neighbors implemented with dbscan:

# load packages
library(Banksy)
library(dbscan)

# load example data
data(hippocampus)
locs <- hippocampus$locations

# fit k-nearest neighbors on spatial coordinates
k_geom <- 15
knn <- dbscan::kNN(locs, k=k_geom)

# matrix where each row gives the indices of the k-nearest spatial neighbors for each cell
knn$id