Adding Spatial Centroids to Existing Object

[n318162971]

Hello,

I have an existing Seurat v5 object containing RNA counts, cell annotations etc.
Seperately, I have the centroid coordinates for the same cells.
How am I to add those coordinates to the object so that I may apply spatial-oriented methods to it?

Thanks in advance

[corrinenief]

You can use the Seurat built in functions. You need to create a "centroid" object, assign the centroid info to an "FOV" object, then store in your seurat object. Here is my code that works for nanostring cosmx data

library(Seurat)
options(Seurat.object.assay.version = "v5")

#an RDS file from a nanostring TMA without spatial data included
myobject = readRDS("~/Documents/COSMX_TMA-seuratObject.RDS")

## add cell segmentation coordinates if you have it (usually stored in polygon-coords.csv for nanostring), otherwise you can skip
segs = read.csv("~/Documents/COSMX_TMA-polygons-coords.csv")
segs <- as.data.frame(segs)
segs<- CreateSegmentation(segs)

#use the x y and cell ID from metadata in the seurat object
cents <- data.frame(x = [email protected][["x_slide_mm"]],y = [email protected][["y_slide_mm"]],cell = [email protected][["cell_id"]])
cents <- CreateCentroids(cents, nsides = Inf, radius = 3)

mycoords <- list("segmentation" = segs,"centroids" = cents)
#mycoords <- list("Centroids" = cents) #if skipping segmentation data

myimg <- CreateFOV(mycoords, type = c("segmentation","centroids"),
                  molecules = NULL,
                  assay = "Spatial",
                  key = "RNA_",
                  name = "uterus")

#pick a name for your image, here I am using 'Uterus'
myobject@images$Uterus = myimg

This post also has an example with vizgen data
https://divingintogeneticsandgenomics.com/post/how-to-construct-a-spatial-object-in-seurat/