Error in run_tSNE in giotto clustering

[Aditya-Venkat]

Hey there !

I'm using Giotto to run spatial clustering on a dataset, but I run into the following error :

ST <- runtSNE(ST, dimensions_to_use = 1:20)
Error in Rtsne.default(X = matrix_to_use, dims = dims, perplexity = perplexity, :
Remove duplicates before running TSNE.

ST is the Giotto object after running PCA and UMAP on it. I have done checks from my side for duplicates in spatial_locs, cell/geneIDs in raw expressions and in the pca/umap matrix, but there are no duplicates.

Please help me resolve this issue.

Thanks!

[RubD]

@Aditya-Venkat can you provide a minimal and reproducible example, otherwise it's impossible for us to figure out what went wrong. Thanks

[Aditya-Venkat]

Sure, this is the code snippet of the portion of the code leading up to the error.

library(Giotto)
library(DescTools)
library(anndata)

ST$var_names_make_unique()
ST$obs_names_make_unique()
st_adata <- t(as.matrix(ST$X))

x<- ST$obsm$spatial[,1]
y<- ST$obsm$spatial[,2]
spatial_location_anndata<- data.frame(x=x, y=y)

row.names(spatial_location_anndata)<-colnames(st_adata)
my_giotto_object = createGiottoObject(raw_exprs = st_adata,
spatial_locs = spatial_location_anndata)
visium_brain <- filterGiotto(gobject = my_giotto_object,
expression_threshold = 1,
min_det_genes_per_cell = 10,
expression_values = c('raw'),
verbose = T)
visium_brain <- normalizeGiotto(gobject = my_giotto_object, scalefactor = 6000, verbose = T)

visium_brain <- calculateHVG(gobject = visium

visium_brain <- runPCA(gobject = visium_brain,
scale_unit = T,
center=T,
method="factominer")

visium_brain <- runUMAP(visium_brain, dimensions_to_use = 1:5)
visium_brain <- runtSNE(visium_brain, dimensions_to_use = 1:5)

I am also attaching a dataset this error is coming up on.
MERFISH_0.04.zip