Renthal-Lab / harmonized_atlas

Code used to construct and annotate the harmonized reference atlas
3 stars 3 forks source link

DRG_nonneurons_release.rds integrated atlas #6

Open cecileherbermann opened 1 month ago

cecileherbermann commented 1 month ago

Hi,

Thank you for creating the harmonized cross-species cell atlases of trigeminal and dorsal root ganglia. I am trying to anchor new human data to the atlases and have downloaded the integrated atlases on https://painseq.shinyapps.io/harmonized_painseq_v1/, but the DRG_nonneurons_release.rds seems to be the non-integrated one (whereas the other atlases are integrated). Can you confirm that this is the wrong file indeed and would it be possible to upload the integrated version? Thank you very much in advance!

Kind regards,

Cecile

shamsbhuiyan commented 1 month ago

Hi Cecile,

Thank you for reaching out! Would you be able to share a screenshot of what the DRG non-neuronal atlas looks like on your end? When I download it and run the following command, I get the attached UMAP projection:

DimPlot(DRG_nn,raster=F)

image

If easier to share via email, please feel free to email me: sbhuiyan1@bwh.harvard.edu

Cheers, Shams

cecileherbermann commented 1 month ago

Hi Shams,

My apologies, I got confused about the lay-out of the Seurat object of DRG nonneurons since it is different from the other atlases. When running the line atlas <- subset(neuron_seurat, subset = Species %in% c("C. Macaque","R. Macaque","Human")) for human data anchoring, it gave some errors, e.g. Error: Cannot add graphs with unordered cells (this can be fixed by just emptying the graphs with atlas@graphs <- list()) and this error after scaling the atlas:

Centering and scaling data matrix
  |=================================================================================================================================================================================================================| 100%
Error in validObject(object = value) : 
  invalid class “Assay” object: meta.features' must have the same features as 'data' 

I thought this had to do with the data not being integrated, but now I realize that it probably has to do with subsetting the Seurat object, even though it worked for the other atlases. If you have a script successfully anchoring new human data to the nonneuronal DRG atlas, please let me know!

Kind regards,

Cecile

shamsbhuiyan commented 1 month ago

Hi Cecile! I think I know the cause of this error, but I want to double check two things: 1) Does the script anchor when you do not subset for the primate data? 2) Which Seurat version are you using?

Cheers, Shams