Open XpelC opened 1 year ago
Amazing, the only thing left is to give the specific cluster identities. Tomorrow after you have done that we should meet.
I suggest to use SingleR on your clusters for a final confirmation.
Sperm cells are mixed with T cell clusters (cluster 23). Should I label the cluster as sperm or T cell?
Amazing, the only thing left is to give the specific cluster identities. Tomorrow after you have done that we should meet. I suggest to use SingleR on your clusters for a final confirmation.
Sperm cells are mixed with T cell clusters (cluster 23). Should I label the cluster as sperm or T cell?
Forget about pserm, we will drop them from the dataset.
For epithelial, just divide them in luminal and basal. There must be markers for that.
divide them in luminal and basal.
I'll search for it.
- Color the umap by cancer and normal sample.
Also it's important to do this
- FeaturePlot: Use the original umap (split by cell marker, and sample).
- Use features and the markers:
The marker for basal cell (KRT 17)
Also please refer to the original data (the basal cell is found in cluster 4,6,9,16)
Great, any marker of luminal?
Great, any marker of luminal?
If we treat luminal as epithelial, then EPCAM for luminal, KRT17 for basal
If we treat luminal as epithelial, then EPCAM for luminal
Fair enough. Let me know when do you want to meet. Please before also complete the colouring by benign/cancer sampels.
Fair enough. Let me know when do you want to meet. Please before also complete the colouring by benign/cancer sampels.
I'll work on it now
Fair enough. Let me know when do you want to meet. Please before also complete the colouring by benign/cancer sampels.
tumor vs normal sample. If it is possible, let's meet at 14:00.
reference =celldex::MonacoImmuneData()
counts |>
as.SingleCellExperiment() |>
# Annotate
logNormCounts(assay.type = "X") |>
SingleR(
ref = monocyte_reference,
assay.type.test=1,
labels = reference$label.fine
)
Please provide the seurat object with the macrocluster column in the metadata into the GP_tranfer
1) create a directory in /stornext/General/scratch/GP_Transfer/
2) copy the seurat object into that directory
Please provide the seurat object with the macrocluster column in the metadata into the GP_tranfer
done dir: /stornext/General/scratch/GP_Transfer/xinpu_chen
Please provide the seurat object with the macrocluster column in the metadata into the GP_tranfer
done dir: /stornext/General/scratch/GP_Transfer/xinpu_chen
Thanks @XpelC ,
I noticed you did not label all the macrocluster we discussed
integrated_sample_annotated |> distinct(macrocluster)
# A tibble: 5 × 1
macrocluster
<chr>
1 epithelial
2 endothelial
3 fibroblast
4 other
5 T
Please modify referring to this (we should have 7 macroclusters as we have 7 circles in the image)
Please modify referring to this (we should have 7 macroclusters as we have 7 circles in the image)
done
Please modify referring to this (we should have 7 macroclusters as we have 7 circles in the image)
done
The file
/stornext/General/scratch/GP_Transfer/xinpu_chen/integrated_sample.rds
seems to be unchanged.
Could you please give me the path of your seurat object and execute this on your own file
integrated_sample_annotated <- readRDS("integrated_sample_annotated.rds")
integrated_sample_annotated |>
distinct(macrocluster)
Could you please give me the path of your seurat object and execute this on your own file
path: /stornext/General/scratch/GP_Transfer/xinpu_chen/integrated_sample.rds This time it should work.
Hello Xinpu,
the macro cluster distribution seems a bit off, can you confirm that you meant this subdivision?
the macro cluster distribution seems a bit off, can you confirm that you meant this subdivision?
This part has been fixed
Thanks, that's not the only problematic part. Also epithelial are shared between two macroclusters.
I think I’m able to annotate the immune cell. I’ll look through the epithelial dataset and let you know shortly.
Best wishes, Xinpu
On Sep 21, 2022, at 9:59 AM, Stefano Mangiola @.**@.>> wrote:
Great
other_ cell
Are you able to annotate Immune clusters?
epithelial
Would you be able to color by
— Reply to this email directly, view it on GitHubhttps://github.com/stemangiola/cellsig/issues/69#issuecomment-1253031845, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ALBA442DONM7VMMNX2I4IXTV7JFVLANCNFSM6AAAAAAQK6OQTQ. You are receiving this because you were mentioned.Message ID: @.***>
Thanks, please send back the fixed macrocluster UMAP first. We want to do step by step, to not redo any work. Please reply to this message with the UMAP image.
Thanks.
Thanks, please send back the fixed macrocluster UMAP first. We want to do step by step, to not redo any work. Please reply to this message with the UMAP image.
reference =celldex::MonacoImmuneData()
The singleR result. Should I left join it with the cell UMI in the original seurat object?
reference =celldex::MonacoImmuneData()
The singleR result. Should I left join it with the cell UMI in the original seurat object?
Yes, this is an evidence that you should match with manual curation. Automatic classification is not perfect but helps.
You can choose between: going on the manual curation, or integrate the other datasets that we collected first.
singleR result
singleR result.
I'll try to produce macroclusters basing on the singleR result now.
cell type name formatting The umap of the new integrated data
The umap of separated datasets
new cell types in each cluster seurat_clusters cell_type n
Clean the dataset
cell type name formatting
[x] Two columns: one column is original cell type names, the other column is the formatted names.
[x] These name should be lower case, with no space, singular (eliminate inconsistency)
[x] Left join the table with the data.
Check the dataset:
Cluster the cells separately [NOT NEEDED ANYMORE]
cell type decision
https://satijalab.org/seurat/articles/pbmc3k_tutorial.html
Sanity check
double check: