search

RausellLab / CelliD

Gene signature extraction and cell-identity recognition at individual cell level from single-cell RNA-seq.
GNU General Public License v3.0
87 stars 19 forks source link

RunCellHGT() error when annotating 16,000 cells from different tissues #13

Open ivy-yuan opened 2 years ago

ivy-yuan commented 2 years ago

I selected about 16,000 cells from 9 tissues of mouse and wanted to annotate with Cell ID. I followed the example code but got an error in RunCellHGT() function with default parameter. Here is what the error saying Error in RunCellHGT.Seurat(expr_data, pathways = all_gs) : All pathways have less than 10 features in common with the data. What's the problem? Could anyone help me with this issue?

Cortalak commented 2 years ago

Hi,

I do believe that as your are using mouse dataset your data has a gene name that is proper to mouse but the signature that your are using has the human notation, and hence the program think that you have supplied gene signatures that has no equivalent in your actual data.

Cortalak commented 2 years ago

try to compare the gene notation in your data rownames(your_seurat_data) against your signature and see if it matches.

Cortalak commented 2 years ago

Note that if your are using gene symbols, human will have the gene names all in capital letter whereas mouse will only have the first letter.

Cortalak commented 2 years ago

If that is the case you have to find signatures that has mouse gene nomenclature or you can try to convert your gene signatures. The proper way to convert is to use orthologs with package like biomart but you can also just try to transform the signatures from full capital letter to only first letter capital.