Closed Abnaveed closed 2 years ago
Hello Abdullah,
the package internally uses a conversion table for human/mouse orthologs stored at ProjecTILs::Hs2Mm.convert.table
. This table is derived from BioMart and should be fairly complete, but we know that some genes are missing. That is the case for your gene of interest KLRC1:
tab <- ProjecTILs::Hs2Mm.convert.table
tab[grep("^KLR",tab$Gene.HS),]
Gene.stable.ID.HS Gene.HS Gene.MM Alt.symbol Alt.symbol.HS
2876 ENSG00000134539 KLRD1 Klrd1 Klrd1 KLRD1
3103 ENSG00000183542 KLRC4 Clec2m Clec2m,Klre1,Klri1,Klri2 KLRC4,KLRC4-KLRK1,KLRC3,KLRC2,KLRC1,AC068775.1
3232 ENSG00000213809 KLRK1 Klrk1 Klrk1 KLRK1
6487 ENSG00000139187 KLRG1 Klrg1 Klrg1 KLRG1
8195 ENSG00000111796 KLRB1 Klrb1 Klrb1c,Klrb1,Klrb1a,Klrb1b,Klrb1f,Gm44511 KLRB1
11877 ENSG00000188883 KLRG2 Klrg2 Klrg2 KLRG2
I have added a new parameter to the function make.projection
(in the development version of ProjecTILs, to be installed with remotes::install_github("carmonalab/ProjecTILs", ref="dev")
) that should allow you to specify a custom table of orthologs. For example, you could add KLRC1 to the existing table with:
mytab <- rbind(ProjecTILs::Hs2Mm.convert.table, c("ENSG00000134545","KLRC1","Klrc1","Klrc1","KLRC1"))
and then specify this new table when you run the projection:
query.projected <- make.projection(query.seurat, ref=ref, ortholog_table=mytab)
Let us know if this works for you!
-m
Hey
Whenever I run ProjecTILS on a Seurat object with my T cells, the resulting file loses KLRC1 signal. It says there is no expression of it in any assay. I know from feature plots and gene expression data that it is highly expressed in the cluster subset I am projecting on ProjecTILS.
I am assuming this has something to do with how gene names are being converted. Is there a way to fix this ?
Thank you