I have barcodes from a CITE-seq experiment that I would like to assign cell labels to. To accomplish this task, I used the Monaco reference database as input into SingleR based on the RNA expression data.
I would now like to assign cell types to barcodes based on the ADT expression data. I recently read a paper by Stuart et al., 2019 that describes a method for transferring this type of information to other datasets using canonical correlation analysis (CCA), which is implemented in Seurat.
I have seen this integration method used to transfer cell annotations to scATAC data, but have not seen it extended to surface protein marker data. Does anyone have any experience using this method for ADT data that they could share or alternative approaches?
I am not sure what you're asking. The Monaco reference is not a CITE-seq dataset and generally you either infer celltypes by the surface antibodies, that is then basically using FACS on single-cell data, or you infer celltypes based on the gene expression data. So what exactly is the problem?
Actually, I do not know if I even need to do the cell typing separately for the antibody capture data.. I already assigned cell types to barcodes based on the gene expression data. The barcodes for the antibody capture data are the same, so I already have the information I want.