Hello everyone,
I'm trying integrative analysis of two types 10x scRNA seq data, GEX and Cite-seq (ADT) using Seurat .
These 2 types data each contains different cell number because GEX data was "filtered" form as a consequence of the mapping and quality control.
I would like to use cells present in both ADT and GEX for analysis and integrate data from both. Is this the correct way?
In my mind, I suppose the ffollowing command is consistent with the purpose of my analysis, but of course it does mot work because number of columns (= cell barcodes) are different between each other.
sample1_GEX<-Read10X(data.dir = "")
sample1_GEX_seurat<-CreateSeuratObject(counts=sample1_GEX)
sample1_ADT<-Read10X(data.dir = "")
sample1_GEX_seurat[['Protein']] = CreateAssayObject(counts = sample1_ADT$Antibody Capture)
I am a beginner on this field, so I am afraid this is a very basic question.. , but I would appreciate it if you could teach me if there is a way to extract and analyze cells that are common (=having the same cell barcode) between GEX and ADT data.
Best regards,
Is this CITE-seq from the same exact cells that were assayed for gene expression or do you try to somehow squeeze in CITE-seq from another study? If from the exact same run then tools like CellRanger will produce one count matrix where CITE-seq counts are appended on the bottom, so you don't need to do any matching of cells/barcodes.
Thank you very much for your kind advice.
I apologize for my lack of explanation.
We separately prepared two types of library on the same sample; one is for GEX, the other is for ADT assay.
As you mentioned, ADT assay includes both "gene expression" and "antibody capture" data.
However, I want to use GEX data as gene expression data because we have already conducted an alignment procedure for the GEX data using customized reference genome.
I think the proper way is conducting alignment for ADT data, but it's difficult due to technical problems...
We conduct mapping using STARsolo, but as far as I know, there is no established way to properly map cite-seq data in this pipeline. (Cellranger cannot be used due to PC environment problems.)
If you have any suggestions, please let me know.
Thank you so much.
Essentially, what other tools like salmon-alevin do is to map GEX normally, using only the 'reliable' barcodes/cells. For CITE-seq you map against the feature barcode reference and you return all detected barcodes, regardless of reliability. Then you take the intersect. It's really that simple. With STARsolo that should work just the same. Does that make sense? I used that strategy before and it works well. Probably CellRanger does something similar. Consider using Docker or a VM if you want to give CellRanger a try.