How to demultiplex single cell rna-seq based on SNP mutation?

Question

[ahmad mousavi]

Hello,

I have a 10x sample which composed of three pooled cell line, although I know about the hash tags and other methods, we just mixed them with any prior consideration.

As we know number of known SNP for each of these cell lines, I am curious to know is there any way to demultiplex them based on those SNP or not?

I want to catch cell ID of each cell line to embed them into Seurat object.

I have tried scSplit but I don't how to get the cell ID to import it in Seurat.

Thanks.

[ATpoint]

Be aware that 10X samples have on-target sequencing of only the 3'UTR, so going for mutations in noise and low-coverage scRNA-seq appears to me very difficult. Still, since different cell lines should harbor very different transcriptomes they should cluster much apart. Can't you simply make the standard clustering and then see whether you basically get three big clusters, each corresponding to one cellline? Checking then for obvious markers that make the celllines unique could do the trick, no? Hashing them would have been much easier though.

[ahmad mousavi]

Thanks for answer, anyway can we use scSplit to find relative genotype of each cell line based on common SNPs ?

About the low reads for scRNA-SEQ you are right, but at the moment we have no choice we should keep on these data, if I got it p roperly you mean we need to select each cell line based on specific markers expression? and then extract those cells from Seurat object?