How to identify cell identities in mixed scRNA-seq using mutational signatures from bulk RNA-seq

Question

[Chilly]

Hi folks,

I have two coral individuals of the exact same species. I dissociated their cells, mixed them together and performed 10x chromium scRNA-seq (all in one library). I then performed bulk RNA-seq on both individuals separately.

How can I find mutational signatures for each coral individual (I am initially considering only SNPs) through bulk RNA-seq datasets and use these mutational signatures to identify the cell identity in the mixed scRNA-seq dataset? I want to distinguish the cells from the two sources, just like performing scRNA-seq on two libraries separately.

I know how to find mutations from single cells, such as SCmut and SComatic. I also know how to find mutations from bulk RNA-seq, such as Mutect2. But I have no idea how to combine the two.

Does anyone know if anyone has published a similar study? Or does anyone have a clear solution?

Thanks in advance.

Regards,
Dee

Answer 1

[jared.andrews07]

Call mutations in your bulk RNA-seq data for each sample. Find a few mutations that distinguish the individuals, preferably in a broadly and relatively highly expressed gene. Genotype those specific mutations in your scRNA-seq data via vartrix. Then visualize or whatnot based on genotype in your scRNA-seq data, and it should be pretty clear.

I've got a bit more detailed tutorial here.

[Chilly]

Hi Jared,

Thx for your reply. Awesome, that's exactly what I need. Is there any easy and direct way to generate accurate vcfs for SNPs from bulk RNA-seq?

[jared.andrews07]

I'd take a look at the GATK Best Practices Workflow for RNA-seq and go from there.

[Chilly]

Great, that's my plan too. Looks like I've chosen the best direction.