scRNAseq: Split FASTQ into cell specific FASTQ Files

Question

[niklas.lang]

Hi ! I did scRNAseq (10XChromium) of cancer cells and now I am looking for a simple way to split my FASTQ Files into celll specific FASTQ Files. Alternativley split my BAM File into cell specific BAM Files - because I need one FASTQ (BAM) File per cell in order to do perform variant calling on the single cell level. My goal is to obtain one single VCF file per cell.

Is there a build-in function in CellRanger for that? Are there any other tools to demultiplex my FASTQ/BAM Files using cell barcodes and split them into cell specific FASTQ/BAM Files?

Any help is highly appreciated! Thank you so much!

[Devon Ryan]

It is incredibly unlikely that this will prove useful. 10X reads will be enriched around the polyA site and be highly redundant.

[niklas.lang]

My idea was to split the BAM File into several cell-specific BAM File by using the "CB" Tag. I don't know how this could possibly interfer the PolyA enrichment, so I would be glad to hear what your concerns are?

[Devon Ryan]

The question is whether single cell variant calling with this type of data is likely to be useful. The answer is "no", because you're mostly sequencing around the TES.

[Friederike]

...meaning you're limited to detecting (exonic) variants close to the TES

Answer 1

[igor]

If you want to call variants from 10x data, there is a specialized tool by 10x that is optimized for their data and uses the combined BAM file (no need to modify your BAM file): https://github.com/10XGenomics/vartrix

If you are still interested in splitting a BAM file by barcode, there are some approaches outlined in this blog post: https://divingintogeneticsandgenomics.rbind.io/post/split-a-10xscatac-bam-file-by-cluster/

Answer 2

[swbarnes2]

You could split on the CB tag, but coverage in single cell datasets is so poor, you aren't going to find high quality variants.