What is the difference between `possorted_genome_bam.bam` in the output of `cellranger count` and `sample_alignments.bam` in the output of `cellranger multi`?

Question

[Dan]

Hello:

I am trying to run single-cell RNA-seq velocity analysis by velocyto run10x (https://velocyto.org/velocyto.py/tutorial/cli.html) using the output of cellranger multi with hashtag multiplex.

Could someone please let me know what the difference between "possorted_genome_bam.bam" in the output of cellranger count with no multiplex, and sample_alignments.bam in the output of cellranger multi with multiplex?

Can I use the sample_alignments.bam in the output of cellranger multi to run velocyto run10x?

Thanks a lot.

[kanika.151]

As far as I see from the tutorial, you just need to give the cellranger output folder and it will do the rest.

bamfile = os.path.join(samplefolder, "outs", "possorted_genome_bam.bam")

velocyto run10x needs the above bam file in the output folder so I think sample_alignments.bam file will not work.

But did you find the answer to this question:

Could someone please let me know what the difference between "possorted_genome_bam.bam" in the output of cellranger count with no multiplex, and sample_alignments.bam in the output of cellranger multi with multiplex?

Answer 1

[GenoMax]

https://www.10xgenomics.com/support/software/cell-ranger/latest/analysis/outputs/cr-outputs-bam has the explanation of the BAM files produced by 10x cellranger pipelines.

The cellranger count pipeline generates an indexed BAM file named possorted_genome_bam.bam. This file contains position-sorted reads aligned to the genome and transcriptome, along with unassigned reads. If Feature Barcode libraries are included in the analysis, the BAM file will also include both aligned and unaligned records for each library type (e.g., Antibody Capture, CRISPR Guide Capture, 3' Cell Multiplexing, Antigen Capture).

The cellranger multi pipeline produces two BAM files. It separates the reads into unassigned reads, which are stored in unassigned_alignments.bam, and reads that are assigned to samples, which are stored in sample_alignments.bam.

[kanika.151]

Thank you! I actually read that but I wanted to understand if I want to run mpileup on the output of cellranger I should use the sample_alignments.bam, right? We are trying to see if a particular variant is seen in a cell type.

[GenoMax]

I would think so. I assume you will be looking at the CB tags then.

You could also subset your BAM by cell barcode: Separate single cell BAM file by the cell barcode

[jared.andrews07]

Or use vartrix, which can be a bit simpler than dealing with the barcodes manually (and lets you carry mutation info into downstream analyses more easily). I've got a tutorial here.