counting from a single BAM file with multiple samples

Question

[dr-device]

I've got a BAM file that contains RNA-seq data from about 20 samples. The samples are differentiated by one another via the use of a cell barcode - which is appended to the end of the readID on the first column via an underscore (for example, READ:ID:WITH:ABUNCH:OF:COLONS_GCCCTTAT).

Is there any way to use featureCounts or umi tools to generate a gene count matrix with each sample in a different column? I know umi tools can do that but I don't have any UMIs (unique molecular identifiers), just cell barcodes, and I don't see a way of using it without UMIs.

[Ram]

You mention cell barcodes. Is this an scRNA-seq BAM?

[dr-device]

Yes, its from a plate-based protocol but with literally only 20 cells. Im therefore planning on just using edgeR for my analysis rather than doing traditional single cell analyses.

[bk11]

You can do sth like that has been discussed here.

[LChart]

I think you can use UMI-tools to do this with something like --extract_method=regex --bc-pattern=(?P<cell_1>8)$ --ignore-umi. However I don't think --extract_method=regex is supported nativelyby umi-tools count.

Answer 1

[dr-device]

I ended up splitting the BAM file by cell barcode using my own custom script and then using featureCounts individually on each BAM file before combining the results into a count Matrix in R.