Bam File Barcodes and UMIs not Formatted Correctly For Variant Calling

Question

[d]

I am working with bam files for single-cell RNA sequencing data that have barcodes under the tag -XX and UMIs under the tag -SS. However, I am trying to use this bam file as input into multiple variant-calling workflows (freebayes+vartrix, cellsnp-lite, etc.), but am running into issues, as the outputs of all of these are empty and don't contain variants. I know these workflows work with bam files outputted by Cellranger, and am wondering if this is an issue with the way my Bam file is formatted. Since Cellraner uses the -CB tag for barcodes, how can I change my -XX to by -CB? Or is there another workaround?

[hxj5]

You may want to try --cellTAG and --UMItag options of cellsnp-lite to specify tag_for_cell_barcode and tag_for_UMI (check a list of full parameters with cellsnp-lite -h).

For example, cellsnp-lite <other options> --cellTAG XX --UMItag SS

Answer 1

[i.sudbery]

If you are comfortable with python, a quick script with pysam should sort this

import pysam

inbam = pysam.AlignmnetFile("input_file.bam", "rb")
outbam = pysam.AlignmentFile("output_bam_file.bam", "wb", template=inbam)

for read in inbam.fetch(until_eof=True):

    try:
        cell_barcode = read.get_tag("XX")
    except KeyError:
        # No XX tag - either skip read
        continue
        # or output as is without tag (delete one)
        outbam.write(read)

    read.set_tag("CB", cell_barcode)
    outbam.write(read)

outbam.close()

You could also use sed

$ samtools view -H input_file.bam | sed 's/XX:Z/CB:Z/g' | samtools view -b > output_bam_file.bam

However, this relies on the string XX:Z not appearing anywhere else in the file, other than in the tags. This is likely true, but its just possible it could appear in the a read name.