Preset for long-read RNAseq with cell barcode

[Euchiz]

Hi everyone,

I am currently working with single-cell Oxford Nanopore RNAseq data that contains both a 10x cell barcode and a UMI at the beginning of each read. Unfortunately, there seem to be no existing presets (or could anyone point out which should I use) that cater to this specific combination of single-cell barcoding and long-read sequencing parameters.

While I have reviewed the existing long-read RNAseq presets from the website, they do not include options for handling single-cell barcodes. I attempted to modify the YAML configuration files to create a custom preset that combines the 10x single-cell workflow with long-read sequencing parameters.

I have reviewed a previous discussion on a related topic (Discussion #1024), but the provided guidance was based on outdated YAML configurations.

Could you please provide assistance or guidance on creating a custom preset for this specific use case?
FYI here is the read structure I am working with:

"^(CELL:N{16})(UMI:N{12})(R1:*)"

Any help with the correct configuration or a new preset would be greatly appreciated.

By the way, I have already used (wf-single-cell) to refine my nanopore reads. Does it mean I can remove the refinement step and use the result bam file from wf-single-cell directly for alignment and assembling?

Thank you for your support and for developing such a powerful tool!

[mizraelson]

Hi,
I assume you use 10xV3 kit, correct? Are you sure the reads do not contain any ONT adapter sequences before the cell barcode at the beginning? Assuming the pattern can be in reverse complement, as is usually the case with ONT, I would try the following custom preset. Unzip the YAML file and put it the directory where you run MiXCR from or to the ~/.mixcr/presets folder. The command would then be:

mixcr analyze local:10x-ont \
    --species hsa \
    input_R1.fastq.gz \
    input_R1.fastq.gz \
    output

Use the latest develop MiXCR version.

I recommend using the raw dataset, not preprocessed by wf-single-cell.

[Euchiz]

Thank you so much! You are right there should be adapter before the barcode. I changed the expression and it worked.

[mizraelson]

Great! Maybe you can share the adjusted preset (with the adapter) here in case other users struggling with the same protocol.

[JackYGhannam]

Hello!

Thank you @mizraelson for all of your hard work on this. A follow-up question: would this custom preset also work if one wanted to process 10x 5'v2 single-cell data for which TCRs were subsequently enriched and sequenced via nanopore sequencing? Thanks again!

[mizraelson]

Yes, but you should change the pattern and the cell barcodes whitelist accordingly to ^(CELL:N{16})(UMI:N{10})\^(R2:*) and builtin:737K-august-2016 in the preset file.

[austinc26]

HI!

Thanks @mizraelson for helpful insight on this! I was just wondering if I have a 10x Multiome dataset that was later sequenced using Oxford Nanopore, could I use this preset if I trimmed the TSO and R1 adapters and discarded reads that did not have either adapter and also used the 737k-arc-v1 cell barcode whitelist in the yaml file. Thanks!

[mizraelson]

I am not sure I understand the data structure you will get. But the 737k-arc-v1 barcodes list is not build into MiXCR because it is not used for VDJ libraries. We might add it though in the upcoming version.