Fix for barcode reuse?

[AMechels]

Dear developers,

I wanted to use your tool enclone on 2 bcr files from 1 donor, before and after treatment. However, I get following message:

Significant barcode reuse detected.  If at least 25% of the barcodes in one dataset
are present in another dataset, is is likely that two datasets arising from the
same library were included as input to enclone.  Since this would normally occur
only by accident, enclone exits.  If you wish to override this behavior,
please rerun with the argument ACCEPT_REUSE.

Here are the instances of reuse that were observed:

bcrPBMC_JUQ064_hg19, bcrPBMC_JUQ065_hg19 ==> 245 of 972, 829 barcodes (29.6%)

The barcodes don't have a prefix or suffix, so it's only normal that some are present in both. Is there an argument I can set that adds a string to the barcode using the META csv file?

Thanks in advance, Aurelie

[DavidBJaffe]

Hi Aurelie,

That is statistically way more than one would expect by chance. The most likely explanation is that contamination (laboratory or conceivably informatic) occurred somewhere.

I would suggest looking to see if the contig sequences themselves are the same.

Regardless, proceed with caution, because ignoring the warning could lead to incorrect scientific conclusions.

All the best,

David

On Mon, Jul 18, 2022 at 9:45 AM AMechels @.***> wrote:

Dear developers,

I wanted to use your tool enclone on 2 bcr files from 1 donor, before and after treatment. However, I get following message:

Significant barcode reuse detected. If at least 25% of the barcodes in one dataset are present in another dataset, is is likely that two datasets arising from the same library were included as input to enclone. Since this would normally occur only by accident, enclone exits. If you wish to override this behavior, please rerun with the argument ACCEPT_REUSE.

Here are the instances of reuse that were observed:

bcrPBMC_JUQ064_hg19, bcrPBMC_JUQ065_hg19 ==> 245 of 972, 829 barcodes (29.6%)

The barcodes don't have a prefix or suffix, so it's only normal that some are present in both. Is there an argument I can set that adds a string to the barcode using the META csv file?

Thanks in advance, Aurelie

— Reply to this email directly, view it on GitHub https://github.com/10XGenomics/enclone/issues/494, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC37EKCPQVJBYDC6ADF4E4TVUWC3BANCNFSM5344VWXQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[nh3]

Hello,

I encountered the same warning but under a different experiment design. We used cell-hashing to pool different donors and demultiplexed the data according to https://www.10xgenomics.com/resources/analysis-guides/demultiplexing-and-analyzing-5%E2%80%99-immune-profiling-libraries-pooled-with-hashtags . Then, when running enclone across each demultiplexed sample's cellranger-multi outputs using the "p1;p2;p3" notion, it complains "Significant barcode reuse detected". It appears that "all_contig_annotations.json" generated from demultiplexed samples contain the same set of contigs only differ in the is_cell/is_gex_cell/is_asm_cell fields, and enclone seems to ignore those. Is it valid to filter down to cell-only contigs and supply those to enclone?

Many thanks!