all_contig_dandelion.tsv

[dt583]

Description of the question

Hello,

I'm following the tutorial for BCR analysis and ran into an issue.

During the 'Singularity preprocessing' step, the file all_contig_dandelion.tsv was generated. However, in the next step, 'Loading data for analysis,' the tutorial uses filtered_contig_dandelion.tsv instead of all_contig_dandelion.tsv.

When I ran the following command using all_contig_dandelion.tsv:

singularity run -B $PWD /path/to/sc-dandelion_latest.sif changeo-clonotypes --h5ddl vdj.h5ddl

I encountered the following error message:

Error in (function (db, sequenceColumn = "junction", vCallColumn = "v_call", :
3835 cell(s) with multiple heavy chains found. One heavy chain per cell is expected.

Is the step "vdj, adata = ddl.pp.check_contigs(vdj, adata)" sufficient for filtering the V(D)J data (derived from the all_contig_dandelion.tsv), or are additional filtering steps required?

Also, since I am using Seurat for GEX data analysis and have already completed the QC steps there, I would like to confirm: Is an AnnData object converted from Seurat suitable for the adata argument in the ddl.pp.check_contigs function?

Thanks in advance for your help!

Minimal example

Any error message produced by the code above

OS information

No response

Version information

No response

Additional context

No response

[zktuong]

hi @dt583 , apologies for the late response as i thought i had responded.

During the 'Singularity preprocessing' step, the file all_contig_dandelion.tsv was generated. However, in the next step, 'Loading data for analysis,' the tutorial uses filtered_contig_dandelion.tsv instead of all_contig_dandelion.tsv.

When I ran the following command using all_contig_dandelion.tsv:

singularity run -B $PWD /path/to/sc-dandelion_latest.sif changeo-clonotypes --h5ddl vdj.h5ddl

I encountered the following error message:

Error in (function (db, sequenceColumn = "junction", vCallColumn = "v_call", :
3835 cell(s) with multiple heavy chains found. One heavy chain per cell is expected.

So this issue is a new thing that's being enforced from the R code in the interaction with the immcantation packages, so it's behaving like a bug just for this aspect. It's nothing to do with all/filtered data as input.

I'm in the process of refactoring this part, along with other aspects of the code and i hope to be able to release the new version in the next couple of weeks.

Is the step "vdj, adata = ddl.pp.check_contigs(vdj, adata)" sufficient for filtering the V(D)J data (derived from the all_contig_dandelion.tsv), or are additional filtering steps required?

yes this should be sufficient. there are a few parameters to toggle e.g. how many heavy/light chains per cell to keep in this function

Also, since I am using Seurat for GEX data analysis and have already completed the QC steps there, I would like to confirm: Is an AnnData object converted from Seurat suitable for the adata argument in the ddl.pp.check_contigs function?

Yes! as long as the cell barcodes match, it should work.