findAlleles light chain alleles not changing from 00

[liel-cohen]

Hello,

I ran the analyze command on 10x BCR data, sliced the clns file, and ran findAlleles on the sliced clns (MiXCR version and commands attached below).
For IGH clones, the results appeared correct: most clones had enough information with statuses of "DE_NOVO", "FOUND_KNOWN_VARIANT", or "NOT_CHANGED_AFTER_SEARCH", and alleles appropriately changed from *00 to the identified allele (including novel allele variant annotations).
However, for IGL and IGK clones, while most had sufficient information and showed statuses of "DE_NOVO" or "NOT_CHANGED_AFTER_SEARCH", none of the alleles changed from 00. Even when novel alleles were identified and novel allele variant annotations were added to the gene name, the allele number remained 00 (e.g., IGKV3-2000-M1-BMSZIWO2SGTE2U7U5ZU5EXAV).

Why do light chain allele numbers remain at *00 despite successful allele detection?
How can I extract the allele number used in the mixcr base gene library for the 00 allele?

Technical details

I'm using the dev version below due to an issue I have with the main 4.7.0 (#1847 (comment)):
MiXCR v4.7.0-104-develop (built Fri Sep 27 14:22:11 PDT 2024; rev=8b1ab7f480; branch=develop; host=fv-az1249-2) RepSeq.IO v2.5.0 (rev=06fa1852ee) MiLib v3.5.0-3-mitool-tag-types (rev=7d638011d4) Built-in V/D/J/C library: repseqio.v5.1

These are the commands I used:

mixcr -Xmx200g analyze 10x-sc-xcr-vdj -s hsa O_1{{a}}_L{{n}}_{{R}}_001.fastq.gz ~/folder/ 

mixcr slice -Xmx150g \
--ids-file ~/folder/donor_cloneIds/cloneIds_donorX.txt \
--verbose \
~/folder/assembledCells.clns \
~/folder/assembledCells_donorX.clns

mixcr -Xmx150g findAlleles \
    --output-template {file_dir_path}/{file_name}_w_alleles.clns \
    --export-library ~/folder/donor_alleles/alleles_donorX.fasta \
    --export-alleles-mutations ~/folder/donor_alleles/alleles_mutations_donorX.tsv \
    --report ~/folder/donor_alleles/alleles_report_donorX.txt \
    --verbose \
    --use-local-temp \
    ~/folder/assembledCells_donorX.clns

Thank you so much for any help!
Liel

[mizraelson]

Hi Liel,
The thing is, we don’t rely solely on IMGT alleles because we’ve found a number of errors in their dataset. For the heavy chain, we manually verified the alleles using raw data. We also published a paper on this, which you can find here.

However, we haven’t done the same verification for the light chain. As a result, the number remains 00, though the alleles are still being correctly identified-they’re just all marked as DE_NOVO.

Mark

[liel-cohen]

I see, thanks!