Missing tsv files in MiXCR 4.7.0 output for 10X GEX pipeline

[genome2020]

Hi,

I just ran latest version (4.7.0) of MiXCR on my 10X 5' GEX data. However, I am missing several .tsv files that were present when I ran older version of MiXCR on the same data. Specifically the following files are missing.

S48.clone.groups_mixed.tsv
S48.clone.groups_TRAB.tsv
S48.clone.groups_TRGD.tsv

The command I used is:

mixcr -Xmx100g analyze 10x-sc-5gex --threads 16 --species hsa S48_R1_001.fastq.gz S48_R2_001.fastq.gz S48

Could you please help with the issue? Thanks!

[mizraelson]

Hi, what files did you get in the output?

[mizraelson]

Can you share the last few lines of stdout to confirm that the analysis has finished successfully? I don’t see the assembleContigs report, which might mean that the run failed due to insufficient resources, for example.

[genome2020]

Here are the last lines for a run with a different sample that I set up with more memory. The tsv files are still missing

[genome2020]

@mizraelson any further insights into why this is happening?

[mizraelson]

Hi, yes. Sorry for the delay. The thing is that the behavior of the mixcr assembleCells command has changed. Now it requires all clones to be covered by the same feature, while the 10x-sc-5gex preset returns clones that are covered as much as possible for each individual sequence.

You can generate the grouped clonotype tables by adding --assemble-contigs-by VDJRegion (or another feature) to the analyze command to make sure the clones in the output are equally covered. Additionally, include parameters that add the assembleCells and exportCloneGroups steps to the pipeline:

mixcr analyze 10x-sc-5gex \
-s hsa \
--assemble-contigs-by VDJRegion \
--add-step assembleCells \
--add-step exportCloneGroups \
input_R1.fastq.gz \
input_R2.fastq.gz \
output

[genome2020]

Thanks for your help.