Issues getting full output with ONT data

[bioinfoAG]

Hello,
Thank you for creating such an efficient tool!
I am doing HLA typing on ONT data and I installed SpecImmune using Docker. I fail to get all the output files including the visualization pdf and sample_id.GENE.final.type.result.formatted.txt. I managed to see the promising results from sample_id.GENE.type.result.txt file, but I would like to fix the issue and get the full output.
The only error present in the log is connected to dysgu. This is the error message I get:
KeyError: 'nanopore'
dysgu is empty, use sniffles and cutesv
Is there a way to fix dysgu so that it takes 'nanopore' as a flag?
Thank you in advance for your help!

[troyeeee]

Hi,
Sorry for the late reply. dysgu does support the nanopore flag, and the message “dysgu is empty” usually means no SVs were found in that region.

Could you share a small subset of your data that reproduces the issue? That would help us check what is going on.

[bioinfoAG]

Hi, thank you for your help!
Here's a subset of a cram that I used.

subset.sample.cram.gz

[bioinfoAG]

Hello!
In the end, I managed to get the output for some samples. However, in some cases, I only manage to get the sample.HLA.type.result and the issue is this:

Traceback (most recent call last):
File "/usr/lib/python3.12/multiprocessing/pool.py", line 125, in worker
result = (True, func(*args, **kwds))
^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/multiprocessing/pool.py", line 48, in mapstar
return list(map(*args))
^^^^^^^^^^^^^^^^
File "/usr/local/bin/cuteSV", line 750, in process_process_sigs_type
with open("%ssignatures/%s%s.pickle"%(temporary_dir,pid,sv_type), "rb") as f:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '/path/sample/Genes_step2//HLA-A_work/cutesv/signatures/930695DUP.pickle'

Could you please help me understand the underlying issue with the pickle file?
Thank you in advance!

[troyeeee]

Hi, thank you for your help! Here's a subset of a cram that I used.

subset.sample.cram.gz

Hi, I tried to extract reads from the provided subset.sample.cram.gz, but it requires the original hg38_chr_withM.fasta reference, which I don’t have and can’t match via MD5 to any public GRCh38 build. Could you instead provide either the corresponding FASTQ files or a BAM aligned to a standard public reference , so I can reliably reproduce and debug the dysgu issue?

[troyeeee]

Hello! In the end, I managed to get the output for some samples. However, in some cases, I only manage to get the sample.HLA.type.result and the issue is this:

Traceback (most recent call last): File "/usr/lib/python3.12/multiprocessing/pool.py", line 125, in worker result = (True, func(*args, **kwds)) ^^^^^^^^^^^^^^^^^^^ File "/usr/lib/python3.12/multiprocessing/pool.py", line 48, in mapstar return list(map(*args)) ^^^^^^^^^^^^^^^^ File "/usr/local/bin/cuteSV", line 750, in process_process_sigs_type with open("%ssignatures/%s%s.pickle"%(temporary_dir,pid,sv_type), "rb") as f: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: [Errno 2] No such file or directory: '/path/sample/Genes_step2//HLA-A_work/cutesv/signatures/930695DUP.pickle'

Could you please help me understand the underlying issue with the pickle file? Thank you in advance!

Hi,

This issue is likely caused by a strict check in cuteSV. In particular, it can happen when the reference used for read alignment is not exactly the same as the reference used when running cuteSV. Please double‑check that the reference chromosomes (names and sequences) in your alignment and in the SpecImmune run are identical.

Hope this helps!

[troyeeee]

Hi, thank you for your help! Here's a subset of a cram that I used.

subset.sample.cram.gz

Hi,
I extracted the reads following your command in the email and ran SpecImmune successfully in the conda environment.

Are you using the latest code to build the Docker image? There might be version differences causing the issue.

Also, the CRAM file looks a bit unusual — the alignments don't seem to be in the MHC region, and most reads are secondary alignments with MAPQ=0.

[troyeeee]

Hello! In the end, I managed to get the output for some samples. However, in some cases, I only manage to get the sample.HLA.type.result and the issue is this:

Traceback (most recent call last): File "/usr/lib/python3.12/multiprocessing/pool.py", line 125, in worker result = (True, func(*args, **kwds)) ^^^^^^^^^^^^^^^^^^^ File "/usr/lib/python3.12/multiprocessing/pool.py", line 48, in mapstar return list(map(*args)) ^^^^^^^^^^^^^^^^ File "/usr/local/bin/cuteSV", line 750, in process_process_sigs_type with open("%ssignatures/%s%s.pickle"%(temporary_dir,pid,sv_type), "rb") as f: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: [Errno 2] No such file or directory: '/path/sample/Genes_step2//HLA-A_work/cutesv/signatures/930695DUP.pickle'

Could you please help me understand the underlying issue with the pickle file? Thank you in advance!

Hello!

We have updated the code regarding SV processing. The pipeline will now skip cases where read quality is too poor, as SV results in such scenarios are generally unreliable. This prevents the crash you are seeing due to missing intermediate files.

Please pull the latest code and rebuild your environment to resolve this issue.

Best regards,