AnnotSV passing wrong arguments to variant convert

[GACGAMA]

Hello!
I'm using variantconvert 2.0.1 and the latest AnnotSV within a conda installation.

When generating MANTA annotations, tsv files are produced correctly with:

AnnotSV -SvinputFile file1.vcf -annotationsDir /scratch4/nsobrei2/references/annotsv/AnnotSV_annotations/ -genomeBuild GRCh38 -outputFile file1_annotated -vcf 1 >& file1.log &

But variantconvert gives a range of bugs:

python3 /home/ggama1/.conda/envs/annotsv/share/python3/variantconvert//src/variantconvert convert -i file1.vcf -o file_annotated.vcf -c /home/ggama1/.conda/envs/annotsv/share/python3/variantconvert/src/variantconvert/configs/GRCh38/annotsv3_from_vcf.combined.local.json

2024-07-10 13:54:13 [INFO] running variantconvert 2.0.1
/scratch4/nsobrei2/programs/variantconvert/src/variantconvert/converters/vcf_from_annotsv.py:62: FutureWarning: Setting an item of incompatible dtype is deprecated and will raise in a future error of pandas. Value '[0.0 0.0 1.0 ... 0.0 0.0 0.0]' has dtype incompatible with float64, please explicitly cast to a compatible dtype first.

Traceback (most recent call last):
  File "/home/ggama1/.conda/envs/annotsv/lib/python3.12/site-packages/pyfaidx/__init__.py", line 391, in __init__
    self.file = open(filename, 'r+b' if mutable else 'rb')
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '/scratch4/nsobrei2/references/annotsv/Annotations_Human/BreakpointsAnnotations/GCcontent/GRCh38/GRCh38_chromFa.fasta'

Part of this error is because annotsv is passing the wrong location for GRCh38 annotation fasta file, as it should be:
/scratch4/nsobrei2/references/annotsv/AnnotSV_annotations/ and not /scratch4/nsobrei2/references/annotsv
as specified in my first command.

[lgmgeo]

It looks like you followed the Bioconda install (https://lbgi.fr/AnnotSV/downloads) correctly.

Can you show me the output header? (in particular the -annotationsDir value displayed)

@nvnieuwk, can you help us? Could this be a bug related to Conda?

[GACGAMA]

Hi @lgmgeo
I will rerun the data and get the header. Running manually with:

cd /home/ggama1/.config/variantconvert/2.0.1
variantconvert config -c GRCh38/* --set GENOME.path=/scratch4/nsobrei2/references/annotsv/AnnotSV_annotations/BreakpointsAnnotations/GCcontent/GRCh38/GRCh38_chromFa.fasta --fill_genome_header

variantconvert convert -i file1_annotated.tsv -o file1_annotated.vcf -c GRCh38/annotsv3_from_vcf.json >& file1_annotated.log &

Did work, so it seems like a bug in the conda installation of AnnotSV (probably)

[lgmgeo]

Maybe I have an idea...
Did you run AnnotSV without using the -annotationsDir option first?

Anyway, can you try:

find /home/ggama1/.conda/envs/annotsv/share/python3/variantconvert/ -name "*local*"

Then remove all "*local*" files from the installation.

Finally, re-run AnnotSV with the -annotationsDir option.

Does it work?

[sandratra-rab]

Hi @lgmgeo,
I have encountered the same issue. I gave the wrong -annotationsDir during my first execution of AnnotSV using the -vcf 1 parameter. Even though I did correct it for my next jobs, it kept returning me an error and the wrong path for the chomFa.fasta file.

I have applied your suggestion by removing the files annotsv3_from_bed.combined.local.json and annotsv3_from_vcf.combined.local.json from my python3/variantconvert/configs/GRCh38/ directory and I launched a re-run of AnnotSV. It worked and it generated the same json files but with the correct path this time.

[lgmgeo]

Great, thank you for your feedback!

I will think about a new check in the code to avoid such behavior in case of bad installation.