AnnotSV
AnnotSV copied to clipboard
VCF field format issue
Hi, The new AnnotSV VCF output format has an issue -- many numeric fields are being given Type=String. This breaks downstream pipelines that then do not allow proper numeric type filtering on "String" fields.
For example: ##INFO=<ID=B_inv_AFmax,Number=1,Type=String... ##INFO=<ID=B_gain_AFmax,Number=1,Type=String...
Can you please fix this?
Thanks
Hi,
Unfortunately, I can't help you with such an issue. The VCF output is produced with variantconvert. Can you switch and create an issue on the variantconvert GitHub?
Best, Véronique
Ok
Hi, I am reaching out to variantconvert, on this issue (https://github.com/SamuelNicaise/variantconvert/issues/27), but am not hearing back. Is there another way to fix this bug?
I contacted the author, I know he is actively working on a new version. He should give you an answer quickly.
Hello, I'm curious if you heard an update about this?
yes, but unfortunately I don't have any time slot to provide...
The variantconvert developper expects a fix to be available by the end of the year at worst. https://github.com/SamuelNicaise/variantconvert/issues/27#issuecomment-1784828701
Hi, The above issue with variantconvert has still not been fixed which makes it very hard to use the VCF output of annotSV. I'm having trouble having a response there. Is there any other way to address this issue? Thanks!
This bothers me a lot, but I don't have the skills to debug this. It's really specific to variantconvert. I will ask the developer if this can be done quickly.
Thanks for the new version! Any news on the variantconvert bug?
variantconvert: A new version should be posted in 1 or 2 weeks
Is there a different way for annotSV to generate VCF either with a different tool or writing this to be part of annotSV?
Because it doesn't seem that variantconvert is being maintained anymore, and I doubt that the bug fix will be made.
Thanks
I don't know any other tool to generate VCF.
If variantconvert is not updated, I will take the time to develop a module to generate VCFs this summer.
Ok thank you!
Hi, Variantconvert was finally updated: https://github.com/SamuelNicaise/variantconvert/issues/27#issuecomment-2063711704
Can you now incorporate this into the next version of AnnotSV? When might that be? Thanks!
Yes, that's great. variantconvert version 2.0.0 is posted and this should fix the bug. I'm going on vacation this evening. I'll take care of it when I get back.
AnnotSV 3.4.1 is posted with an updated version of variantconvert (v2.0.1).
The variantconvert conversion mode can be set by the user in AnnotSV (see the -variantconvertMode options). Instead of combining all full and split annotations (default mode: combined), variantconvert can represent each full or split annotation on one line (mode: full&split), or only "full" annotation can be kept (mode: full).
Regarding on your needs, please set the -variantconvertMode
option to full
or fullsplit
.
Can you confirm that this fixes your bug?
Thanks! I will check
Hi, any news?
We will test in a few days
Is there a way to upgrade annotSV while keeping our custom annotations / configurations and also the genecards annotation?
Sorry, no way to do a "real" upgrade.
You need to install the new version:
make PREFIX=. install
Then:
- Either use your previous annotations with the
-annotationsDir
option. - Or install the annotations de novo with:
make PREFIX=. install-human-annotation
and replace the $ANNOTSV/share/AnnotSV/Annotations_Human/Users/ directory with the previous one as well as the $ANNOTSV/etc/AnnotSV/configfile with the previous one.
Using the latest annotSV version, for some reason it is not finding variantconvert. Python 3.9 is giving an error that it couldn't find the variantconvert module.
Does the following directory exist?
ls -l $ANNOTSV/share/python3/variantconvert
Can you show me the complete output?
Yes it does. I already deleted the log so I don't have the complete output.
What I have done is I now went into this directory: $ANNOTSV/share/python3/variantconvert
and then I ran pip install -e .
Would this be an ok attempt to fix the problem?
This step should be done when running AnnotSV for the first time and this should create the $ANNOTSV/share/python3/variantconvert/pipinstall.flag
file.
Can you give me the content of this file?
The first time I ran it I had python 3.8. I think that explains the error.
But then the next time I ran it, I had python 3.9 but it couldn't find variantconvert.
Is it possible that the prior error pipinstall.flag makes AnnotSV think that it installed even though it didn't?
In that case, is my fix of manually installing it ok? Or should I uninstall variantconvert and try again?
Defaulting to user installation because normal site-packages is not writeable
Obtaining file:///gpfs/data/evronylab/bin/AnnotSV/share/python3/variantconvert
Installing build dependencies: started
Installing build dependencies: finished with status 'done'
Getting requirements to build wheel: started
Getting requirements to build wheel: finished with status 'done'
Preparing wheel metadata: started
Preparing wheel metadata: finished with status 'done'
ERROR: Package 'variantconvert' requires a different Python: 3.7.2 not in '>=3.8'
WARNING: You are using pip version 20.2.2; however, version 24.0 is available.
You should consider upgrading via the '/gpfs/share/apps/python/cpu/3.7.2/bin/python3.7 -m pip install --upgrade pip' command.
I think I will just uninstall variantconvert and delete the pipinstall.flag file and rerun and see if that works.
Ok that seems to have fixed the problem.
Is it possible that the prior error pipinstall.flag makes AnnotSV think that it installed even though it didn't?
Absolutely
In that case, is my fix of manually installing it ok?
Yes
I think I will just uninstall variantconvert and delete the pipinstall.flag file and rerun and see if that works.
Perfect too
In your output, I see:
ERROR: Package 'variantconvert' requires a different Python: 3.7.2 not in '>=3.8'
Python 3.8 is required. This seems not to be the case on your run.