TRGTdb

[ACEnglish]

Adding code for coverting a TRGT output VCF into a database. See tdb_tutorial.md for usage details.

TODOs:

• This code uses truvari v4.0-dev (min commit 6edb54f). This requirement is because there were changes made to truvari.vcf2df. Until v4.0 is released, truvari will need to be manually installed. After truvari v4.0 is cut, we can simply uncomment the line from trgt/setup.py that installs it (line 31).
• Currently, the trgt python package is versioned as 0.0.1. This should be synchronized with the main trgt executables' version.
• trgt.database.dbutils.pull_saps assumes the allele length range is stored in the vcf as FORMAT/ALLR. However trgt v0.3.4 writes FORMAT/ALCI. Therefore, this code isn't compatible with trgt v0.3.4.
• There is a placeholder in trgt.__main__ for wrapping the trgt main executable. If we want to distribute trgt with a single command line interface (e.g. trgt run, trgt viz, trgt db), we'll need to place the executables into the repository, update MANIFEST.in to package those executables, and then make external calls from run_main (e.g. Popen(os.path.join(trgt.__file__, 'bin', 'trgt'))
• There has been no license work on this code beyond setting the whole package to "BSD 3-Clause Clear License". I haven't checked that all of the dependencies are okay for PacBio to use. For example, this code uses pyarrow which is distributed by apache with an 'Apache-2.0 license`, so that's probably okay. But there are a total of 20 dependencies that would need to be checked. To get the full list of dependencies:

git clone <trgt>
cd trgt/
#Making an isolated virtualenv for trgt
python3 -m venv mtrgtpy  
source mtrgtpy/bin/activate
# manually installing truvari v4.0-dev
cd ../truvari/
python3 -m pip install .
cd -
# installing trgt and its dependencies
python3 -m pip install .
python3 -m pip freeze | wc -l

[zqfang]

Hi @ACEnglish,

I got an issue runing the following cmd

trgt db create -o strains.tdb trgt_out/SJL.sorted.vcf.gz   

The error message is ValueError: cannot reindex on an axis with duplicate labels

Traceback (most recent call last):
  File "/home/fangzq/.conda/envs/trgt/bin/trgt", line 33, in <module>
    sys.exit(load_entry_point('TRGT', 'console_scripts', 'trgt')())
  File "/home/fangzq/github/trgt/trgt/__main__.py", line 44, in main
    CMDS[args.cmd](args.options)
  File "/home/fangzq/github/trgt/trgt/dbcmds.py", line 28, in db_main
    CMDS[args.cmd][1](args.options)
  File "/home/fangzq/github/trgt/trgt/database/create.py", line 74, in create_main
    n_data = trgt.load_tdb(i) if i.rstrip('/').endswith(".tdb") else trgt.vcf_to_tdb(i)
  File "/home/fangzq/github/trgt/trgt/database/dbutils.py", line 209, in vcf_to_tdb
    allele_df = pull_alleles(data)
  File "/home/fangzq/github/trgt/trgt/database/dbutils.py", line 156, in pull_alleles
    alleles["LocusID"] = data["LocusID"]

Do you have any preprocessing step for importing trgt output to trgtdb ?

my trgt cmd is

./trgt-v0.8.0-linux_x86_64 --genome mm10.fa --repeats tr_catalog.adjusted.mm10.bed --reads SJL.aligned.sorted.bam --output-prefix trgt_out/SJL --threads 6

[ACEnglish]

Database tool has been refactored and placed into a repository at https://github.com/ACEnglish/tdb.

@zqfang - Please try from that repository and if the error still happens, open a ticket there.