pyvariantdb 1.3

A python package to download, process and access dbSNP data.

dbSNP to Parquet Converter

A python-package focused rewrite of Weinstock's dbSNP to parquet repository. This package allows convenient download, processing and access to data from dbSNP. A simple python interface can be used to work with the data once it is processed.

Installation

pyvariantdb is available on PyPI and can be installed from there with the package management tool of choice. For development we use pixi:

# install pixi
curl -fsSL https://pixi.sh/install.sh | sh
pixi update && pixi install

Usage

We recommend to prepare the data from the command line before using the package since download and processing takes quit some time. Per default the data is stored at ~/.cache/pyvariantdb. This can be changed through the usage of environment variables:

export PYVARIANTDB_HOME = "/raid/cache/pyvariantdb"

Execution of the pipeline can be done with:

# activate pixi
pixi shell -e default
# download dbsnp
pyvariantdb-download
# transform the database to parquets
pyvariantdb-make-dbsnp 

We aim to provide an easy interface to extract genomic coordinates from dbSNP. For this, the main functionality is extremely simple. Just provide the rs-identifier (with or without chromosome) and we fetch the respective coordindates. However, to achieve this we need to run the pipeline mentioned above.

Extracting coordinates is simple:

from pyvariantdb.lookup import SNPLookup

lookup = SNPLookup()
# P53 mutations, chromosome 17
rsids = ["rs1042522", "rs17878362", "rs1800372"]
df_all = lookup.query_all(rsids)
df_chr = lookup.query_chromosome(rsids, "17")
print(df_all)
print(df_chr) 

Processing Pipeline

pyvariantdb offers some quality of life improvements for working with dbSNP and the original repository. The original pipeline remains the same:

1. Downloads dbSNP data (GRCh38 build 156)
2. Filters for SNVs only
3. Converts chromosome contigs to standard naming
4. Splits data by chromosome
5. Creates Parquet lookup tables with RSID mappings

Output

The script entrypoint generates the following files on-disk to the defined cache dir.

• dbSNP_156.bcf - Full filtered BCF file
• dbSNP_156.chr*.bcf - Per-chromosome BCF files
• dbSNP_156.chr*.lookup.parquet - Per-chromosome RSID lookup tables
• dbSNP_156.lookup.parquet - Combined RSID lookup table