name: dbnsfp
version: 0.0.2
output_asset_class: dbnsfp
description: |
  A database developed for functional prediction and annotation of all
  potential non-synonymous single-nucleotide variants (nsSNVs) in the
  human genome (Gencode release 29/Ensembl 94).
input_files:
  dbnsfp:
    description: zipped dbNSFP database file
input_params:
  threads:
    default: "8"
    description: Number of threads to use for parallel computing
input_assets: {}
docker_image: quay.io/biocontainers/htslib:1.24--ha79157c_0
command_templates:
  - cp {{values.files["dbnsfp"]}} {{values.output_folder}}/{{values.genome_digest}}.zip
  - unzip {{values.output_folder}}/{{values.genome_digest}}.zip -d {{values.output_folder}}
  - gunzip -v {{values.output_folder}}/*variant.chr*.gz
  - head -n1 {{values.output_folder}}/dbNSFP*_variant.chr1 > {{values.output_folder}}/{{values.genome_digest}}_dbNSFP.txt
  - cat {{values.output_folder}}/dbNSFP*variant.chr* | grep -v '#' >> {{values.output_folder}}/{{values.genome_digest}}_dbNSFP.txt
  - rm {{values.output_folder}}/dbNSFP*_variant.chr*
  - bgzip -@ {{values.params["threads"]}} {{values.output_folder}}/{{values.genome_digest}}_dbNSFP.txt
  - tabix -s 1 -b 2 -e 2 {{values.output_folder}}/{{values.genome_digest}}_dbNSFP.txt.gz
  - find {{values.output_folder}} -type f -not -path '{{values.output_folder}}/_refgenie_build*'
    -not -path '{{values.output_folder}}/{{values.genome_digest}}_dbNSFP.txt.*' -exec rm -f {} +
custom_seek_keys:
  # sed -E, not grep -P: the biocontainers image's grep is BusyBox (no -P),
  # and the whole pipeline runs inside that one image now.
  version: tabix --version 2>/dev/null | sed -nE 's/^tabix \(htslib\) ([0-9][^ ]*).*/\1/p' | head -1
default_asset: "{{values.custom_seek_keys.version}}"

tags:
- variant-annotation
- database
- human
outputs:
- pattern: '*_dbNSFP.txt.gz'
  description: dbNSFP database file (bgzipped)
- pattern: '*_dbNSFP.txt.gz.tbi'
  description: Tabix index file
test:
  commands:
  - test -f {output_dir}/{genome}_dbNSFP.txt.gz
  - test -f {output_dir}/{genome}_dbNSFP.txt.gz.tbi
resources:
  memory: 8GB
  disk: 100GB
  time: 4h
metadata:
  author: nsheff
  created: '2026-05-20'
  license: BSD-2-Clause