The purpose of this project is to convert databases that are required by varfish into a format that can be easily imported (i.e. TSV files with a header containing the column names of the corresponding varfish database table).

• bcftools

We recommend the installation of the requirements via conda:

conda install bcftools bedops samtools

git clone [email protected]:bihealth/varfish-anno.git
cd varfish-anno

make init

This step creates the folder structure in databases/.

make download

The downloads will be stored in databases/<database_name>/download/. The reference is placed in downloads/.

Please note that the download routine is not sophisticated. You might want to double check the process, especially in case something breaks. It is thought as extensive instructions to download the required databases. If the files are already available to you, you can place them in the corresponding download folder and omit this step. Note that in this case the conversion scripts might need some adaption to match the correct file name (see next section).

The download links are defined in downloads/Makefile and the variable names are prefixed with URL_. Those variables are safe to change (if the downloaded file contains the expected format).

• The KEGG database is not automatically downloadable. Instructions are printed to obtain the required files (or see instructions below). They need to be placed in databases/kegg/downloads.
• The case files are in .ped format and are individual depending on your project. You need to place them in databases/case/download.
• Copy the resulting -vars file from Varhab to databases/annotation
• Copy the resulting -gts file from Varhab to databases/smallvariant

Note that ExAC, gnomAD and dbSNP databases are rather large files and will take time to download.

• https://genome.ucsc.edu
• Tools -> Table Browser
  • group: All tables
  • assembly: GRCh37
  • table: keggPathway
  • output format: selected fields from primary and related tables
  • output file: genetokegg.tsv
  • get output
    • Linked Tables -> knownGene -> allow selection from checked tables
    • Linked Tables -> ensGtp -> allow selection from checked tables
    • Select Fields from keggPathway -> mapID
    • ensGtp fields -> gene
    • get output

• https://genome.ucsc.edu
• Tools -> Table Browser
  • group: All tables
  • assembly: GRCh37
  • table: keggMapDesc
  • output format: selected fields from primary and related tables
  • output file: kegginfo.tsv
  • get output
    • Select Fields from keggMapDesc -> mapID
    • Select Fields from keggMapDesc -> description
    • get output

make convert

Every script defines a HEADER, INPUT and OUTPUT variable, and, if needed, a REF variable. The names should be self-explanatory. They are preset to the downloaded files. You can change the INPUT and REF, if needed.

Note that ExAC, gnomAD and dbSNP databases are rather large files and especially dbSNP will take time to convert.