VCF to RDF Mapping

VCF to RDF Mapping

The team worked to produce an ontology mapping and supporting software to expose Variant Call Format (VCF) files as linked data, in order to facilitate offline batch conversion of VCF to various RDF formats as well as to enable online SPARQL query access to compressed VCF files directly.

BioHackathon 2015 Team

• Worked on by Raoul, Arto, and Kieron, with some support from Pjotr and Jerven.
• Built on previous work at BioHackathon 2014 and RDF Summit by Raoul, Francesco, Kieron, James, Will and Simon.

Software Produced

• https://github.com/ruby-rdf/rdf-vcf (RDF::VCF plugin for RDF.rb)
  • Released on RubyGems, installable with jruby -S gem install rdf-vcf. (Note: Requires JRuby 9.0+.)
  • Includes a vcf2rdf program to transform VCF files into RDF as a batch job.
  • Implements an RDF.rb reader for VCF and BCF files, supporting also bgzipped and tabix-indexed files.
• https://github.com/helios/bio-sparql-otf (OTF: On-the-Fly conversion)
  • Implements a proof-of-concept-quality SPARQL backend for VCF files, based on RDF.rb and RDF::VCF.
  • Continued from BioHackathon 2014 work: https://github.com/dbcls/bh14/wiki/On-The-Fly-RDF-converter
• https://github.com/JervenBolleman/sparql-vcf (Java+Sesame)
  • Implements a proof-of-concept-quality SPARQL backend for VCF files, based on Sesame.

Mapping

There are currently two distinct mappings:

• Ensembl/FALDO. This originates from BioHackathon 2014 work and is presently described in the RDF::VCF source code.

• GFVO (Genomic Feature and Variation Ontology)

Examples

Command-Line Interface (CLI)

The CLI utility called vcf2rdf transforms VCF files into RDF (currently outputting N-Triples):

vcf2rdf Homo_sapiens.1.vcf.gz Homo_sapiens.2.vcf.gz ...

The input files can be either plain-text VCF or compressed by bgzip (as in the above example).

Application Programming Interface (API)

The RDF::VCF gem can be used like any other RDF.rb reader plugin:

# Load the RDF::VCF library:
require 'rdf/vcf'

# Open a VCF file for reading:
RDF::VCF::Reader.open('Homo_sapiens.vcf.gz') do |reader|

  # Loop over all generated RDF statements:
  reader.each_statement do |statement|

    # Print the RDF statement to the screen:
    $stdout.puts statement.inspect
  end
end

SPARQL Query Interface

SELECT ?s ?quality WHERE {
  ?s faldo:location ?location .
  ?location faldo:reference [ dc:identifier "Y" ] .
  ?location faldo:begin [ faldo:position ?begin ] .
  ?location faldo:end [ faldo:position ?end ] .
  ?s vcf:quality ?quality .
  FILTER(?begin >= 2749180)
  FILTER(?end <= 2755180)
}