Skip to content

Beacon - Discovery Services for Genomic Data

Beacon IconThe Beacon protocol defines an open standard for genomics data discovery, developed by members of the Global Alliance for Genomics & Health. Since 2016, the Beacon protocols is being developed through the ELIXIR Beacon project as a GA4GH driver project.

As part of the project, since early 2016 the Computational Cytogenetics and Oncogenomics Group at the University of Zurich develops the Beacon+ demonstrator, to demonstrate current functionality and future Beacon protocol extensions.

The Beacon+ implementation is a custom front end on top of the Progenetix dataset, with emphasis on structural genome variations from cancer samples.

On 2020-01-20, Beacon+ became part of the ELIXIR Beacon Network.

BeaconPlus Data / Query Model

The Progenetix / Beaconplus query model utilises the GA4GH core data model for genomic and (biomedical, procedural) queries and data delivery.

The GA4GH data model for genomics recommends the use of a principle object hierarchy, consisting of

  • variant (a.k.a. genomicVariation)
    • a single molecular observation, e.g. a genomic variant observed in the analysis of the DNA from a biosample
    • mostly corresponding to the "allele" concept, but with alternate use similar to that in VCF (e.g. CNV are no typical "allelic variants")
    • in Progenetix identical variants from different sampleas are identified through a compact digest (variantInternalId) and can be used to retrieve those distinct variants (c.f. "line in VCF")
  • callset
    • the entirety of all variants, observed in a single experiment on a single sample
    • a callset can be compared to a data column in a VCF variant annotation file
    • callset has an optional position in the object hierarchy, since variants describe biological observations in a biosample
  • biosample
    • a reference to a physical biological specimen on which analyses are performed
  • individual
    • in a typical use a human subject from which the biosample(s) was/were extracted

In the Progenetix backend we mirror the GA4GH data model in the storage system, consisting of the corresponding

  • variants
  • callsets (compares to runs + analyses)
  • biosamples
  • individuals

collections of MongoDB databases. These collections are addressed by scoped queries.

filters Filters / Filtering Terms

Filters represent a way to allow the resource provider to direct "self-scoped" query values to the corresponding attributes in their backend resource. In the Progenetix implementation, a lookup table followed by scope assignment is used to map prefixed filter values to the correct attributes and collections. Most of the filter options are based on ontology terms or identifiers in CURIE format (e.g. NCIT:C4033, cellosaurus:CVCL_0030 or PMID:16004614). For use case examples please look below; documentation of available ontologies and how to find out about available terms can be found on the Classifications and Ontologies page.

In Beacon v2, the new FilteringTerms schema adds options to specify different types of filters (OntologyFilter, AlphanumericFilter, CustomFilter) which can contain a number of parameters to define e.g. scope or matching behaviour. These more complex terms are only available through PUT requests.

Example
"filters": [
    {
        "id": "NCIT:C4536",
        "scope": "biosamples",
        "includeDescendantTerms": false
    }
],

Hierarchical Terminologies in Filter Queries

Hierarchical terminologies allow queries at different levels, to include all its children terms. The Progenetix query filter system adopts this inclusion logic if the classification / code type is hierarchy-based. However, the includeDescendantTerms pragma can be used to modify this behaviour - globally if provided in a GET request (&includeDescendantTerms=false) or as part of filter objects (see above).

Examples for codes with hierarchical treatment within the filter space are:

  • NCIt
  • true, deep hierarchical ontology of cancer classifications
  • Cellosaurus
  • derived cell lines are also accessible through the code of their parental line

Beacon API

Beacon-style JSON responses

The Progenetix resource's API utilizes the bycon framework for data query and delivery and represents a custom implementation of the Beacon v2 API.

The standard format for JSON responses corresponds to a generic Beacon v2 response, with the meta and response root elements. Depending on the endpoint, the main data will be a list of objects either inside response.results or (mostly) in response.resultSets.results. Additionally, most API responses (e.g. for biosamples or variants) provide access to data using handover objects.

Example responses can be genrated through the path examples below.

Please be aware that Beacon responses use camelCased parameter names.

Beacon v2: Path Examples in Progenetix

The Beacon v2 protocol uses a REST path structure for consistant data access.

The Beacon+ implementation - developed oin the bycon project implements an expanding set of those Beacon v2 paths for the Progenetix resource.


Base /

The root path provides the standard BeaconInfoResponse.


Base /filtering_terms

/filtering_terms/
/filtering_terms/ + query

Base /biosamples

/biosamples/ + query
/biosamples/{id}/
/biosamples/{id}/variants/ & /biosamples/{id}/variants_in_sample/
/biosamples/{id}/analyses/

Base /individuals

/individuals/ + query
  • /individuals/?filters=NCIT:C7541
  • this example retrieves all individuals having an annotation associated with NCIT:C7541 (retinoblastoma)
  • in Progenetix, this particular code will be part of the annotation for the biosample(s) associated with the returned individual
  • /individuals/?filters=PATO:0020001,NCIT:C9291
  • this query returns information about individuals with an anal carcinoma (NCIT:C9291) and a known male genotypic sex (PATO:0020001)
  • in Progenetix, the information about its sex is associated with the Individual object (and rtherefore in the individuals collection), whereas the information about the cancer type is a property of the Biosample (and therefore stored in the biosamples collection)
/individuals/{id}/
/individuals/{id}/variants/ & /individuals/{id}/variants_in_sample/

Base /variants

There is currently (April 2021) still some discussion about the implementation and naming of the different types of genomic variant endpoints. Since the Progenetix collections follow a "variant observations" principle all variant requests are directed against the local variants collection.

If using g_variants or variants_in_sample, those will be treated as aliases.

/variants/ + query
/variants/{id}/ or /variants_in_sample/{id} or /g_variants/{id}/
/variants/{id}/biosamples/ & variants_in_sample/{id}/biosamples/

Base /analyses (or /callsets)

The Beacon v2 /analyses endpoint accesses the Progenetix callsets collection documents, i.e. information about the genomic variants derived from a single analysis. In Progenetix the main use of these documents is the storage of e.g. CNV statistics or binned genome calls.

/analyses/ + query

bycon Beacon Server

The bycon project provides a combination of a Beacon-protocol based API with additional API services, used as backend and middleware for the Progenetix resource.

bycon has been developed to support Beacon protocol development following earlier implementations of Beacon+ ("beaconPlus") with now deprected Perl libraries. The work tightly integrates with the ELIXIR Beacon project.