Informational endpoint for variant parameters

[mbaudis]

While the /filtering_terms endpoint provides a standard way to inform about terms being used/queryable for a beacon there is currently no information about supported genomic variation query parameters or the possible values for some of them. Especially w/ view towards aggregators & networks it would make sense to provide this information through a dedicated endpoint or as part of info.

Examples:

referenceName

  RefSeqId:
    description: >-
      Reference sequence id for genomic reference sequence in which variant
      coordinates are given, e.g. "refseq:NC_000009.12" for human chromosome 9 in the
      GRCh38 assembly. The use of the assembly specific RefSeqId is recommended
      although alternatively names, synonymous or aliases e.g. "chr9" could be
      used in conjunction with an `Assembly` parameter.
    type: string
    example:
      - refseq:NC_000009.12
      - chr9
      - NC_012920.1

... is rather polymorphic - it is not clear what a instance (has to) support(s) - in Progenetix we support 9, chr9, refseq:NC_000009.12, NC_000009.12 and CM000671.2 ... for human chromosome 9, GRCh38; but nothing else (and I guess other implementations may not even go that far).

---

variantType

      description: >-
        The `variantType` is used to query variants which are not defined through
        a sequence of one or more bases using the `alternateBases` parameter.
        Examples here are e.g. structural variants:
        * DUP
          - increased allelic count of material from the genomic region between
            `start` and `end` positions
          - no assumption about the placement of the additional sequences is being
            made (i.e. no _in situ_ requirement as tandem duplications)
        * DEL: deletion of sequence following `start`
        The Beacon model is not prescriptive with regard to the values allowed
        for `variantType` with use of extended types (such as from `EFO:0030063`)
        being permitted. However, a support for the basic CNV types above - where
        represented in the data - is recommended.
      type: string
      examples:
        - EFO:0030070
        - DUP
        - DEL
        - EFO:0030069

Which variant types and which vocabularies for expressing them are supported by a beacon?

---

start, end and their use

Does the beacon support range (1 start, 1 end) or bracket (2 starts, 2 ends) queries?

---

Extended / alternative variant parameters

Does the beacon support the use of:

*variantMinLength, variantMaxLength

• mateName
• geneId
• aminoacidChange
• genomicAlleleShortForm

---

[redmitry]

Hi all,

This is the problem of polymorphic types definition in JSON schema.
genomicVariations endpoint points to the "common" parameters which is just JSON Object
We have (syntactically erroneous) requestParameters.json defined in the model, but because of polymorphism these are not reflected. I agree that logical solution would be having ´/parameters´ endpoint along the ´/filtering_terms´ ... or no parameters at all.
Actually, from the implementation point of view I see no much difference between parameters and filters. We can use filters as they were parameters.

Cheers,

Dmitry

[mbaudis]

Actually, from the implementation point of view I see no much difference between parameters and filters. We can use filters as they were parameters.

Exactly.