Re: Protein representation with and without BioChemEntity

Hi,

On 06/11/2017 19:28, Anders Riutta wrote:
> Hi all,
>
>     If we are going to use a second context with aliases defined by
>     Bioschemas profiles, would it not make sense for that second
>     context be one provided by Bioschemas? If we have a Bioschemas
>     context, that context will take care of define the aliases and
>     collate the possible corresponding ontology terms. In this way we
>     keep it simple for users and avoid multiple cases (camel and
>     others), for minimum and recommended properties at least. A third
>     party property would always be possible for optional properties,
>     giving proority to any of those already coined in schema.org
>     <http://schema.org>
>
>
> I love any idea that keeps things simple! If I understand correctly, 
> this proposal would push the Bioschemas community to converge on 
> choices for both terms and IRIs, but it would still allow for using 
> different terms.

Yes, that would be the idea. I will work on a Bioschemas context for the 
protein case and will share when it is ready so we can have a better 
idea about how it would look and work.

Regards,
>
> It could look like this for annotating a pre-existing JSON API 
> <https://www.oclc.org/developer/news/2016/retrofitting-an-existing-api-with-linked-data.en.html>. The 
> Bioschemas context would specify one alias, e.g.:
>
>     { ..., 
>
>     transcribedFrom:
>     http://semanticscience.org/resource/is-transcribed-from
>     <http://semanticscience.org/resource/is-transcribed-from>, 
>
>       ... }
>
>
> But if the pre-existing API uses the term "transcfrom" then it would 
> need to use a combined context like this:
>
>     [ 
>
>       { ..., 
>
>     transcfrom:
>     http://semanticscience.org/resource/is-transcribed-from
>     <http://semanticscience.org/resource/is-transcribed-from>, 
>
>         ... } ,
>
>       "http://bioschemas.org/context.jsonld"
>
>     ]
>
>
> One point of caution: the JSON-LD spec says 
> <https://json-ld.org/spec/latest/json-ld/#advanced-context-usage>, 
> "Multiple contexts may be combined using an array, which is processed 
> in order," but that doesn't necessarily mean the Bioschemas term would 
> take precedence for the combined context above. For example, if the 
> output from the API were to be expanded and re-compacted with that 
> combined context, the result could use the term "transcfrom" instead 
> of "transcribedFrom", as specified by the term selection algorithm 
> <https://www.w3.org/TR/json-ld-api/#term-selection> and discussed in 
> this comment 
> <https://github.com/digitalbazaar/jsonld.js/issues/75#issuecomment-61841449>.
>
> Regards,
> Anders Riutta
>