search

EnvironmentOntology / envo

A community-driven ontology for the representation of environments
http://www.environmentontology.org
Creative Commons Zero v1.0 Universal
132 stars 51 forks source link

Map to NERC VS #731

Open pbuttigieg opened 5 years ago

pbuttigieg commented 5 years ago

In https://github.com/EnvironmentOntology/envo/pull/728, created during a BODC / NERC-VS, BCO-DMO, and ENVO meeting at the Marine Institute in Dublin, we demonstrated how we would map ontology terms to the SKOS resources in the NERC VS.

We'd like to do this en masse as has been done for the SWEET-ENVO mapping by @cmungall: https://github.com/cmungall/sweet-obo-alignment

Indeed, the NERC Vocabs can be mapped to UBERON and several other OBO resources too.

@gwemon could you suggest which vocabs we should target first? Things that deal with environmental entities (geographic features, materials, etc) and anatomy would be a good place to start.

Ping: @gwemon @ashepherd @adamml

cmungall commented 5 years ago

Is there a link to download the skos?

Looks like I could query it via sparql https://www.bodc.ac.uk/resources/products/web_services/vocab/ but slightly more convenient if there is a download link (preferably stable URLs for different versions)

pbuttigieg commented 5 years ago

@gwemon?

I know your collections are here, but what's the right way to get the SKOS?

alko-k commented 5 years ago

Hi @pbuttigieg if you click this: view-source:http://vocab.nerc.ac.uk/collection/S27/current/ it will give you the skos. Or if you set your browser/software to receive rdf/xml. Is this what you want?

gwemon commented 5 years ago

@pbuttigieg @cmungall You could start with the following: Matrix domain entities: http://vocab.nerc.ac.uk/collection/S21/current/ Matrix phase entities: http://vocab.nerc.ac.uk/collection/S23/current/ The one mentioned by Alex above is our chemical substance vocab: http://vocab.nerc.ac.uk/collection/S27/current/ The physical entity name and its subgroups: http://vocab.nerc.ac.uk/collection/S18/current/ and http://vocab.nerc.ac.uk/collection/S19/current/ The datum vocab: http://vocab.nerc.ac.uk/collection/S20/current/ The one for organs and various body parts is: http://vocab.nerc.ac.uk/collection/S12/current/ Any questions pls let us know.

pbuttigieg commented 5 years ago

Thanks @alko-k @gwemon! I think that will work, we'll ping you here if there are any issues

cmungall commented 5 years ago

Ah, I see, conneg is used so I can just download the SKOS RDF easily, this is v clear, thanks!

On Thu, Mar 21, 2019 at 10:30 AM Gwen Moncoiffé notifications@github.com wrote:

@pbuttigieg https://github.com/pbuttigieg @cmungall https://github.com/cmungall You could start with the following: Matrix domain entities: http://vocab.nerc.ac.uk/collection/S21/current/ Matrix phase entities: http://vocab.nerc.ac.uk/collection/S23/current/ The one mentioned by Alex above is our chemical substance vocab: http://vocab.nerc.ac.uk/collection/S27/current/ The physical entity name and its subgroups: http://vocab.nerc.ac.uk/collection/S18/current/ and http://vocab.nerc.ac.uk/collection/S19/current/ The datum vocab: http://vocab.nerc.ac.uk/collection/S20/current/ The one for organs and various body parts is: http://vocab.nerc.ac.uk/collection/S12/current/ Any questions pls let us know.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/EnvironmentOntology/envo/issues/731#issuecomment-475328636, or mute the thread https://github.com/notifications/unsubscribe-auth/AADGOe-Pj8D2mRb3h6MsZfvxWpE-D7Biks5vY8G8gaJpZM4cAQIy .

mabablue commented 5 years ago

hi all, I would like to draw your attention to the Interoperability of observable properties description group at RDA. We are aiming to get endorsed as a WG. @gwemon is co-chairing this group, but I really like @adamml, @adamshepherd and @cmungall to look at this: https://rd-alliance.org/groups/harmonizing-fair-descriptions-observational-data-wg and eventually subscribe. Right now we are deciding about meeting days here: we will meet twice a month (first and third week of the month), first week: European/American friendly (18.00 CEST, Vienna/9:00 PDT, San Francisco). Please check doodle: https://doodle.com/poll/6u7469nr5sf3agsu to choose the week day third week: European/Australian friendly (9:00 CEST. Vienna/18:00 AEDT, Melbourne). Please check doodle: https://doodle.com/poll/gb45cmsy74pnwmsg to choose the week day !! Would be great to have Adam Leadbetter involved as we are thinking to reuse the Complex Property Model here... thanks barbara (barbara.magagna@umweltbundesamt.at)

ashepherd commented 5 years ago

thanks @mabablue for drawing this to our attention. we were sorry to miss the BoF, but i've joined the RDA group

adamml commented 5 years ago

@mabablue thanks for flagging this.

I'll take a look at the teleco times and see if I can join. It'll probably only be one a month due to other "real-life" commitments.

On Wed, 10 Apr 2019, 20:51 Adam Shepherd, notifications@github.com wrote:

thanks @mabablue https://github.com/mabablue for drawing this to our attention. we were sorry to miss the BoF, but i've joined the RDA group

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/EnvironmentOntology/envo/issues/731#issuecomment-481817384, or mute the thread https://github.com/notifications/unsubscribe-auth/ADNjW6LNWEPa6Zrc3rSZ1jgpNrEgm1X6ks5vfjKngaJpZM4cAQIy .

pbuttigieg commented 5 years ago

Thanks @mabablue - the mapping we're doing here can be an example of hard-wiring interoperability across terminology resources. Let's see if we can make this a common practice and a precursor to co-development as standard!

pbuttigieg commented 5 years ago

Work in Progress! This comment will be edited.

Mapping to S23

The idea of "spheres" in general seems a bit alchemical, and is not likely to be maintained in the ENVO architecture. Further, many of the definitions do not actually describe phases, but portions of physical bodies themselves (e.g. "gaseous" is actually referring to the gas parts of the atmosphere). The semantics should be stable, however, as I'm aligning by the definitions.

NERC term OBO IRI S23 IRI Notes
aerosol http://purl.obolibrary.org/obo/ENVO_01001652 http://vocab.nerc.ac.uk/collection/S23/current/S23C005/ 1
amorphous http://purl.obolibrary.org/obo/ENVO_01001646 http://vocab.nerc.ac.uk/collection/S23/current/S23C014/ 2
biogenic carbonate http://purl.obolibrary.org/obo/ENVO_01001644 http://vocab.nerc.ac.uk/collection/S23/current/S23C021/ 3
particulate http://purl.obolibrary.org/obo/ENVO_01001651 http://vocab.nerc.ac.uk/collection/S23/current/S23C003/ 4
gaseous http://purl.obolibrary.org/obo/ENVO_01001645 http://vocab.nerc.ac.uk/collection/S23/current/S23C013/ 5
colloidal http://purl.obolibrary.org/obo/ENVO_01001647 http://vocab.nerc.ac.uk/collection/S23/current/S23C004/ 6
dissolved plus reactive particulate NA http://vocab.nerc.ac.uk/collection/S23/current/S23C010/ 7
heavy fraction http://purl.obolibrary.org/obo/ENVO_01001649 http://vocab.nerc.ac.uk/collection/S23/current/S23C011/ 8
light fraction http://purl.obolibrary.org/obo/ENVO_01001650 http://vocab.nerc.ac.uk/collection/S23/current/S23C012/ 8
non-algal particle NA http://vocab.nerc.ac.uk/collection/S23/current/S23C007/ 9
non-carbonate NA http://vocab.nerc.ac.uk/collection/S23/current/S23C015/ 10
not applicable NA http://vocab.nerc.ac.uk/collection/S23/current/S23C008/ 0
pigment http://purl.obolibrary.org/obo/CHEBI_26130 http://vocab.nerc.ac.uk/collection/S23/current/S23C009/ 11
shell and shell fragment http://purl.obolibrary.org/obo/UBERON_0006612 http://vocab.nerc.ac.uk/collection/S23/current/S23C018/ 12
unknown NA http://vocab.nerc.ac.uk/collection/S23/current/S23C001/ 0

Notes

The notes below clarify why S23 terms may be aligned to ENVO terms with somewhat different labels. In all cases, the semantics reflect the S23 definitions, rather than their label.

  1. These are values of a categorical variable rather than references to bona fide entities. We wouldn't create classes in ENVO for these.
  2. The definition in S23 specifies that this is part of an atmosphere, rather than any aerosol.
  3. The definition in S23 doesn't sound like a quality/property, but references the actual solids which are amorphous.
  4. This has been axiomatised with GO:biological_process and CHEBI carbonates.
  5. This class has been linked to the hydrosphere via parthood
  6. As mentioned above, this term's definition does not describe the phase, but the gaseous portions of the atmosphere .
  7. As with "particulate", this class is hard linked to the hydrosphere
  8. The definition of this class is hard to work with due to ambiguity, holding off for the moment
  9. The relational semantics here (relative to bromoform) are currently tricky to express with OBO conventions (to my knowledge), but a discussion on this is ongoing. Nonetheless, the classes are present.
  10. A strange class. It's a complement of a fuzzy and undefined taxonomic classification. It seems to be referring to just a single particle and the definition isn't helpful. Unlike "particulate", this is part of the "solid sphere" (sediments etc, but I'm not sure what SPM is). Also, note that "algal" is taxonomically scattered. ENVO has a temporary class for "alga" in the NCBITaxon hierarchy, but this should be ceded to PCO or ECOCORE. My feeling is that classes like this shouldn't be added (especially because of their definition by complement).
  11. I don't think we can work with this class - it's the complement of the output of a method, and not very useful for annotation.
  12. The listed CHEBI class should be used.
  13. The listed UBERON class should be used for shell. The union class for shell + shell fragement will be created in the next PR.
pbuttigieg commented 5 years ago

Note to self: shell fragmentwas a straggler from the previous S23 mapping, be sure to catch now that UBERON shell is imported.

gwemon commented 5 years ago

Great work @pbuttigieg! I will have a look at your comments. Note I could not access the ENVO links. It returns term not found.

pbuttigieg commented 5 years ago

Great work @pbuttigieg! I will have a look at your comments. Note I could not access the ENVO links. It returns term not found.

Hi @gwemon the PURLs are not live yet - they will be activated shortly after the release. Right now, they are "reserved" for the classes.

If you want to view the classes, you can open the envo-edit.owl file in our src/envo/ directory. That's our pre-release dev space.

dr-shorthair commented 5 years ago

How will these mappings be packaged? Will it be in a separate graph? Is there an RDF representation of the mappings?

pbuttigieg commented 5 years ago

Hi @dr-shorthair

For the moment, I'm adding them as dbxrefs on the respective ENVO classes and will then request NERC to do the same.

I favour this over an external file as this is a "hard mapping" done by curators and it should be advertised with the classes themselves.

xref https://github.com/ESIPFed/sweet/issues/126

We certainly can release an additional RDF file, however.

cmungall commented 5 years ago

Partially dependent on https://github.com/ontodev/robot/issues/312

gwemon commented 5 years ago

@pbuttigieg Regarding "dissolved plus reactive particulate" , this is effectively "dissolved" i.e. the "liquid phase". However this is the theory and in practice the "purity" of the liquid phase will be dependent on the method used to separate the liquid from the solid phases. When I was at uni the accepted cut-off in oceanography text books was 0.45um. But some people used 0.4um filters and in the field GF/F (nominal pore size of 0.7um) or even sometimes GF/C (even coarser) was used . Still anything measured in the filtrate would be labelled "dissolved". So this is why we also define the filter size/type in a separate vocab (http://vocab.nerc.ac.uk/collection/S24/current/) as part of our semantic model.

gwemon commented 5 years ago

Regarding "pigment" and "non-algal particle" I need to double-check the facts behind these with Roy Lowry. The terms were used to build P01 parameter codes for the light absorbance at various wavelength of material filtered and treated with various chemical to try and separate the pigmented from non-pigmented particles. I think they can be ignored for the time being.

Also SPM stands for suspended particulate material (we'll need to tidy those definitions at some stage!).

gwemon commented 5 years ago

Finally for "shell and shell fragment", I am pretty sure that this is a misplaced term. It should not be in S23 bus in http://vocab.nerc.ac.uk/collection/S18/current/.