GOLD/EBI Air terms

[kaiiam]

Cross reference to #672. Create ecosystem terms to represent the GOLD/EBI-mgnify root > Environmental > Air classes:

root > Environmental > Air
root > Environmental > Air > Indoor Air 
root > Environmental > Air > Indoor Air > Dust
root > Environmental > Air > Outdoor Air

[kaiiam]

I propose the following terms: @pbuttigieg let me know what you think.

GOLD term: root > Environmental > Air

ENVO:air-associated ecosystem

Def:

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere.

Subclass of ecosystem

Equivalent to

ecosystem
  and ('determined by' some 
    (air 
        and ('part of' some 
            ('nitrogen-oxygen planetary atmosphere'))))

GOLD term: root > Environmental > Air > Indoor Air

ENVO:construction-enclosed air-associated ecosystem

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem
  and ('determined by' some 
    (air 
        and ('part of' some 
            ('nitrogen-oxygen planetary atmosphere'
               and ('located in' some 
                    ('part of' some construction))))))

GOLD term: root > Environmental > Air > Indoor Air > Dust

ENVO:construction-enclosed air-associated dust ecosystem

Def:

An ecosystem which is determined by air-laden dust which is part of a nitrogen-oxygen planetary atmosphere enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem
  and ('determined by' some 
    (air 
        and ('determined by' some 
           (dust 
              and ('part of' some 
                ('nitrogen-oxygen planetary atmosphere'
                   and ('located in' some 
                      ('part of' some construction))))))))

GOLD term: root > Environmental > Air > Outdoor Air

ENVO:outdoor air-associated ecosystem

Def:

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere that is not enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem
  and ('determined by' some 
    (air 
        and ('part of' some 
            ('nitrogen-oxygen planetary atmosphere'
               and (not ('located in' some 
                    ('part of' some construction)))))))

[cmungall]

Those nested determined-bys don't look right

On Fri, Aug 16, 2019 at 4:28 AM Kai Blumberg notifications@github.com wrote:

I propose the following terms: @pbuttigieg https://github.com/pbuttigieg let me know what you think. GOLD term: root > Environmental > Air

ENVO:air-associated ecosystem

Def:

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere.

Subclass of ecosystem

Equivalent to

ecosystem and ('determined by' some (air and ('part of' some ('nitrogen-oxygen planetary atmosphere'))))

GOLD term: root > Environmental > Air > Indoor Air

ENVO:construction-enclosed air-associated ecosystem

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem and ('determined by' some (air and ('part of' some ('nitrogen-oxygen planetary atmosphere' and ('located in' some ('part of' some construction))))))

GOLD term: root > Environmental > Air > Indoor Air > Dust

ENVO:construction-enclosed air-associated dust ecosystem

Def:

An ecosystem which is determined by air-laden dust which is part of a nitrogen-oxygen planetary atmosphere enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem and ('determined by' some (air and ('determined by' some (dust and ('part of' some ('nitrogen-oxygen planetary atmosphere' and ('located in' some ('part of' some construction))))))))

GOLD term: root > Environmental > Air > Outdoor Air

ENVO:outdoor air-associated ecosystem

Def:

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere that is not enclosed within a human built construction.

Subclass of ecosystem

Equivalent to:

ecosystem and ('determined by' some (air and ('part of' some ('nitrogen-oxygen planetary atmosphere' and (not ('located in' some ('part of' some construction)))))))

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/EnvironmentOntology/envo/issues/858?email_source=notifications&email_token=AAAMMOOKLATXMWPRQS7BK7DQE2FOTA5CNFSM4IMGD5OKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4OMBSI#issuecomment-521978057, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAMMOIX7CJSDRBL2JMCTZTQE2FOTANCNFSM4IMGD5OA .

[kaiiam]

@cmungall could you be more specific, I had modeled it after the patterns form the environmental system classes done for the envoMeo subset.

For example:

cnidarian-associated environment has the axiom

'animal-associated environment'
 and ('determined by' some 
    (Cnidaria or ('part of' some Cnidaria)))

[pbuttigieg]

ecosystem
  and ('determined by' some 
    (air 
        and ('determined by' some 
           (dust 
              and ('part of' some 
                ('nitrogen-oxygen planetary atmosphere'
                   and ('located in' some 
                      ('part of' some construction))))))))

@kaiiam I think the issue is that the air is not determined by dust. Determination is between a system and a material entity.

[pbuttigieg]

For

root > Environmental > Air > Indoor Air > Dust

you would first create root > Environmental > Air > Indoor Air like so:

ecosystem
  and 'determined by' some (
    air 
       and (
           'part of' some ( 
                'nitrogen-oxygen planetary atmosphere'
                   and 'located in' some 
                      'building envelope'
                )
           )
       )

Then you can handle the root > Environmental > Air > Indoor Air > Dust leveraging the above:

construction-enclosed air-associated ecosystem
  and 'determined by' some 
   dust

[pbuttigieg]

Also, "Indoor Air" would leverage air in the same way:

First, handle air-associated ecosystem

ecosystem
  and 'determined by' some (
    air 
       and (
           'part of' some
                'nitrogen-oxygen planetary atmosphere
       )
    )

then construction-enclosed air-associated ecosystem can be

air-associated ecosystem
  and 'located in' some 'building envelope'

[kaiiam]

@pbuttigieg thanks for the feedback.

@cmungall do you think this overall strategy of well axiomatized terms would be a good way of handling the GOLD/ EBI MGNIFY purls? It was what @pbuttigieg and I had discussed to be a reasonable solution, as opposed to simply mapping to a single material entity, material feature, system term etc.

[kaiiam]

This strategy would also be moving away from the advice @cmungall gives here https://douroucouli.wordpress.com/2019/06/29/ontotip-learn-the-rector-normalization-technique/ about using explicit design patterns, which is what he and I had originally setup aka the EBI biomes robot template or is it ok as long as I keep the patterns as regular as possible while still being complete, and correct as described here: https://douroucouli.wordpress.com/2019/05/10/ontotip-single-inheritance-principle-considered-dangerous/?

[cmungall]

I don't think the SI vs MI advice is relevant here.

A few things going on in this ticket, let me try and address.

pre vs post-composition: do we make named classes for everything, vs just expressing via combinations of class expressions at annotation time?

If we make named classes, should these go in the ENVO ID space, or a different one?

Whichever path we take, things should follow a consistent design pattern/schema

[kaiiam]

pre vs post-composition: do we make named classes for everything, vs just expressing via combinations of class expressions at annotation time?

In principle they could be pre-composed at annotation time, however, we're operating under the constraint that existing systems, specifically those we are serving GOLD/EBI-mgnify presumably want a single iri per annotation. Hence we were thinking my work could be to create pre-composed classes with thoroughly composed axioms as a compromise between current 1-iri systems and future multiple annotation-accepting systems.

I also think people aren't quite ready to do the post-composition themselves nor realize why it would benefit them, or have systems allowing them to do it ATM. Hence this work could be an example of providing well-axiomatized pre-composed semantics for these good repositories of existing genomics data, with the intention of it being the model for user-defined post-composition in the future. Perhaps we (the Hurwitz lab) could even deliver a software tool to facilitate that in the future.

If we make named classes, should these go in the ENVO ID space, or a different one?

Up to you guys, would a different or new Ontology namespace be in order?

[kaiiam]

New pass at air terms making use of a Robot template design schema while trying to model the terms after the patterns used in the environmental system classes done for the envoMeo subset and what was suggested by @pbuttigieg above.

GOLD term: root > Environmental > Air

ENVO:air-associated ecosystem

An ecosystem which is determined by air which is part of a nitrogen-oxygen planetary atmosphere.

C % (subclass of)	C 'determined by' some %	C 'part of' some %	C 'located in' some %	C 'bearer of' some %
ecosystem	air	nitrogen-oxygen planetary atmosphere

GOLD term: root > Environmental > Air > Indoor Air

ENVO:building envelope-enclosed air-associated ecosystem

An air-associated ecosystem which enclosed within a building envelope.

C % (subclass of)	C 'determined by' some %	C 'part of' some %	C 'located in' some %	C 'bearer of' some %
air-associated ecosystem			building envelope

GOLD term: root > Environmental > Air > Indoor Air > Dust

ENVO:building envelope-enclosed air-associated dust ecosystem

A building envelope-enclosed air-associated ecosystem which is determined by some dust.

C % (subclass of)	C 'determined by' some %	C 'part of' some %	C 'located in' some %	C 'bearer of' some %
building envelope-enclosed air-associated ecosystem	dust

GOLD term: root > Environmental > Air > Outdoor Air

ENVO:outdoor air-associated ecosystem

An air-associated ecosystem which is not enclosed within a human built construction.

Axioms not fitting regular design pattern to be added manually in protege:

air-associated ecosystem

not ('located in' some 
     'building envelope')

[cmungall]

@kaiiam - it looks like the IDs for new terms in your template are 7 digit. All new ENVO IDs should always be zero padded to make 8 digits.

[cmungall]

On the call today I made a recommendation that the values in the spreadsheet should be unfolded as far as possible.

for example, currently we have this:

                  ID: ENVO:3300090
               label: estuarine sediment microbiome
          definition: An estuarine microbiome which is determined by some sediment.
environmental feature: 
environmental material: http://purl.obolibrary.org/obo/ENVO_00002007 *sediment*
environmental system: http://purl.obolibrary.org/obo/ENVO_01001048 *sediment environment*
environmental system 2: 
             quality: 
 biome (subclass of): ENVO:3300089 *estuarine microbiome*
      EBI Biome Path: https://www.ebi.ac.uk/metagenomics/api/v1/biomes/root:Environmental:Aquatic:Estuary:Sediment
EBI Biome Path > delim: root > Environmental > Aquatic > Estuary > Sediment

it's a bit odd how we have the redundancy with sediment. Also the biome class is already a precomposed class in the spreadsheet this should be unfolded to something more atomic.

The way to think about this is what would an ordinary MIxS annotator not necessarily familiar with ENVO curate?

I think we perhaps should have went for a 3-slot system here to begin with, e.g. for this one:

system: ENVO:aquatic feature: ENVO:estuary material: ENVO:sediment

[kaiiam]

@cmungall

@kaiiam - it looks like the IDs for new terms in your template are 7 digit. All new ENVO IDs should always be zero padded to make 8 digits.

I see the issue in the ID field. Will rectify as I make the additions to the robot template. My plan is to break this work up and make issues for logical term sets such as this one for the Air terms. Get revision from @cmungall and or @pbuttigieg then push that to the robot template file.

On the call today I made a recommendation that the values in the spreadsheet should be unfolded as far as possible.

Sounds good I'll make another pass at these terms doing so following the 3 or 3+ slot system laid out above, keeping the MIXS standard in mind (which was the original motivation anyway).