[Gramene] Re: [Po-dev] Fwd: [Obo-discuss] SOY reference, format spec.

[Erick Antezana]

As promised, resending the email(s) below...

On 26 May 2010 14:31, Erick Antezana <erick.antezana at gmail.com> wrote:
> Hi Pankaj,
>
> w.r.t. our interest on PO: we have been collecting and integrating
> some publicly available ontologies (such as GO, ECO, PO, TO, ...) into
> our internal platform to support not only plant genome annotation
> (e.g. rice, maize, tomato, papaya, soy, potato, ...) but also to
> enable data integration. Our bioinformatics group supports activities
> not only in our research centre in Ghent, Belgium but also in other
> centres around the globe (e.g. India, US, South America, Canada, ...).
> We have been using some internal vocabulary/terminology since long ago
> and since some time ago, we embarked on standarising and harmonising
> the use of terms across the different groups. In that sense, we
> thought of sharing such terminology and in that way contribute to the
> plant community so that we could all "speak the same language".
> Concretely, we could contribute with new terms, synonyms, definitions,
> etc. (which in some cases could be organism-specific of course...).
>
> Regarding the approach of explicitly capturing the crop name (or
> label/subset) to which a term belongs, we also thought that is
> actually an excellent way of dealing with various crops. At a first
> glance, it could of course look a bit redundant but given the current
> technical "limitations" (i.e. querying, inferencing [1]), such
> approach is truly justifiable. I think, however, that in the future we
> could run out of "standard" labels for those subsets or simply face
> issues if we want to define a more "specific organism" (subspecies),
> in that sense, I would propose to define those labels as follows :
>
>        subsetdef: CODE   "scientific name; taxon id; common name"
>
> the CODE will basically be inspired by the UniProt convention
> (http://www.uniprot.org/docs/speclist), for instance:
>
>          subsetdef: ORYSJ   "Oryza sativa Japonica; 39947; Rond Rice"
>          subsetdef: SOLLC   "Solanum lycopersicum; 4081; Tomato"
>
> Any thoughts about such proposal?
>
> cheers,
> Erick
> [1]  Actually, it is possible to combine a taxonomy and PO and then
> run a simple transitive closure algorithm to get the common parent
> organism/genus/family, etc, etc. but that will imply an extra
> infrastructure....
>
> On 19 May 2010 18:28, Pankaj Jaiswal (OSU)
> <jaiswalp at science.oregonstate.edu> wrote:
>> Looks like an interesting discussion and
>>
>> On 5/19/2010 6:36 AM, Erick Antezana wrote:
>>>
>>> Hi Laurel,
>>>
>>> thanks for your email. Please find below (in-line) some
>>> comments/questions:
>>>
>>> On 22 April 2010 00:44, Laurel Cooper<cooperl at science.oregonstate.edu>
>>>  wrote:
>>>>
>>>> Hi Erick,
>>>> I have been slow to reply because I have been working on finding some
>>>> answers to your questions below:
>>>>
>>>> * where can I find such ontology (Stages Ontology) with that namespace
>>>> (SOY)?
>>>> See SoyBase: http://soybase.org/ontology.php
>>>
>>> I was able indeed to browse the SOY ontology (both the developmental
>>> and trait ontology using the amigo interface). However, I was not able
>>> to download them. Where can I get an OBO-formatted version of them?
>>
>> Rex nelson is copied on this mail. I am encouraging him to start submitting
>> the SOY in PO SVN. As soon as we have it in svn it will be available to the
>> community.
>>
>>>
>>>> SOY is a link to the SOYBASE growth Stages Ontology. Numbers like
>>>> SOY:0000001 refer to the cross-references (Xrefs) between PO growth
>>>> stages
>>>> and the SOY growth stage ontology.
>>>
>>>> I checked with the curator of SoyBase, Rex Nelson and this was his
>>>> response:
>>>> "I realized that I needed a Soy growth and development ontology and then
>>>> a
>>>> Soy trait ontology (since the Plant TO was so Poa centric) so I thought
>>>> that
>>>> I needed to make a new prefix for all three (SoyWGR, SoyGR and SoyTO
>>>> respectively).  On hindsight that might have been a mistake, I just don't
>>>> know right now.  Yes, the SOY prefix was unilaterally changed by me to
>>>> SoyWGR."
>>>> You should contact him directly if you have further questions.
>>>
>>> Thanks for such information. I think it would be indeed very much
>>> appreciated by the community in general if those prefixes wouldn't be
>>> there... There are other ways to make such ontology ("virtual")
>>> partitions (e.g., using subsets or namespaces, see OBO spec 1.2
>>> http://www.geneontology.org/GO.format.obo-1_2.shtml). We could discuss
>>> further if you are interested in such a refactoring.
>>
>> Rex. please take a note. I would prefer the three classes stay as separate
>> ontology.
>>
>>>
>>> By the way, where can I get the latest "cereal plant development"
>>> ontology (prefix: GRO)? The one I found dates from 2007! :
>>>
>>>
>>> http://obo.cvs.sourceforge.net/viewvc/obo/obo/ontology/developmental/plant_development/cereals/cereals_development.obo?revision=1.6&view=markup
>>>
>>>>
>> just replied to your mail on the obo-discuss list.
>>
>>>> *is the root term of the ontology stored in po_temporal.obo the term with
>>>> name "plant growth and development stages"?  Yes
>>>>
>>>> *I think that term (as well as its direct children) should not be written
>>>> in
>>>> plural ('stage' instead of 'stages'). You are probably correct
>>>>
>>>> *The synonym of the same term (PO:0009012) is "Arabidopsis growth", I
>>>> think
>>>> this is not "totally" correct.  You are probably correct here also
>>
>> The synonym came directly from the arabidopsis growth stage ontology that is
>> now obsolete.
>>
>>>>
>>>> * migrating the po_temporal file from OBO spec 1.0 to OBO 1.2:  I have
>>>> taken
>>>> the latest version of the po_temporal.obo and gotten anontology in 1.2
>>>> (you
>>>> can find it attached)
>>>> Thank you for providing that file.  I talked with Pankaj and we are going
>>>> to
>>>> migrate both PO_anatomy and PO_temporal to 1.2 after the appropriate
>>>> testing.
>>>
>>> I have just seen that there are new versions of those ontologies in
>>> 1.2.  Just some minor remarks:
>>>
>>> For the term: lenticel, the xref of its definition should be
>>> [ISBN:0471245208] instead of [ISBN: 0471245208] (no space after ':' )
>>>
>>
>> fixed see revision 829
>> http://palea.cgrb.oregonstate.edu/viewsvn/Poc/trunk/ontology/OBO_format/po_anatomy.obo?
>>
>>
>>
>>> Instead of the following lines:
>>>
>>>     def: "The end distal from the fruit pedicel." [(SGN:nm), GR:ap]
>>>     def: "The stem end of the fruit." [(SGN:nm), GR:ap]
>>>     def: "Vascular system of the fruit." [(SGN:nm), GR:ap]
>>>     def: "The space of variable size controlled by stomatal guard
>>> cells in response to environmental clues.       " [GR:cwt]
>>>
>>> it should be respectively:
>>>
>>>    def: "The end distal from the fruit pedicel." [GR:ap, SGN:nm]
>>>    def: "The stem end of the fruit." [GR:ap, SGN:nm]
>>>    def: "Vascular system of the fruit." [GR:ap, SGN:nm]
>>>    def: "The space of variable size controlled by stomatal guard cells
>>> in response to environmental clues." [GR:cwt]
>>
>> fixed see revision 829
>> http://palea.cgrb.oregonstate.edu/viewsvn/Poc/trunk/ontology/OBO_format/po_anatomy.obo?
>>
>>>
>>> On the other hand, some terms, such as "hull" (id: PO:0006000) belong
>>> to various subsets which in some cases could be sort of redundant
>>> (overlapping), such as in the case of "hull" (see below), "Rice" and
>>> "Maize" are part of the "Poaceae" family, so it is not necessary to
>>> add Rice and Maize as extra subsets... however, such extra lines
>>> actually could be very helpful while querying the ontology (avoiding
>>> some sort of inference that might take a taxonomy to infer such
>>> subsets...) Anyway, what are your thoughts about it? It would be
>>> interesting to systematically follow a defined way of capturing such
>>> information.
>>
>> I think you got the clue. Though it was an experimental aspect. We want to
>> build a mechanism where a user can search to find all the terms / a subset
>> that is relevant to rice/maize/grasses (poaceae). We would love to hear
>> about this aspect. Our approach would be to create/tag a term using either
>> the defined subset of a xref to the taxonomy_name and taxonomy_ID to find
>> terms relevant to a taxa (any level) in the taxonomic hierarchy. Its a
>> challenge for PO compared to other ontologies (except GO) that usually cater
>> to a single organism/species.
>>
>>>
>>> [Term]
>>> id: PO:0006000
>>> name: hull (sensu Poaceae)
>>> namespace: plant_structure
>>> def: "A leaf like structure often found enclosing the grass fruit,
>>> caryopsis. It is derived from the dried lemma (PO:0009037) and palea
>>> (PO:0009038) floral bracts found in the grass floret (PO:0006318)."
>>> [GR:pj]
>>> subset: Maize
>>> subset: Poaceae
>>> subset: Rice
>>> relationship: develops_from PO:0009037 ! lemma
>>> relationship: develops_from PO:0009038 ! palea
>>> relationship: part_of PO:0006342 ! infructescence
>>>
>>>>
>>>> *Is the latest version the one found at [2] ? (it dates from 2008).  Yes
>>>> it
>>>> is.  It is an active ontology, but it is currently not being revised, but
>>>> it
>>>> will be in the near future.  We are currently revising the PO_anatomy to
>>>> include all plants.
>>>>
>>>> *Could you please comment a bit on the plans you have for the future
>>>> development of this ontology? See above,
>>>>
>>>> *On the other hand, how could we (or the community in general) contribute
>>>> with some new terms/relations/fixes/etc?
>>>> If there are terms, problems etc the best approach is to open a
>>>> SourceForge
>>>> tracker item.
>>>> Here is the link:
>>>> https://sourceforge.net/tracker/?group_id=76834&atid=835555.  Click on
>>>> "add
>>>> new" and be sure to select the appropriate group.  It is best to have one
>>>> tracker per issue.
>>>> If you are interested in collaborating with development etc, please
>>>> contact
>>>> Pankaj Jaiswal who is the head of the project.
>>>
>>> Yes, I will definitely write an email to him expressing our interest
>>> in enriching the current version of PO.
>>>
>>
>> we look forward to. Also let us know know your interest on how you are using
>> the PO and in what resource.
>>
>>> All the best,
>>> Erick
>>>
>>
>>
>