Summary of discussion

[Vincent, Leszek]

Hi colleagues

Here's a summary of the items discussed during our phone conference last week (05/07/'02) - not presented in order of importance or chronological order:

1. Development of a site-neutral ontology browser - Pankaj mentioned that this is to be developed in the near future (during the next few weeks?).

2. Possibility of a phenotype workshop & small POC meeting after the ISMB meeting in Canada - with Richard B. & Michael A. Most folk would be otherwise occupied at this time (& not attending the ISMB meeting) & so would not be available to attend. Reaffirmed the useful role of phone dialogue as a stop-gap.

3. Comments on Temporal ontology:
- consider providing more explicit wording for Root;

- consider providing the scientific name for each taxon referred to - species name if possible but at least generic name. This would facilitate users of ontology using taxon-appropriate controlled vocabulary because taxa would be identified by internationally recognized scientific names (according to the  international code of botanical nomenclature (ICBN) principles). Common name can be retained to facilitate recognition - e.g. arabidopsis (Arabidopsis thaliana Heynh.) & sorghum (Sorghum bicolor (L.) Moench.). This would also contribute to credibility of ontology information. Furthermore, it would enable users to discriminate between closely associated taxa such as S. bicolor & S. propinquum which might be separately referenced in our ontologies, in time;

- Consider standardizing the citing of curatorship info. Currently the format is 'GR:pj37'. Consider 'Curator:Gramene,pj37'. Subsequent comment has been that curatorship info. need only be provided where published term definitions are modified by a curator. The latter is a less-tedious route;

- 'PMID' used in Dbxrefs - consider using 'PUBMED' as in list of acceptable abbreviations;

- The printed reference info. should be provided in the first instance. Where a URL can be cited for a term definition this should not be used in place of the printed source but can be used in addition to the printed source.

4. Leonore would email a copy of the scanned glossary file from Esau (Anatomy of Seed Plants) - around the group.

5. Noted that significant problems existed in the anatomy ontology which comprised the TAIR anatomy ontology 'merged' with some Gramene ontology......

6. Noted that the Temporal ontology (Development/Growth) presented a dicot / monocot dichotomy which was undesirable. Preferable to have dicot / monocot affinities emerge later in the ontologies (if at all?). Is it necessary to cite this dicot / monocot perspective?? The means by which we represent dicot / monocot associations (if desired) will need some puzzling over. Noted some confusion between these taxon-specific ontologies & the desire to produce a generic ontology which leads to taxon-specific ontologies (I think that's what I perceived or did I get that wrong??). 

[Post-meeting comments: what's the intention of the Temporal ontology - generic or taxon-specific or both?? If it is intended to be generic, in the non-taxonomic sense, and taxon-specific, then it would seem that "sensu ..taxon name.." would need to be used for all the taxon-specific info. At present, taxon-specific ontologies are presented]. 

7. Regarding the 'merged' anatomy ontology - Noted some significant points:
a) parent-child 'lineages' need to be tested for adherence to the True Path Rule (http://www.geneontology.org/doc/GO.usage.html);

b) affirmed that testing for adherence to the True Path Rule (TPR) is time consuming & necessitates considerable application & interpretation of basic botanical (& biological) anatomical 'systems';

c) affirmed that it is essential that the content of all of our ontologies be rigorously tested for adherence to the TPR so as to ensure scientific accuracy of the content. This is a very time consuming task but is unavoidable if we are to gain the confidence of potential users of POC ontology products;

d) suggested that ontologies should not be publicly available until each is considered to mirror the current understanding of botanical knowledge domains e.g. anatomy, temporal (development/growth) & have been tested for adherence to the TPR;

e) noted that the anatomy ontology had some word substitutions - different words associated with the same definition (I think we referred to this as 'word duplications' during the ph.call) - consistency needs to be achieved here;

f) certain words were used as terms & not as synonyms;

g) noted that some of the original TAIR definitions had been improperly treated ('partially corrupted');

h) suggested that the "Commit" & "Commit as New Term" buttons were not being consistently used & that perhaps this had contributed to the corruption noted;

i) decided that we would consider how to correct the anatomy ontology that Pankaj had developed - either by Pankaj re-developing the ontology, by building it on the TAIR's anatomy ontology, or by us all evaluating the anatomy ontology & passing on comments to Pankaj. 

[My 'take' on this is that the original developer would hopefully do the redeveloping. Once the difficulties cited in paragraphs e, f, g & h above have been resolved the rest of us should get involved with testing for adherence to the TPR - either by us each reviewing the entire structure or by division of the structure for testing of components amongst us. My preference would be for us each (or as many as possible of us) to check the entire structure for TPR adherence. Either way this will be tedious but I think that's what putting together an excellent product will take].

8. Some other points which I think I raised:
- DAG-Edit arranges terms in alphabetic & numerical order - it would be important to avoid this style of arrangement in browser views i.e. 'children' under a 'parent' should be arranged in a biologically meaningful order. Pankaj thought this could be achieved via the browser s/ware (I hope so as this biologically meaningful ordering of terms will probably be anticipated by most users of our ontologies);

- Developmental stages of Zea need to include 'milk stage' (maizeDB ID# 67373);

- the word 'wither' in TAIR-based ontology could be associated with the word 'senesce';

- typos in some of the TAIR-based definitions;

- Dbxrefs for some synonyms are missing;

- some text absent for some synonyms e.g. for V12-V14 for Zea;

- consider citing the URL for the online medical dictionary for term definitions (secondary reference). The URL is: http://cancerweb.ncl.ac.uk/omd/index.html

9. Leszek mentioned that he hoped to be able to resubmit a proposal to the NSF for funding for the POC.

Subsequent news is that the perceived hurdles on the UM-C campus have been overcome & I've been approved to serve as PI on this proposal. The 'wheels' are already in motion...

How can you contribute to this proposal?? Should you have some 'gut' or considered feedback based on last year's submission then please don't hesitate to share it with me (I can resupply you with a copy of the previous proposal). 

At the same time I think we can all affirm that we have moved considerably down the road in the development of the POC. 

This resubmission to the NSF is due for the last week of June. 

All in all we have even more stimulation to make further good progress on areas we're currently working on. The consequence being that the reviewers of this proposal will have good evidence of 'proof of concept'.

Apologies that this summary 'note' is so long but I'm motivated to capture the detail of our dialogue so as to foster open & constructive collaboration.

Regards,
- Leszek Vincent
xoxoxoxoxoxoxoxoxoxoxoxoxoxoxox
P. Leszek D. Vincent Ph.D., FLS
Plant Science Unit, Dept. of Agronomy, 209 Curtis Hall,
University of Missouri-Columbia, Columbia, MO 65211-7020, USA. Ph: (573) 884-3716 (Agronomy), Fax:(573) 884-7850;
Ph/Fax (Home): (573) 441-1228; 
Email: Leszek at missouri.edu
Plant Systematist on the Maize Mapping Project - NSF award 9872655 -
(http://www.cafnr.missouri.edu/mmp/ and  http://www.agron.missouri.edu/)
xoxoxoxoxoxoxoxoxoxoxoxoxoxoxox