Minutes of the Inaugural Meeting
of the Plant Ontology Consortium
Town & Country Convention Center, Terrace Room #3
9:30 – 12:00, Saturday, January 12, 2002 (beginning of
Plant, Animal & Microbes X genome Conference, San Diego, California,
USA)
Agenda:
• Welcome, Introductions & Apologies
• Additional items for the agenda
• Formation of the Plant Ontology Consortium
Composition
Trademark
Web address
Publications
Networking with other research groups
• Update/outcome of the NSF proposal for funding for the
POC
• Discussion & review of the ‘White Paper’ on the
proposed Plant OntologyTM Consortium
• Plans for 2002
• Other relevant matters
• Opportunity to share & review our progress on ontology
and controlled vocabulary development that various of us are working on
(Reports).
Welcome:
Leszek – Welcomed all the participants to this
“ground breaking” meeting to identify consensus through a “round table
discussion”.
participants:
• Leszek Vincent: MaizeDb/Maize Mapping Project - a plant
systematist/botanist; prime mover in compilation of the “common skeleton”
whitepaper.
• Klaus Mayer: MIPS, Munich – Arabidopsis Genome, now
European PlaNet, long involvement in “functional categories” MIPS
GO, expanding interest in plants.
• Jeff Ning: KSU, Postdoc, Biologist ? Bioinformatics;
project with Jan Leach, Rice Mutants, Scot Hulbert; own project, gene prediction
algorithm development.
• Ken McNally: GRC, IRRI – allele mining of IRRI germplasm
collection using comparative genomics.
• Junjian Ni: Gramene (Cornell Univ.), mutant control
vocabularies with Pankaj Jaiswal.
• XiaoRang Pan: CSHL, Bioinformatics postdoc, computer
background, working on Gramene, database work for Arabidopsis, especially
mutant database (Rob Martiensen?), Arabidopsis / maize.
• Koji Doi: National Institute of Agrobiological Sciences
(NIAS), Japan – construction of rice cDNA project, applying ontology to
rice cDNA.
• Kenichi Higo: National Institute for Agrobiological
Sciences, Associate Director RGP, Gene sequence – Sakata; map-based cloning,
rice knockout, proteome, cDNA project each project construct links;
cis element database (his special project), needs to link to other databases
efficiently.
• Michael Ashburner: GO representative (Drosophila),
Flybase curator, Drosophila melanogaster genome 1998 consortium, former
director of EBI.
• Susan McCouch: Cornell Univ., Gramene Co-PI; 8 years
beforehand, RiceGenes (no controlled vocabulary), systematic approach
QTL & mutant phenotypes, need to solidify traits and characteristics;
trait ontology using GO.
• Pankaj Jaiswal: Cornell Univ./Gramene; trait ontology
for plants (initially rice); contributing to development of controlled
vocabulary - specifically for monocots; GO rice protein database (8,000
proteins so far).
• Richard Bruskiewich: International Rice Research Institute
(IRRI); Bioinformatics Specialist involved in bioinformatics support of
functional genomics and germplasm characterization, especially, the International
Rice Information System (IRIS) for which gene, trait and (mutant) phenotype
ontologies are being developed.
• Lin Huang: UK CropNet (IGER - John Innes), forage grasses
mainly.
Apologies: Leonore Reiser, TAIR
Ed Coe, Mary Polacco – MaizeDB
Additional items for the agenda: No further items
were tabled.
Formation of the Plant Ontology Consortium:
Short history:
- formed through networking; 2 years ago, Leszek
Vincent met Leonore Reiser at the annual Maize Genetics Conference in Idaho;
Leonore asked Leszek about CV/O for maize?
- GO users network
- Contact with Michael Ashburner
- Pankaj @ meeting last year proposed drafting
a common skeleton doc. ( as a “white paper”, one portion of this meeting’s
agenda
- 3 core participants so far:
Gramene (CSHL/Cornell; Lincoln Stein is PI)
IRRI
MaizeDB
TAIR is closely associated with the work of the
POC.
Question: are any other core participants needed?
GO - Open meetings are “self funded” by participants,
not GO. Very little overhead, some central funding for WWW etc needed &
for central coordination; don’t want too many participants in core of POC.
Other potential database participants – keeping core group small is a pragmatic
not autocratic decision.
Trademark – LV contacted GO (Mike Cherry); Recommended
to use unregistered trademark (TM) adjacent to “Plant Ontology” i.e. Plant
OntologyTM.
Pankaj – TM needs to be traded on the commercial
market before it can be registered as such; currently, only have a copyright
(C with a circle). It should be enough.
KM (MIPS) - databases are copyrighted.
Trade Mark(TM) considered by some to be inappropriate,
COPYRIGHT or SERVICE MARK better?
Leszek to follow-up. [Follow-up: A trademark is
a name, symbol, word, or combination of 2 or more of these that a person
places on goods that the person sells or distributes, a container of the
goods, a display associated with the goods, or a label or tag affixed to
the goods, to identify those goods that the person makes or sells, and
to distinguish them from goods that another person makes or sells. Use
of ‘TM’ for Plant Ontology is in reference to the distribution of the products
of the POC. Consequently Plant OntologyTM appears appropriate. Confirmed
by Legal Office at UM-C.]
- Gramene already has a WWW page (www.gramene.org)
with links to plant ontology info.
- The SPRIG site (Specialized Plant Resources for
Informatics and Genomics) at www.bioinformatics.org/sprig/poc
has some provisional info. on the POC prior to the development of the POC
website.
- Gramene also has some web & data resources
which will be shared.
- MaizeDB has purchased www.plantontology.org
for 10 years.
- Publications – would like to publish: RMB – contacted
by Jo Wixon @ John Wiley & Sons; will be seeking PAGM info. for publication
in Comparative & Functional Genomics journal.
- Networking with other research groups: - word
of mouth.
Update/outcome of the NSF proposal for funding
for the POC
Grant proposal submitted to NSF in June 2001 (RCN
program: Research Coordination Network). 21 proposals submitted in
2001, 7 accepted for funding, ours was not, but was recognized as a valuable
and necessary.
Panel summary:
- POC, Excellent idea
- Need “proof of concept” data (live prototyped
database with interfaces)
- Need to identify “de novo” activities;
a clear action plan involving the GO consortium
Susan McCouch – need to actually start doing some
PO work, get a track record of activity & credibility & then reapply
to NSF for funding (by 28 June 2002). Conclusion: let’s get started on
the “real” work.
Discussion & review of the ‘White Paper’ on
the proposed Plant OntologyTM Consortium
Michael Ashburner – philosophical viewpoints…
1. Major GO mistake: Spent a lot of time, but not
enough, on semantics and on the precise meaning of what a “term” in GO
really is. Ashburner & Lewis have written a paper on semantics of terms
& relationships between terms. This topic is not addressed in initial
GO papers.
2. GO terms are equivalent to concepts, not a physical
substance, but a concept-subconcept, semantic relationships, 2 relationships
currently used in GO (ISA, PARTOF); not used consistently from beginning
of GO work.
3. Read “WordNet” article – reference, by MIT press,
a lexical database, early chapter is very instructive & provides good
info. on relationships (ISA & PARTOF).
4. ISA and PARTOF relationships (represented by
% and < symbols respectively in syntax) are very complex: in the cellular
component, usually means component of an organelle. In process, is a subprocess.
Cognitive scientists distinguish half a dozen “PARTOF” relationships.
5. Need for work on mutant phenotypes:
Michael Ashburner provided the following table
representing the skeleton of thoughts regarding recording descriptive info.
in ontologies.
| Noun |
|
|
|
|
| |
have |
|
|
|
| |
|
determined by |
|
|
| Attributes |
|
|
by assay |
|
| |
have |
|
|
Done under |
| Values |
|
|
conditions |
|
| |
have |
|
|
|
| Qualifiers |
|
|
|
|
• GOBO (Global Open Biology Ontologies) – overall family
of biological ontologies (get from MA)
Would be the following
things:
Web site:
Best practice guidelines:
Open source
Share a syntax, tools – DAG, Amigo, DAML-OIL
Namespace
CVS repository
Namespace server for prefixes (XML…
Gene
Gene Attribute
Gene Structure
Variation (EBI)
Gene product
GO
Protein family (Interpro)
Chemical
i) Biochemical substance (LIGAND)
ii) Pathway (KEGG, etc)
Development
{models} – GO/temporal
Anatomy {models}
i) Gross
ii) Tissue
iii) Cell-type
Phenotype
Mutant
Trait
Pathology
Disease
Experimental
conditions {MGED (MIAME/MAGE-ML), MIAFG}
Taxonomy
{NCBI - TAXMAN}
Cross-Product
methodologies to combine ontologies, select two subgraphs, make cross product
in DAG-Edit, inheritance rules?
‘WHITE’ PAPER - Discussion:
-
Mike A. – use database abbreviations in GO
-
Mike A. – use NCBI THXMAN for taxonomy
-
Pankaj, Susan – Use accepted botanical terms
-
Mike A. – GO cannot have “same string” but not same identifier;
identifier & definition; “sensu” only used for concept with identical
lexical strings but in different domains, not used to restrict the domain
of application, but lexical use of the term; David Botstein argument early
in GO suggested that taxonomic restrictions in GO are not required
- Leszek – “true path route”? Mike
– yes…
-
Michael A., Richard B. (Pankaj) – anatomy and morphology
are synonymous, a continuum.
-
Susan – languages other than (American) English?
-
- Mike – No. Multilanguage thesauri?
-
Susan –
- Three dimensional (feature space, scale,
relative (no reference)) matrix: exhibiting how mutants differ from wild-type:
- Spatial
- Ectopic, extra anatomical parts
- Temporally:
flowering time
- Quality (attributes)
more intense color, aroma, QTL,
etc.
????
- Lesion mimic?
Reports
GRAMENE report – handout provided. 3D feature
space…More info at http://www.gramene.org/plant_ontology/
IRRI – embedding GO schema into IRIS; moving IRIS
terms into database (with Gramene); hired an MSc. Biologist to help so
more productivity.
MaizeDB:
• Worked on white paper
• Working on Maize-specific ontologies
• Discussion with TAIR: bottom-up approache – work on
taxon specific needs first, then merge.
• Working with John Richter – discovering bugs
with DAG tool.
MIPS: Phenotype database for Arabidopsis, 1200
described AT phenotypes, gene names
Post-meeting comment:
1. The discussion & review of the 'White Paper’
on the proposed Plant OntologyTM Consortium must be considered to be largely
incomplete. It may be difficult to provide an opportunity to explicitly
‘ratify’ this document amongst POC collaborators.
![[Home Page]](/images/poc_logo.gif){kind=logo}
