POC-MorphoBank Webex meeting 07-05-2011
In attendance:
POC members: Laurel Cooper (OSU), Pankaj Jaiswal (OSU), Ramona Walls (NYBG); Barry Smith (University at Buffalo, NY), Justin Preece (OSU)
Collaborators: Maureen O'Leary
The purpose of this meeting is to discuss potential collaboration between the Plant Ontology and Morphobank.org.
Background Information @ Morphobank
Current utility and goals.
From the MorphoBank website:
MorphoBank (MB) is a web application for conducting phylogenetics or cladistics research on morphology that enables teams of scientists to work over the web - in real time - and to do research they could not easily do using desktop programs alone.
MorphoBank displays dynamic phylogenetic matrices of morphological characters with labeled images demonstrating homology statements, and implements the data editing functions of widely used desktop programs (e.g., Mesquite, Nexus Data Editor) over the web in a password protected environment. It is an environment for virtual collaboration by teams of researchers building phylogenetic matrices with affiliated image data.
MorphoBank can also draw on images in existing 2D and 3D digital libraries. If a scientist has images that are not deposited in other digital libraries, MorphoBank uses its database to store images (including films and CT scans) submitted by scientists, and allows contributors to label anatomical structures on the images. MorphoBank records metadata for the images.
MOL gave us a introduction to her site and some additional background. Started about 10 years ago to keep track of projects, media and be a way to work interactively. She commented that it would be extremely useful for have an example matrix, using the PO, in order to demonstrate its utility to users. Showed us an example of a character (x axis) by taxon (y axis) matrix. One idea would be converting an existing plant matrix to a Morphobank example. Focus on interactive nature, web based
Challenges exist in getting users to adopt ontologies, even have resistance to sharing matrices
The character states being used are basically free text, can be anything you want, not necessarily consistent across projects. Some has been done in ATOL groups, virtual, interactive nature of MB promotes this. Used for the generation of phylogenetic trees, using TNT and PAUP software.
Currently, continuous characters have to be converted to discrete characters to display, but they are working on algorithms to deal with them. See references in Journal "Cladistics".
Here is a link to a FAQ on the site with more info: Morphobank FAQ
Plant Ontology
Current utility and goals
Goals (from current grant):
- Develop and enrich the PO to describe plant anatomy and developmental stages of all plants including bryophyte, pteridophytes, and gymnosperms, as well as angiosperms
- Develop the PO as reference ontology of plant structures and growth stages by developing mappings to similar vocabularies in use by plant databases
- Outreach and education activities to extend the PO usage and awareness about Plant Biology research
- Provide the vocabulary needed for phenotypic descriptors (qualities, as well as entities), by enriching the PO or PATO with quality terms. This is needed for:
-description of mutant phenotypes or natural variation for genetic or genomic research, or for large-scale phenomic screenings
-description of systematic/phylogenetic characters
- Provide user-friendly tools for online data curation using ontologies such as the PO.
- Provide links to images for terms to aid in understanding definition.
Current uses:
Primarily used by model organism databases to associate gene expression and phenotypic variation to plant anatomical entities or plant growth and development stages.
Used in genomic analysis tools (e.g., VirtualPlant, eFP Browser) to aggregate data according to ontological rules.
Ideas and Questions re. collaboration between PO and Morphobank
Graphical interface for entering systematic data using ontologies
Morphobank could provide software to store data (specimen data, images, etc.) and create character matrices based on ontology terms.
PO could provides EQ (entity-quality) data for plants.
From MOL: "If you were to describe what would be an ideal tool or link, can you say what that would be? In other words, we are familiar and comfortable working in the matrix environment, if we were to open that matrix or character list, what tool or function would be ideal to link to PATO (and PO or other anatomy ontologies)?"
PO would like to develop a species-neutral upper-level matrix, with characters on Y axis and states on X axis. Then people could use this to annotate characters. The whole matrix would not be exposed to users. Data would be stored in a multi-dimensional matrix, but users would interact with something more like a series of cascading drop-down menus.
Ideas/Plans from PO: two prong approach:
Top-down Approach: Large character matrix, plant-based, species-independent, all anatomical entities (~500) from PO, assayed for different characters for example: leaf shape, thickness, size, color; parts of leaf e.g. tip, lobes,
Some of these matrices are already available in textbooks, etc. This could be used by the individual projects and they could request terms that are missing, or create something on their own. Encourage communication between the two levels, and between projects to find common reference points
bottom-up approaches: project-wide, e.g. Monocot or Gymnosperm tree of life projects
Three sides (dimensions) of a cube: taxonomy + anatomical entities + state;
The MB window: 'Matrix Character Viewer' would be the local place to put in ontology terms, best to put in as drop-down menus
Ontology button: place to put rules; relate traits to each other, for e.g. if this structure is absent, these traits are not relevant
Exemplar project would be very helpful- e.g. Gymnosperm TOL, premade matrix would save a lot time and effort
We need a data set to test the concept. Can DWS or MAG provide one?
Links to images
Adding links to images for PO terms is a priority of the PO now. Morphobank has an image data base. Can we set up links between PO terms and Morphobank images?
Need a stable url for individual images, in standard format, to set up as dbxref in GO database file.
Examples:
abbreviation: GR_REF
database: Gramene: A Comparative Mapping Resource for Grains.
object: Reference example: GR_REF:659
generic_url: http://www.gramene.org/
url_syntax: http://www.gramene.org/perl/pub_search?ref_id=ID number
example: http://www.gramene.org/perl/pub_search?ref_id=659
abbreviation: GR_MUT
database: Gramene: A Comparative Mapping Resource for Grains.
object: Mutants
example: GR_MUT:GR:0060198
generic_url: http://www.gramene.org/
url_syntax: http://www.gramene.org/perl/mutant/search_mutant?id=ID number
example: http://www.gramene.org/perl/mutant/search_mutant?id=GR:0060198
Other
References and Links
Dahdul et al. 2010 - Paper on using EQ statements to describe legacy data set of morphological characters for fish.
