FAQ
From HAO Wiki
General
What is an ontology?
- See a brief overview here.
Is this a glossary?
- No. We can extract certain parts of the HAO to act as a glossary, but the whole of the project is a fundamentally different construct.
How does an ontology differ from an glossary?
- One of the major differences is that we add layer of logic over concepts (= things you might find in a glossary). This level of logic puts certain restrictions on what is in the HAO, but it also provides a new level of utility, i.e. we can reason across the ontology to make inferences.
What is an ontogeny?
- An ontogeny is not an ontology, but the two are sometimes confused when "ontology" is brought up.
HAO Content
How can I suggest a change or new data for the HAO?
- A simple email to our listserv will suffice. We ask that you please include a published reference for your proposed changes. See How do I contribute?
I think you've misinterpreted my publication within the HAO, how can this be fixed?
- See 'How can I suggest a change or new data for the HAO?'.
How do you deal with homology?
- We discuss this in part in the PLoS ONE paper (see links below). In short, while it sounds counter-intuitive, we explicitly try to avoid invoking references to homology in the circumscription of the concepts in the HAO. To understand why this is so it's important to remember that perhaps the most important role of the HAO is to allow users to identify particular anatomical features, i.e. the first step in making a homology hypothesis prediction. Only when we can identify some "thing" on an actual insect can we move forward to formulate a comparative statement. In some ways one can think of the HAO as a set of engineering specifications. An analogy- if you were to build a new house, you would require a set of plans, an actual blueprint- you wouldn't be able to accomplish the building of house by saying "I want a house that is homologous_to that one". The HAO seeks to provide the foundation from which more robust homology hypotheses can be created. Homology statements can be derived with reference to the HAO by making statements like "In my taxon the HAO concept 1234 is homologous to your taxon's HAO concept 4567." Note that we could also build up complex statements: "In my taxon, the HAO concept 1234 that is located on HAO concept 4567 is homologous to the HAO concept 1234 located on your taxon's concept 8910".
How do you determine the preferred term for a concept?
- In short, we don't, or at least making the distinction of a "preferred-term" is not necessary for the day-to-day curation of the ontology. One of the strengths of an ontology such as the HAO is that we allow concepts to be referenced in the absence of labels- what you want to talk about is something completely separate from what you want to call the thing you want to talk about. There is however, some desire to have a set of preferred terms. To facilitate the "anointing" of such terms an ongoing discussion has been initiated on the listserv, and we plan to document this process, and possible adoption mechanisms on the wiki page linked above.
How do you choose among synonymous terms?
- See "How do you determine the preferred term for a concept?".
What is a "class" or "concept" in the HAO?
- A class is the basic data point in the HAO. A class is a representation (one can think of it as a definition, blueprint or perhaps an assay) of an anatomical structure. Each class is defined by a definition in genus-differentia format. Class are tied to one another via logically-based/defined relationships. Classes in the HAO are taxon-agnostic. An individual hymenopteran has an "instance" of a class if the author studying that individual can read the circumscription of the class (i.e. it's definition and relationships) and it matches what they are presently observing.
What is a "label" in the HAO?
- A label is a word or words that reference a particular anatomical concept (structure/character/class).
What is a "sensu" in the HAO?
- A sensu ties a label to a class via a reference.
What is a "relationship" in the HAO?
- A relationship in the HAO is a logical connection between two concepts.
What logical relationships are you currently using?
- is_a, part_of, and attached_to
Where do you get your terms (=labels)?
- We mine the literature both manually and with extraction/analyzing tools built into mx, review pre-press publications, and propose new labels when necessary to clarify concepts. We also take proposals to our listserv, and welcome individual suggestions by email or otherwise are very welcome.
How extensive is your coverage?
- TODO
How often is the HAO updated?
- There are two categories of updates:
- Day to day development of the HAO, these updates happen almost every day. You can track it's development via its "pulse", where you can also subscribe to an RSS feed.
- Official releases to the OBO foundary. These releases happen typically every several months and are typically correlated with major additions or edits.
Use
How do I get involved?
- See here HAO_to. It's best to start by posting by contacting the community through the list-serve.
What is the hymglossary?
- The hymglossary is a public front end to the HAO. Through this interface you can access nearly all of the information in the underlying database. Some housekeeping data (e.g. todo tags) may not be visible via the hymglossary.
How do I cite the HAO?
- See here.
How do I cite the NSF ABI grant funding the HAO?
- NSF DBI 0850223 to Andy Deans
Building ontologies
What software did you use to build the HAO?
We primarily use mx, with validation checking in OBO Edit, and more recently, Protege.
Can I use that software?
Yes. All the code used to develop the HAO is open source. The application is at http://mx.phenomix.org. Several other libraries have spun out of building the HAO, see the Ruby libraries here in particular for an OBO parser and associated utilities.
