Statement to describe the efforts of SBO to attain OBO Foundry status
This document describes the current status of the Systems Biology Ontology with regard to its membership of the OBO family of ontologies. More information about the purpose and use of SBO can be found here.
The OBO initiative is an open, community-level collaborative effort to create and apply standardised methods in ontology development. Authors of ontologies belonging to this effort are committed to maintenance and continual improvements to their ontology, based on community feedback and advancements in their scientific field. More information on how to join this effort can be found here. Practical information, and previous meeting details can be found on the OBO Foundry wiki page.
SBO as a 'candidate' ontology
Initially, ontologies that are accepted into the OBO initiative are regarded as 'candidate' ontologies. This reflects the committment of the authors of that ontology to continual improvement of their ontology, according to OBO principles. A list of current candidate ontologies is presented on the second table of this page. The Systems Biology Ontology (SBO) is an OBO Foundry candidate ontology.
Moving from candidate to Foundry membership
The OBO Foundry has established a set of principles for the development of scientific ontologies, to assist in the creation of interoperable and orthogonal reference ontologies in the biomedical domain. Sufficient compliance of a candidate ontology with these principles allows the candidate ontology to become a fully fledged 'Foundry' ontology. The analysis of the compliance level of a candidate ontology with OBO principles is carried out as part of a formal review, usually by an OBO Foundry co-ordinator. The current OBO Foundry ontologies are listed in the first table on this page.
The OBO Foundry review of SBO
The third annual OBO Foundry workshop, hosted at the EBI, reviewed a number of candidate ontologies, including SBO. The formal review process comprised the presentation of an impact assessment of the ontology itself, followed by the findings of the reviewer regarding the compliance of the ontology with the OBO Foundry principles. The review for SBO was carried out by Alan Ruttenberg on Monday 15th February, 2010. The SBO impact assessment can be downloaded from the OBO Foundry wiki or from the EBI. Similarly, the review document can be found on the OBO Foundry wiki or from the EBI.
Summary of OBO Foundry review findings
A summary of the seven main issues highlighted in the review are listed below, together with links to actions we are currently taking, or plan to take shortly:
- Unique content - where terms are not unique to SBO, they should be imported from other ontologies.
- 'Is_a' relationships - need to review some subclass relationships.
- Placement of obsolete terms - some terms are incorrectly positioned in the ontology.
- Orthogonality - identify and handle orthogonal content present in other ontologies, e.g. SO, GO, PATO.
- Representation of unknowns or uncertainty - need to better characterise terms, or else remove them.
- Multiple inheritance - handle asserted multiple inheritance better.
- Expose MathML relationships - relationships encoded in MathML should be more directly available in the ontology itself.
Current and forthcoming remediation steps can be found here.
SBO steps to address review findings
The SBO team plans to address all the highlighted issues and to attain OBO Foundry status. While a number of the issues detailed above are relatively simple to rectify, some require further discussion within the SBO team, with other OBO ontology developers, or with OBO co-ordinators (reviewers):
- We are currently looking at MIREOT as a means to import terms from other ontologies.
- We are currently exploring potentially better ways in which certain is_a relationships could be expressed in SBO.
- We have identified some inconsistencies with obsolete term placement within SBO. These issues are currently being addressed, and require some retreival of information from archives.
- We see some potential means through which we defer to knowledge from domains covered by, for example, SO and GO. We will be contacting the authors of those ontologies soon.
- We are investigating the issues highlighted from the review. We envisage obsoletion of some terms, and probably a better placement/definition of others.
- We are looking into handling some of our asserted multiple inheritance problems using 'properties'.
- We agree that it would be beneficial to make more transparent the relationships that are currently hidden within the MathML of mathematical terms. Initially, we will be discussing potential solutions to this problem internally.