Ontobull and BFOConvert: Web-based programs to support automatic ontology conversion Edison Ong1, Zuoshuang Xiang1, Jie Zheng2, Barry Smith3, Yongqun He1* 1 University of Michigan, Ann Arbor, Michigan, USA; 2 University of Pennsylvania, Philadelphia, PA, USA; 3 University at Buffalo, Buffalo, NY, USA. ABSTRACT previously developed the BFOConvert software. In light of When a widely reused ontology appears in a new version which is not the fact that the IRI conversion issue exists also in other compatible with older versions, the ontologies reusing it need to be updated accordingly. Ontobull (http://ontobull.hegroup.org) has been developed to cases, we have now expanded the BFOConvert program to automatically update ontologies with new term IRI(s) and associated create Ontobull, a program with greater flexibility and a metadata to take account of such version changes. To use the Ontobull web more user-friendly web interface. interface a user is required to (i) upload one or more ontology OWL source files; (ii) input an ontology term IRI mapping; and (where needed) (iii) In addition to ontology version updating, Ontobull can provide update settings for ontology headers and XML namespace IDs. be used for other purposes, including term IRI updating and Using this information, the backend Ontobull Java program automatically updating of ontology metadata such as ontology header updates the OWL ontology files with desired term IRIs and ontology (e.g., owl:imports) and XML namespaces (xmlns) informa- metadata. The Ontobull subprogram BFOConvert supports the conversion of an ontology that imports a previous version of BFO. A use case is pro- tion (https://www.w3.org/TR/owl-guide/). The owl:imports vided to demonstrate the features of Ontobull and BFOConvert. attribute in ontology header specifies an OWL file to import into current ontology. The xmlns attribute qualifies element 1 INTRODUCTION and attribute names of the ontology. Biological ontologies are sets of computer- and human- In this software demonstration, we will introduce the interpretable terms and relations that represent entities in the tool design and use cases of Ontobull and BFOConvert. biological world and how they relate to each other. Hundreds of ontologies have been developed. Over 150 2 ONTOBULL SOFTWARE DESIGN biomedical ontologies have been developed following the The Ontobull web program provides a user-friendly Open Biological and Biomedical Ontologies (OBO) interface for input of an ontology file and conversion Foundry principles [1]. NCBO BioPortal [2] currently settings. The Ontobull Java program, developed using the comprises more than 400 ontologies including both OBO Spring Model-view-controller (MVC) framework and and non-OBO ontologies. Thymeleaf template engine, processes the user’s requests The Basic Formal Ontology (BFO) [1, 3] has been used and generates an Ontobull OWL output file. The as the top ontology by some 175 other ontologies BFOConvert program is a subprogram of Ontobull (http://ifomis.uni-saarland.de/bfo/users), with the goal of specifically targeting conversions involving a move from facilitating interoperability among these ontologies. In addi- use of BFO 1.1 to BFO 2.0. tion to BFO, other OBO Foundry ontologies use relations \ defined in the OBO Relation Ontology (RO) (https://github.com/oborel/obo-relations). For example, the information Artifact Ontology (IAO) and the Ontology for Biomedical Investigations (OBI) [4] both import BFO and RO. One challenge in reusing external ontologies is that once the reused source ontology is updated with modified IRIs, it is difficult to update the target ontologies. Although these reference ontologies are designed to be maximally stable, they do sometimes still undergo major changes. For example, BFO 1.0 and 1.1 do not follow OBO Foundry ID policy (http://obofoundry.org/id-policy.html). As a result, BFO has been updated from BFO 1.0/1.1 to 2.0. To support automatic conversion of ontologies aligned with BFO, we Fig. 1. Ontobull web interface. The circled settings are used in the use case demonstration described next. * To whom correspondence should be addressed: yongqunh@umich.edu 1 Ong et al. 3 ONTOBULL WEB INTERFACE Ontobull can also be used to integrate multiple OWL The use of the Ontobull web interface (Fig. 1) involves three files. The OWL ontology files may have different IRIs re- steps: ferring to the same entities. For integration, we need to re- Upload an ontology source file: An OWL file or a zip place these different IRIs with common ones across all sali- file including an OWL file(s) can be selected and uploaded. ent files. For example, the OWL representation of Beta cell The zip file format is typically used when an ontology in- genomics data (http://www.betacell.org/gbco/) integrates cludes many unmerged OWL files. In this case, Ontobull information from two resources that include terms assigned can automatically process all the OWL files in succession. with different IRIs. Ontobull was used to update IRIs to Define ontology term IRI mapping: The user is required enable this integration. to provide at least one pair consisting of an original ontolo- gy term IRI and a new IRI that will be used to replace it. 5 BFOCONVERT SUBPROGRAM Multiple mapping pairs can be specified using the dynamic The BFOConvert subprogram is now available at: web user-interface. Alternatively, a user can choose a file http://ontobull.hegroup.org/bfoconvert. It provides a pre- that includes all the IRI mapping pairs organized in a simple designed mapping file according to the current stage of tab-delimited format. BFO/RO development. Provide an ontology header and XML namespace updating setting (optional): This setting is used to remove, 6 SOURCE CODE AND LICENSE add, or replace xmlns or owl:imports contents in the OWL The Ontobull source code is openly available at: output file. https://github.com/OntoZoo/Ontobull. The Ontobull source Use cases are shown in what follows. code license is Apache License 2.0. 4 ONTOBULL USE CASE DEMONSTRATION 7 SUMMARY Fig. 2 illustrates how Ontobull is used to convert the With ever increasing needs for ontology updating, the web- Parasite Life Cycle ontology (OPL) which uses terms from based Ontobull and BFOConvert programs provide a timely the Relation Ontology from an old version to a new version. platform for automatic ontology conversion. OPL is an ontology designed to provide a consistent representation of lifecycle stages in parasitic organisms [5]. ACKNOWLEDGEMENTS The “OPL.owl” highlighted in the first circle in Fig. 1, This research is supported by an NIH R01 grant above, represents the OWL file of the original version of (1R01AI081062). We acknowledge the assistance of James this ontology. The second highlighted circle in Fig. 1 A. Overton and Alan Ruttenberg in the development of represents the setting for an IRI mapping. Fig. 2A represents BFOConvert. the results of the conversion of the IRI for the RO expression ‘part of ’. The third highlighted circle in Fig. 1 represents one header cleanup option, illustrated in Fig. 2B. REFERENCES [1] B. Smith, M. Ashburner, C. Rosse, J. Bard, W. Bug, et al., A more detailed tutorial and more use case demonstra- “The OBO Foundry: coordinated evolution of ontologies to tions are available on the Ontobull tutorial website at: support biomedical data integration,” Nat Biotechnol, vol. 25, http://ontobull.hegroup.org/tutorial. BFOConvert has been no. 11, pp. 1251-5, Nov, 2007. used to support the conversion of many ontologies that used [2] P. L. Whetzel, N. F. Noy, N. H. Shah, P. R. Alexander, C. old versions of BFO and RO. Nyulas, T. Tudorache, and M. A. Musen, “BioPortal: enhanced functionality via new Web services from the old new National Center for Biomedical Ontology to access and use ontologies in software applications,” Nucleic Acids Res, vol. 39, no. Web Server issue, pp. W541-5, Jul, 2011. [3] P. Grenon, and B. Smith, “SNAP and SPAN: Towards (A) IRI update (shown in Protégé) Dynamic Spatial Ontology,” Spatial Cognition and old Computation, vol. 4, no. 1, pp. 69-103, 2004. [4] A. Bandrowski, R. Brinkman, M. Brochhausen, M. H. Brush, new B. Bug, M, C, Chibucos, et al, “The Ontology for Biomedical Investigations,” PLoS One, vol. 11, no. 4, pp. e0154556, 2016. (B) Header update (shown in a text editor) [5] P. P. Parikh, J. Zheng, F. Logan-Klumpler, C. J. Stoeckert, Jr., Fig. 2. Ontobull usage demo. The input (OPL.owl) and settings C. Louis, P. Topalis, et al, “The Ontology for Parasite are shown in Fig. 1. (A) The RO object property ‘part of’ IRI is Lifecycle (OPL): towards a consistent vocabulary of lifecycle updated. The results are displayed using the Protégé-OWL editor. stages in parasitic organisms,” J Biomed Semantics, vol. 3, no. (B) Header updated. The source file updates are shown. 1, pp. 5, 2012. 2