Golbreich Christine, Zhang Songmao, Bodenreider Olivier
LIM, University Rennes 1, 35043 Rennes, France.
Web Semant. 2006;4(3):181-195. doi: 10.1016/j.websem.2006.05.007.
We present the method developed for migrating the Foundational Model of Anatomy (FMA) from its representation with frames in Protégé to its logical representation in OWL and our experience in reasoning with it. Despite the extensive use of metaclasses in Protégé, it proved possible to convert the FMA from Protégé into OWL DL, while capturing most of its original features. The conversion relies on a set of translation and enrichment rules implemented with flexible options. Unsurprisingly, reasoning with the FMA in OWL proved to be a real challenge, due to its sheer size and complexity, and raised significant inference problems in terms of time and memory requirements. However, various smaller versions have been successfully handled by Racer. Some inconsistencies were identified and several classes reclassified. The results obtained so far show the advantage of OWL DL over frames and, more generally, the usefulness of DLs reasoners for building and maintaining the large-scale biomedical ontologies of the future Semantic Web.
我们展示了为将解剖学基础模型（FMA）从其在Protégé中使用框架的表示形式迁移到其在OWL中的逻辑表示形式而开发的方法，以及我们使用它进行推理的经验。尽管Protégé中广泛使用了元类，但事实证明，在保留FMA大部分原始特征的同时，将其从Protégé转换为OWL DL是可行的。这种转换依赖于一组通过灵活选项实现的翻译和丰富规则。不出所料，由于FMA规模庞大且复杂，在OWL中对其进行推理被证明是一项真正的挑战，并且在时间和内存需求方面引发了重大的推理问题。然而，Racer已经成功处理了各种较小版本。识别出了一些不一致之处，并对几个类进行了重新分类。迄今为止获得的结果显示了OWL DL相对于框架的优势，更普遍地说，显示了DL推理器对于构建和维护未来语义网的大规模生物医学本体的有用性。
