Matching arthropod anatomy ontologies to the Hymenoptera Anatomy Ontology: results from a manual alignment

Abstract

Matching is an important step for increasing interoperability between heterogeneous ontologies. Here, we present alignments we produced as domain experts, using a manual mapping process, between the Hymenoptera Anatomy Ontology and other existing arthropod anatomy ontologies (representing spiders, ticks, mosquitoes and Drosophila melanogaster). The resulting alignments contain from 43 to 368 mappings (correspondences), all derived from domain-expert input. Despite the many pairwise correspondences, only 11 correspondences were found in common between all ontologies, suggesting either major intrinsic differences between each ontology or gaps in representing each group's anatomy. Furthermore, we compare our findings with putative correspondences from Bioportal (derived from LOOM software) and summarize the results in a total evidence alignment. We briefly discuss characteristics of the ontologies and issues with the matching process.

Figure 1

Cytoscape visualization showing the HAO full ontology network (black circle nodes and gray lines) with correspondences mapped from the other arthropod anatomy ontologies: SPD (purple diamonds); TADS (green squares); TGMA (red octagons); FBbt (blue triangles). Box A represents an area of general agreement between the ontologies, showing multiple correspondences from each ontology (largely consisting of CARO and many general body classes); further magnification (A2, represented by dashed box in A) reveals nodes with many correspondences from different ontologies (arrow). Box B represents an area with fewer correspondences, mainly from FBbt with some TADS (largely consisting of specific muscle classes not present in TGMA and SPD); further magnification (B2, represented by dashed box in B) reveals instances where one class from FBbt is aligned to multiple HAO classes (arrow; many to one relationship).

Figure 2

Comparison of the number of correspondences found through manual alignment alone, LOOM-based algorithm alignment alone (available from Bioportal) and using both methods. Only source-target alignments with results from both methods (SPD-HAO, TGMA-HAO and FBbt-HAO) are shown. Correspondences found by LOOM alone are further characterized as valid (overlooked during the manual alignment), mismatched (invalid correspondences) or other (errors; see text).