Grain legumes are essential for providing high-quality plant-based proteins, meeting the challenges of a growing world population and promoting One Health. Although diverse, these species are closely related and share many similarities, including their global physiology and the range of limiting factors that can affect their yields. Through translational research, the increasing amount of available -omic and genetic data can be used to transfer knowledge across legume species and accelerate crop improvement. However, vocabulary inconsistencies in dataset metadata are a major barrier to the development of a heterogeneous, multi-species and efficient platform for translational research. Indeed, most research groups use specific and sometimes different terms to refer to entities and processes, making it difficult to compare datasets. Here we show how we decided to take advantage of existing Planteome ontologies and promote their use among users of our Neo4j Knowledge Graph (KG), Ortho_KB, to both standardise descriptions and use their hierarchical structure to link semantically related datasets. Using the Plant Ontology, the Plant Trait Ontology and the Plant Experimental Conditions Ontology, we are able to annotate most of the genetic and gene expression related metadata useful for querying our KG. For example, users can highlight the conservation of genetic determinants across studies and species. They can also explore and compare gene expression profiles using data from different sources.