Bio-Medical Knowledge Graphs (BioKGs) are largely used for the representation of heterogeneous inter-related bio-medical entities that can be exploited for the development of artificial intelligence in medicine. The continuous feeding of BioKGs with new results obtained by laboratory analysis is of paramount importance for the generation of massive datasets on which the AI algorithms can be properly trained and tested. In this work, we proposed a semi-automatic approach for the acquisition of tabular data, their semantic annotation according to a domain Ontology, and translation in RDF triples. After validating the generated graph, it can be included in BioKG.