Traditional information retrieval systems, primarily based on keyword searches, often fail to capture nuanced domain-specific relationships. To address such an issue, we propose using a semantic graph-based approach, which enables enhanced semantic querying capabilities and contextual data retrieval. By mapping entities and their complex interdependencies within a graph, our system allows for sophisticated querying beyond simple keyword matches. It effectively leverages interconnected data to provide contextually relevant responses to complex queries, thereby improving the accuracy and depth of information retrieval. We have effectively demonstrated our solution in the field of nutrigenomics, highlighting how semantic graphs can significantly improve data interpretation and facilitate the creation of highly personalized nutrition plans and therapeutic interventions based on comprehensive and detailed genetic insights. Our code is publicly available on GitHub at https://github.com/CosmoIknosLab/bertopic_graph .