A major challenge lies in discovering, emphasizing, and characterizing human gene-disease and gene-gene associations. The limitations of data on the role of human gene products in substance use disorder (SUD) makes it challenging to transition from genetic associations to actionable insights. The integration of data from multiple diverse sources, including information-dense studies in model organisms, has the potential to address this gap. We demonstrate a modified performance of the Random Walk with Restart algorithm when multi-species data is integrated in the heterogeneous network within the context of SUD. Additionally, our approach distinguishes among disparate pathways derived from the Kyoto Encyclopedia of Genes and Genomes. Thus, we conclude that direct incorporation of multi-species data to an aggregated heterogeneous knowledge graph can adjust RWR's performance and enables users to discover new gene-disease and gene-gene associations.