1. Biological knowledge graphs can be modelled using knowledge graph embedding (KGE) models, which learn low-rank vector representations of graph nodes and edges to preserve the graph's inherent structure.
2. KGE models have superior scalability and accuracy compared to traditional graph exploratory approaches, making them suitable for various biology applications such as drug-target interaction prediction and polypharmacy side effect analysis.
3. Practical considerations for adopting KGE models in biological systems should be taken into account, and there are potential opportunities and challenges in this area.