meta-paths

meta-paths performs meta-path-based similarity search in heterogeneous knowledge graphs to quantify relationships among entities such as drugs, diseases, proteins, side effects, genetic interaction networks, and protein-protein interactions.

Key Features:

• Meta-path Implementation: Implements meta-path-based analysis in R for heterogeneous knowledge graphs.
• Built-in Similarity Metrics: Provides multiple built-in metrics for comparing pairs of nodes.
• Graph Representation Support: Leverages both edge list and adjacency list representations of knowledge graphs for metric computation.
• Auxiliary Aggregation Methods: Includes aggregation methods to measure relationships at the set level.
• Scalable Similarity Modeling: Supports scalable and flexible modeling of network similarities.

Scientific Applications:

• Biomedical association recovery: Recovers meaningful associations within open-source biomedical knowledge graphs.
• Drug–disease association identification: Identifies significant drug and disease-related connections, including associations pertinent to Alzheimer's disease.
• Network biology analysis: Facilitates analysis of genetic interaction networks and protein-protein interaction networks within heterogeneous KGs.

Methodology:

Queries the knowledge graph using specified meta-paths to extract pathways, computes similarity metrics between node pairs using edge list or adjacency list representations, and applies auxiliary aggregation methods to derive set-level relationship measures while supporting scalable modeling of network similarities.