AI-powered TargetMap: Enabling system-level target discovery through full-path reasoning on a unified knowledge graph
Xizhi Jin, BO Yang, Hongxia Xu, Ji Cao, Qiaojun He, Ruijia Wu, Sijie Wang, Jiahe Chen, Shuhao Shen, Fangjie Yan, Jian Wu
iScience
Problems Identified (4)
Incomplete target-network understanding: Modern drug discovery suffers from high attrition because individual targets are not fully understood within complex system-wide biological networks.
Limited global pathway reasoning: Existing computational tools such as GNNs are limited in capturing global semantic context and long-range dependencies in mechanistic pathways.
Incomplete target-network understanding: Modern drug discovery suffers from high attrition because individual targets are not fully understood within complex system-wide biological networks.
Limited global pathway reasoning: Existing computational tools such as GNNs are limited in capturing global semantic context and long-range dependencies in mechanistic pathways.
Proposed Solutions (5)
TargetMap knowledge graph platform: The paper introduces TargetMap, an AI-driven knowledge graph platform for system-level therapeutic target discovery.
LLM full-path graph reasoning: TargetMap uses an LLM-based full-path graph reasoning algorithm to reason over entire biological pathways.
Pathway narrative representation: The approach represents entire biological pathways from structured knowledge graphs as coherent narratives for holistic analysis.
Unified knowledge base with Graph RAG: TargetMap is supported by a unified knowledge base and Graph RAG for mechanistic reasoning over knowledge graphs.
Interactive signaling-network visualization: TargetMap includes interactive visualization for exploring signaling networks.
Results (3)
System-level target discovery:
Global mechanistic analysis:
Immersive signaling-network exploration:
Research Domain
AI-assisted drug discovery and biomedical knowledge graphs