RegNetAgents: A Multi-Agent System for Gene Regulatory Network Interpretation

RegNetAgents is a Python package for automated downstream analysis of pre-computed gene regulatory networks (GRNs) using a multi-agent architecture built on LangGraph. The software coordinates four specialized domain agents (cancer biology, drug discovery, clinical relevance, systems biology) to analyze regulatory relationships from ARACNe-inferred networks. It can be used programmatically via Python or through MCP-compatible clients (Claude Desktop, Cursor, Zed) via natural language. A fixed directed acyclic graph (DAG) controls execution order and data flow, while an optional LLM layer adds natural-language narrative rationale to deterministic rule-based domain assessments, ensuring identical inputs produce identical computational outputs. RegNetAgents is intended for computational biologists, bioinformaticians, and researchers who need automated GRN analysis.