Towards Autonomous Mechanistic Reasoning in Virtual Cells

2026-04-13

AI Summary

Large language models struggle to generate factually reliable and actionable mechanistic explanations in open scientific domains such as biology, limiting their utility in autonomous scientific discovery. This work proposes VCR-Agent, a multi-agent framework that formalizes mechanistic reasoning as verifiable “mechanism action graphs.” By integrating biological knowledge retrieval, validator-based filtering, and supervised fine-tuning, the framework enables the automated generation and verification of biological explanations within a virtual cell environment. The authors introduce the VC-TRACES dataset and demonstrate through experiments that their approach significantly improves factual accuracy in gene expression prediction tasks and enhances the effectiveness of supervisory signals for model training.

Technology Category

• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Multiagent Systems: Mechanism Design
• Computer Vision: Visual Reasoning & Symbolic Representations

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large language models (LLMs) have recently gained significant attention as a promising approach to accelerate scientific discovery. However, their application in open-ended scientific domains such as biology remains limited, primarily due to the lack of factually grounded and actionable explanations. To address this, we introduce a structured explanation formalism for virtual cells that represents biological reasoning as mechanistic action graphs, enabling systematic verification and falsification. Building upon this, we propose VCR-Agent, a multi-agent framework that integrates biologically grounded knowledge retrieval with a verifier-based filtering approach to generate and validate mechanistic reasoning autonomously. Using this framework, we release VC-TRACES dataset, which consists of verified mechanistic explanations derived from the Tahoe-100M atlas. Empirically, we demonstrate that training with these explanations improves factual precision and provides a more effective supervision signal for downstream gene expression prediction. These results underscore the importance of reliable mechanistic reasoning for virtual cells, achieved through the synergy of multi-agent and rigorous verification.

Problem

• mechanistic reasoning
• virtual cells
• scientific discovery
• factual grounding
• biological explanation

Innovation

• mechanistic reasoning
• virtual cells
• multi-agent framework
• action graphs
• knowledge verification