SpaCellAgent: A Self-Evolving LLM-Based Multi-Agent Framework for Trajectory Analysis

2026-07-08

AI Summary

Current cell trajectory inference methods rely heavily on manual intervention and heterogeneous tools, resulting in high barriers to entry and low efficiency. This work proposes a large language model–based multi-agent framework that enables end-to-end automation—from natural language instructions to spatiotemporal trajectory analysis and narrative generation—through workflow planning, dynamic tool orchestration, and a feedback-driven self-evolution mechanism. Evaluated on six diverse and complex datasets, the proposed approach achieves over 40% improvement in analytical efficiency while maintaining accuracy comparable to expert-level analysis. To the best of our knowledge, this is the first method to fully automate the entire trajectory inference pipeline without human oversight.

Abstract

Spatial and Single-cell transcriptomics are transformative in deciphering cellular dynamics. As the fundamental paradigm for reconstructing cell developmental paths, trajectory inference (TI) is critical. However, existing methods require extensive manual intervention and proficiency in heterogeneous tools, posing a significant barrier to efficient TI analysis. To bridge this gap, we propose SpaCellAgent, an autonomous large language model (LLM) multi-agent framework that automates end-to-end spatiotemporal analysis and narrative generation. SpaCellAgent utilizes a multi-agent architecture for strategic workflow planning, a dynamic tool-orchestration engine for adaptive algorithm selection, and a self-evolution module that iteratively refines performance through feedback. We evaluate SpaCellAgent on six heterogeneous datasets encompassing complex temporal developmental trajectories, diverse sequencing platforms, and spatially-resolved tissue architectures. SpaCellAgent consistently demonstrates over 40\% improvement in analytical efficiency while maintaining expert-aligned performance. By converting natural language specifications into optimized analytical workflows and fully automating the pipeline, SpaCellAgent democratizes advanced spatiotemporal modeling and establishes a scalable, agent-driven paradigm for computational biology. The code and materials are available at https://github.com/LittleXH-shw/SpaCellAgent.

Problem

• trajectory inference
• single-cell transcriptomics
• spatial transcriptomics
• manual intervention
• heterogeneous tools

Innovation

• multi-agent framework
• self-evolving LLM
• trajectory inference
• tool orchestration
• spatiotemporal analysis