MAT-Cell: A Multi-Agent Tree-Structured Reasoning Framework for Batch-Level Single-Cell Annotation

2026-04-07

AI Summary

This work addresses the limited generalizability of supervised methods and the susceptibility of large language models to noise due to insufficient biological priors in single-cell annotation. To overcome these challenges, the authors propose a neuro-symbolic reasoning framework that reformulates cell-type annotation as a verifiable logical proof generation process. By integrating biological prior knowledge into a symbolic constraint mechanism, the approach leverages multi-agent collaboration, dialectical validation via homogeneous refutation agents, tree-structured syllogistic reasoning, and adaptive retrieval-augmented generation (RAG) to produce logically consistent and robust annotations. Evaluated on large-scale cross-species benchmarks, the method significantly outperforms state-of-the-art models and demonstrates stable high performance under out-of-distribution conditions.

Abstract

Automated cellular reasoning faces a core dichotomy: supervised methods fall into the Reference Trap and fail to generalize to out-of-distribution cell states, while large language models (LLMs), without grounded biological priors, suffer from a Signal-to-Noise Paradox that produces spurious associations. We propose MAT-Cell, a neuro-symbolic reasoning framework that reframes single-cell analysis from black-box classification into constructive, verifiable proof generation. MAT-Cell injects symbolic constraints through adaptive Retrieval-Augmented Generation (RAG) to ground neural reasoning in biological axioms and reduce transcriptomic noise. It further employs a dialectic verification process with homogeneous rebuttal agents to audit and prune reasoning paths, forming syllogistic derivation trees that enforce logical consistency.Across large-scale and cross-species benchmarks, MAT-Cell significantly outperforms state-of-the-art (SOTA) models and maintains robust per-formance in challenging scenarios where baselinemethods severely degrade. Code is available at https://gith ub.com/jiangliu91/MAT-Cell-A-Mul ti-Agent-Tree-Structured-Reasoni ng-Framework-for-Batch-Level-Sin gle-Cell-Annotation.

Problem

• single-cell annotation
• Reference Trap
• Signal-to-Noise Paradox
• out-of-distribution generalization
• biological priors

Innovation

• neuro-symbolic reasoning
• Retrieval-Augmented Generation
• multi-agent verification
• single-cell annotation
• syllogistic derivation tree