MARBLE: Multi-Agent Reasoning for Bioinformatics Learning and Evolution

TL;DR

MARBLE is a framework that enables stable, iterative refinement of bioinformatics models through structured debate, empirical evaluation, and memory updates, improving performance while ensuring robustness.

Contribution

It introduces a novel autonomous model refinement framework combining literature-aware reference selection, debate-driven reasoning, and performance-grounded updates for bioinformatics models.

Findings

Consistently improves performance across multiple bioinformatics tasks.

Maintains high robustness and low regression during iterative refinement.

Structured debate and memory are key for stable model evolution.

Abstract

Motivation: Developing high-performing bioinformatics models typically requires repeated cycles of hypothesis formulation, architectural redesign, and empirical validation, making progress slow, labor-intensive, and difficult to reproduce. Although recent LLM-based assistants can automate isolated steps, they lack performance-grounded reasoning and stability-aware mechanisms required for reliable, iterative model improvement in bioinformatics workflows. Results: We introduce MARBLE, an execution-stable autonomous model refinement framework for bioinformatics models. MARBLE couples literature-aware reference selection with structured, debate-driven architectural reasoning among role-specialized agents, followed by autonomous execution, evaluation, and memory updates explicitly grounded in empirical performance. Across spatial transcriptomics domain segmentation, drug-target interaction prediction, and drug response prediction, MARBLE consistently achieves sustained performance improvements over strong baselines across multiple refinement cycles, while maintaining high execution robustness and low regression rates. Framework-level analyses demonstrate that structured debate, balanced evidence selection, and performance-grounded memory are critical for stable, repeatable model evolution, rather than single-run or brittle gains. Availability: Source code, data and Supplementary Information are available at https://github.com/PRISM-DGU/MARBLE.