A Self-Evolving Agentic System for Automated Generation and Execution of Biological Protocols

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Yankai Jiang, Weiting Tang, Haoran Sun, Zhenyu Tang, Yuejie Hou, Yingnan Han, Rubo Wang, Yueyuxiao Yang, Cheng Liang, Lilong Wang, Wenjie Lou, Xiaosong Wang, Lei Bai, Meng Yang
• Venue: arXiv: cs.AI
• Published: 2026-06-30
• Collected: 2026-07-01
• URL: http://arxiv.org/abs/2606.31763v1
• PDF: https://arxiv.org/pdf/2606.31763v1
• DOI: UNKNOWN

Plain-Language Brief

The paper presents ProtoPilot, a self-evolving multi-agent system designed to automate the generation and execution of biological protocols. The system achieved high success rates in various tasks compared to OpenTrons-AI, and was validated through wet-lab tests.

What the Document Says

• The paper introduces ProtoPilot, a self-evolving multi-agent system for automating biological protocols.
• ProtoPilot includes an expert-grounded benchmark and evaluation framework for testing the conversion process as an experimental automation problem.
• The framework consists of 294 tasks derived from 98 gold-standard protocols, wet-lab expert rubrics, device-level validity gates, and real experimental tests.
• ProtoPilot has layer-wise verifiability, multi-agent orchestration, and a runtime-updated skill library for generating protocols, expanding SOPs, synthesizing SDK-compliant code, and revising workflows based on wet-lab feedback.
• The system achieved a Top@3 expert-preference rate of 90.2%, an overall protocol-to-code gate pass rate of 89.5%, and an Opentrons pass rate of 88.24%.
• These results are compared with OpenTrons-AI, which had a pass rate of 32.35%.
• Wet-lab validation was conducted, producing interpretable readouts, Sanger-confirmed products, and feedback-corrected PCA-assembled DNA targets.
• The authors claim that the evaluation framework captures execution-relevant requirements for autonomous wet-lab automation, and that ProtoPilot can meet them by converting protocol and code generation into validated execution and feedback-guided revision.

Key Concepts Mentioned

• autonomous wet-lab experimentation
• ProtoPilot
• multi-agent system
• expert-grounded benchmark
• evaluation framework
• synthetic-biology tasks
• molecular-biology tasks
• wet-lab feedback
• OpenTrons-AI

Claims Made by the Document

• ProtoPilot can meet the execution-relevant requirements for autonomous wet-lab automation.
• ProtoPilot achieved higher success rates than OpenTrons-AI in various tasks.

Evidence Used

• Top@3 expert-preference rate of 90.2%
• Overall protocol-to-code gate pass rate of 89.5%
• Opentrons pass rate of 88.24%
• Results from wet-lab validation

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Autonomous wet-lab experimentation requires more than plausible protocol text: biological intent, quantitative procedures, device constraints and experimental feedback must remain aligned from protocol and SOP design to code and physical execution. We developed ProtoPilot, a self-evolving multi-agent system, together with an expert-grounded benchmark and evaluation framework for testing this conversion as an experimental automation problem. The framework spans 294 synthetic-biology and molecular-biology tasks derived from 98 gold-standard protocols, wet-lab expert rubrics, device-level validity gates and real experimental tests. ProtoPilot incorporates layer-wise verifiability, multi-agent orchestration and a runtime-updated skill library to generate protocols, expand SOPs, synthesize SDK-compliant code and revise workflows from wet-lab feedback. It achieved a Top@3 expert-preference rate of 90.2%, an overall protocol-to-code gate pass rate of 89.5% and an Opentrons pass rate of 88.24%, compared with 32.35% for OpenTrons-AI. Wet-lab validation produced interpretable readouts, Sanger-confirmed products and feedback-corrected PCA-assembled DNA targets, establishing a verifiable route to autonomous experimentation. Together, these results show that the evaluation framework captures execution-relevant requirements for autonomous wet-lab automation, and that ProtoPilot can meet them by converting protocol and code generation into validated execution and feedback-guided revision.