Shanghai jiao tong university develops ai-native lab to automate molecular biology experiments, published in nature

Researchers from the IPADS Lab at Shanghai Jiao Tong University, led by Chen Haibo and Wang Zhaoguo, collaborated with teams from the School of Chemistry and Chemical Engineering and the Transformative Molecular Frontier Science Center, including Fan Chunhai and Wang Fei. This collaboration, utilizing the DNA Storage Research Center at Zhangjiang Higher Institute, resulted in a publication in Nature, which also featured a homepage highlight. The team engineered the AutoDNA apparatus, an AI-native scientific laboratory, enabling autonomous molecular biology experimentation. This system overcomes challenges in complex process organization, multi-instrument coordination, and precision required for micro-scale liquid handling, which plague existing AI labs that often rely on 'bolt-on' architectures. AutoDNA employs a model-experiment-instrument co-design approach. It uses a multi-agent architecture to break down complex tasks—such as literature review, reagent selection, and code generation—into coordinated steps. Furthermore, it abstracts diverse instruments into 'atomic services' understandable by the model, allowing the AI to generate and correct cross-device control code. The platform integrates high-precision hardware with a scheduling system, and has been validated in applications like nucleic acid testing and cancer detection, running on an OpenHarmony operating system.