Biomni, released on July 9, 2026, by the Stanford University SNAP Lab, is a bio-specialized general-purpose AI agent system that autonomously plans and executes complex and extensive life science research workflows. This tool operates in a manner similar to an experienced bioinformatician skillfully navigating various biological databases and local analysis tools to design optimal pipelines. Biomni integrates the inference capabilities of large language models (LLMs), retrieval-augmented planning, and code execution in a real-world sandbox environment to formulate hypotheses based on researcher queries and automatically generate and validate the necessary code. Traditional bioinformatic analysis methods have been cumbersome, requiring researchers to manually search dozens of different web databases, such as NCBI and ChEMBL, or to intricately combine various specialized analysis software packages based on R and Python. To overcome the limitations of this fragmented research environment, Biomni seamlessly integrates over 150 bio-specialized analysis tools, 59 curated databases, and 105+ software packages into a single intelligent agent platform. This is akin to a conductor organically connecting numerous individual research tools, going beyond simply listing API calls to understand complex data flows, autonomously correct errors through multi-step reasoning, and derive optimal analysis results. Life science researchers can use Biomni to drastically shorten the entire process, from gene editing design to the discovery of new drug candidates. For example, a researcher performing a CRISPR knock-out screen analysis to identify immune checkpoint modulators can set MAGeCK parameters (alphacutoff=0.05) with natural language commands and automatically extract 32 significant regulatory genes from the data lake. Furthermore, by inputting single-cell RNA sequencing (scRNA-seq) expression matrix data, the system can complete everything from preprocessing with cell filtering parameters (mingenes=200) to automated cell annotation using the CellMarker database within 1-2 seconds, providing hypothesis validation and visualization reports simultaneously.

Biomni, released on July 9, 2026, by the Stanford University SNAP Lab, is a bio-specialized general-purpose AI agent system that autonomously plans and executes complex and extensive life science research workflows. This tool operates in a manner similar to an experienced bioinformatician skillfully navigating various biological databases and local analysis tools to design optimal pipelines. Biomni leverages the inference capabilities of large language models (LLMs) and retrieval-augmented planning.

🔬 CRISPR Screen Design and Target Gene Selection: The Biomni agent analyzes immune checkpoint genes from 59 biological databases to establish a CRISPR knock-out screen protocol, sets MAGeCK parameters (alpha_cutoff=0.05), and ultimately selects and recommends 32 optimal guide RNA (gRNA) candidates that can inhibit T-cell depletion.