This project provides a framework for AI-human collaboration in scientific research, enabling researchers to work with a team of LLM agents. It facilitates scientific discovery through structured team meetings and individual agent interactions, as demonstrated by its application in designing SARS-CoV-2 nanobodies.

How It Works

The Virtual Lab orchestrates a team of LLM agents to tackle scientific research tasks. It supports both group discussions among agents and focused one-on-one sessions with individual agents. This approach allows for complex problem-solving by leveraging the collective intelligence of multiple AI agents, guided by human researchers.

Quick Start & Requirements

• Install via pip: pip install virtual-lab
• Alternatively, clone the repo and install locally: git clone https://github.com/zou-group/virtual_lab.git && cd virtual_lab && pip install -e .
• Requires Python 3.12.
• Requires an OpenAI API key set as the OPENAI_API_KEY environment variable, currently utilizing GPT-4o.
• Setup is estimated to take a couple of minutes.

Highlighted Details

• Successfully designed 92 experimentally validated nanobodies for SARS-CoV-2.
• Integrates computational tools like ESM, AlphaFold-Multimer, and Rosetta.
• Detailed example notebook (nanobody_design/run_nanobody_design.ipynb) available for nanobody design workflow.
• Published in Nature (DOI: https://doi.org/10.1038/s41586-025-09442-9).

Maintenance & Community

No specific community channels or roadmap details are provided in the README.

Licensing & Compatibility

The README does not specify a license. Compatibility for commercial or closed-source use is not detailed.

Limitations & Caveats

The project currently relies exclusively on OpenAI's GPT-4o model and requires an OpenAI API key, limiting its use to those with access and budget for this service. No information is provided regarding support for other LLM providers or open-source models.