Briefly Bio, a London-based startup, has secured a significant $1.2 million funding round from Compound VC and others, aiming to enhance the reproducibility of scientific experiments and data.
This initiative addresses the prevalent reproducibility crisis in research, where countless scientific results are difficult or impossible to replicate.
In response, Briefly Bio leverages large language models (LLMs), the technology behind AI applications like ChatGPT, to transform complex lab documentation into a clear, structured format. This platform enables scientists to easily utilize and build upon original documents within their labs.
The company has already attracted initial customers for its paid offering and plans to develop a public version of the platform for open scientific data sharing, akin to GitHub’s code repository for open-source software.
“As GitHub facilitated collaboration among software engineers, we believe Briefly can empower scientists to enhance their experimental work,” said Harry Rickerby, CEO and co-founder of Briefly Bio.
Addressing the Reproducibility Issue
In scientific research, various approaches are employed to tackle complex biological challenges. While the protocols outlining objectives, methodologies, and statistics are meticulously documented, inconsistencies often arise in how different scientists present their work.
This lack of standardization can lead to ambiguity, making collaboration difficult and contributing to an annual cost exceeding $50 billion due to unsuccessful reproducibility efforts.
Rickerby, along with LabGenius colleagues Katya Putintseva and Staffan Piledahl, experienced these challenges firsthand. Rickerby explained, “Katya faced difficulties in reusing published work, while Staffan required precise lab workflow definitions for automation. We discovered that the core issue was the inconsistent documentation of lab practices.”
To combat this, they launched Briefly Bio, providing a platform that converts any scientifically documented protocol into a standardized format rich with step-by-step information. Users can input raw text from original documents, and the tool generates a structured output detailing how to reproduce or build upon the experiment.
“Powered by generative AI, Briefly transforms plain text into a hierarchical representation, automatically extracting and categorizing essential information into processes, actions, and parameters. This creates a visual representation that is far more accessible than traditional text,” Rickerby explained.
Enhancing Scientific Collaboration
Briefly Bio not only organizes existing scientific descriptions but also features an AI copilot capable of recognizing errors and generating relevant parameters quickly. This enhancement enriches the hierarchical representation, making it suitable for reuse in experiments.
While specifics about the underlying models remain undisclosed, Rickerby mentioned these models are enhanced with additional experimental context for better comprehension.
Teams can utilize Briefly Bio’s workspace to manage the structured method efficiently, allowing members to track the progress of each step with relevant calculations and layout details.
Future Aspirations: A GitHub-like Platform for Science
Though still in its early stages, Briefly Bio is experiencing initial revenue growth through a per-user SaaS model.
“Our main users include wet lab scientists engaged in early-stage research and development across academia, biotech, and pharma, who seek clearer documentation and collaboration tools,” Rickerby said. “We’ve also seen interest from laboratory automation professionals collaborating with scientists to define workflows before robot programming.”
Looking ahead, the company aims to create a public platform for sharing experiments and protocols, enabling scientists to access reproducible methodologies they can adapt in their labs—similar to how GitHub transformed open-source software development.
Ranking
