Culture Biosciences, a developer of automated bioprocess development tools, has announced a strategic collaboration with Google Cloud that will integrate Culture’s real-time data monitoring and cloud-connected bioreactors with Gemini AI and other advanced data tools developed by Google.
The collaboration is part of a three-phase plan that involves investing in both cloud computing infrastructure and artificial intelligence tools that will help “democratize early upstream process development,” Darcy Birse, PhD, Culture’s CEO, told GEN in an interview. He added that the company is developing tools that are capable of not only capturing data from bioprocessing experiments in real-time but also extracting insights from the data that can be fed back into the system to optimize the manufacturing process. Culture is also working on tools for building and running in silico models and simulations of novel experimental designs, as part of the partnership.
“Our relationship with Google will transform life science research and development by combining real upstream cloud-connected bioreactor experiments with data science to propel the future of the industry,” Birse said. “Together, we are empowering scientists to achieve faster, more accurate results.”
Applying AI to bioprocess development
Culture opened its doors about eight years ago offering contract development and manufacturing organization (CDMO) services including clonal screening, microbial and mammalian process development and optimization, and some process characterization. It also works on viral vectors for use in AAVs.
The company, which has its offices in the Google Verily campus in San Francisco, initially started in the synthetic biology space but then transitioned into the biologics space about two years ago. It now works with a range of customers including pharmaceutical and biotechnology companies working on cell and gene therapies and other therapy types.
Birse, who joined Culture about a year ago, told GEN that the company has about 300 250 mL and 5 L bioreactors, each with single-use vessels for mammalian cell lines or microbial fermentation. In line with its vision to bridge real-world experimentation with data science, Culture’s bioreactors are connected to cloud infrastructure where data from runs is captured and stored. The company also provides Console, a proprietary cloud-based software for designing, monitoring, and analyzing experiments run using its bioreactors.
In the increasingly competitive therapeutic market, tools that provide some sort of edge are a boon for drug developers. Capturing data in real-time can provide significant value to customers if companies can find a way to extract and use its insights. Prior to partnering with Google, Culture had done a lot of in silico simulations and modeling of different processes and gathered data from about 25,000 bioreactor runs. All of “that data tells us something [and] we’ve got to find smarter ways to use the data,” Birse said
The first step was integrating and cleaning the data as well as storing it in a centralized location. This was part of the first phase of Culture’s investment in computational infrastructure. To accomplish those goals, the company worked with an external partner to ensure the fidelity and consistency of its internal data as well as any data from customer projects and other sources.
Going through the process of cleaning and prepping the data set up the second and third phases of Culture’s computational plans, which will be undertaken in the Google Cloud team. Under the terms of the agreement, the company will be able to integrate Google Cloud tools like Gemini AI, BigQuery, and Looker into Console and provide advanced data management capabilities and automated process control.
Bringing these capabilities to bioprocessing is a game changer for the field, in Birse’s view. With real-time monitoring during and after experiments, scientists can tweak and optimize their processes while they are happening as well as identify failure points. They can also be more thoughtful in terms of experimental design and create smarter, faster experiments. Culture is also able to incorporate insights from customers’ data alongside their own and use the combined information to further optimize experiments and processes.
Being able to speed up experimental design and planning provides real value for customers, Birse told GEN. Scientists will be able to run the fewest number of experiments and get the best outcomes in less time. “Anytime you can drive that [number] down but get equal or better value from the experiments, that’s a big win.”
Furthermore, bringing in machine learning and artificial intelligence capabilities will help the company do significantly more with the data and metadata captured from their processes. One of the ways these tools can add value is in the development of digital twins that can be used to, among other things, model fluid dynamics in bioreactors as well as specific manufacturing processes.
This can be especially beneficial for helping customers successfully scale-up their projects in an efficient way. “You want the highest confidence and probability that you’re gonna have success out of the gate. We’ve got some data that actually allows us to go straight from our 250 mL right up to a 2000 L without having to do any interim processing steps,” Birse said. That capability is available in the Console software currently. Access to Google’s tools “will allow us to do significantly more.”
For the third phase of its plan, Culture will work with the Google Cloud team on a data manipulation and visualization application that will leverage Looker and Gemini’s natural language query capabilities. They plan to make it available under a software-as-a-service model in the Google Cloud marketplace. Separately, the company is also developing a proprietary bioreactor platform with a single-use vessel that will also include its Console software slated for release in the first quarter of 2025. Culture is also in the process of preparing for a Series C funding round.
