Among the many challenges faced in Bioprocessing 4.0, two of them are complexity and data integration. In Cambridge, MA, Ganymede founder and CEO Nathan Clark is helping bioprocessors overcome these issues. He describes Ganymede’s cloud infrastructure as “a complete software and data backbone in a box for an entire biotech company, because we are not just storing the data but we are also building real-time instrument integrations and analysis automations for apps like LIMS and ELNs.”
Data integration has plagued bioprocessing for years. “It’s easy enough to go grab the data, but then what do you do with that file once you have it?” Clark asks. “How do you translate that file into a standard harmonized format and make it FAIR?”
By FAIR, Clark means meeting the FAIR principles for handling data, which were established in 2016. So, Ganymede has developed a platform that not just integrates the data, but harmonizes it and automates downstream scientific analysis along the way.
In working with a synthetic organ development group, Clark’s team faced a series of complex operations: growing cells, validating them, developing organs on scaffolds, and so on. “You had a slew of integrations from common flow cytometry to completely custom devices,” he says. “We started by building a bunch of point-to-point integrations.”
In that project and others, Ganymede takes what Clark calls a three-pillar approach: get the data off the instruments and parse it into Ganymede’s open tabular format; automate the scientific analysis; and then automate data entry into the result’s destination, like an ELN.
With the data all in one place, Clark explains that a company could use it in various ways and go beyond point-to-point integrations, including the development of a digital twin. “You actually have the real-time data inside the integration at your fingertips as a dataframe, and then you can build Python models—everything from simple dashboards up to machine-learning models,” he says.
Instead of just making a digital twin of an instrument or bioprocess, Clark envisions bigger applications. “An interesting case would be digital twins of the entire internals of a CRO or CDMO,” he notes. “There’s no magic wand, no silver bullet for something like that, but we’re making inroads on this with our current clients.”