Ginkgo Bioworks wants “computers to speak DNA,” beefs up AI capabilities with Google’s help
"Ginkgo's mission has really been to make biology easier to engineer. And we've done so much in terms of the investments that we've made in our platform, in the technology, in Foundry, in Codebase, and now being able to complement that with AI is really going to help us to lean into and fulfill and make our mission a reality even quicker and be able to benefit many more people," said Mervyn de Souza, senior director at Ginkgo Bioworks.
Building an AI foundation with Google's help
With its automated bio lab called Foundry, Ginkgo Bioworks had a tremendous amount of data on protein, enzyme discovery and optimization, but it needed the cloud-based infrastructure to expand its AI capabilities, said Dmitriy Ryaboy, VP of AI enablement for Ginkgo Bioworks. With its Google partnership, Ginkgo Bioworks now has access to Google Cloud and its expertise, which will provide crucial support to build out its AI capabilities, he added.
“Google is an absolute leader in this space. A lot of the technologies that are used in particular in generative AI -- the new wave of AI -- came out of Google... And they've developed everything from the algorithms to the hardware to run this stuff efficiently.”
And on the other side of the AI foundation, Ginkgo Bioworks has “built a massive machine for generating ... new biological data,” and has more than 15 years of data, which includes information on “how different genetic modifications affect [an] end result,” Ryaboy said. Currently, the company leverages AI for “meta genomics sourcing,” where a user has found a protein of interest and wants to model different variants, he explained.
"We want to teach computers to speak DNA, so that becomes table stakes, and that requires a whole different level of investment and data generation and data processing, and it opens up all kinds of doors."
Unlocking the future of food, beverage with AI
With the combination of Google Cloud and its data, Ginkgo Bioworks will address bigger industry concerns like sustainability, while ultimately addressing the concerns and wants of the consumer, de Souza said.
“When you think about food, ... its got to be delicious, its got to be affordable, its got to be nutritious, and then, it's got to be sustainable. And I think the sustainability factors been burned into us with all of the issues with climate and making sure we have a planet for our kids and for future generations. But customers do not want to pay for sustainability.”
For instance, Ginkgo Bioworks recently partnered with Israeli food tech company Ambrosia Bio to develop an expression strain of its enzymes used to convert feedstock into the sweetener allulose, he said. While the ingredient is widely used in food and beverage formulations, allulose hasn’t been “made available widely to consumers, because it is expensive, and the production process expensive,” he explained.
“With the sheer volume and magnitude of information that we have, with our Codebase, with knowledge about enzymes, with knowledge about protein sequences -- linking that to functionality -- I think we're really well poised in this partnership to be able to seriously impact cost, bring it down, decrease the time that it takes to discover some of these novel ingredients, and get them to be used in the food industry -- so speed to market without compromising the final quality of the output.”
[Editor's Note: Interested in learning more about how manufacturers and retailers are improving their technology stack through digital transformation strategies and AI to improve operations and address historical food industry challenges? Join FoodNavigator-USA at our FREE virtual Futureproofing The Food System Summit, which will take a deep dive into food tech, food as medicine and the circular economy. Find all the details and register HERE.]
