Oct 24, 2020 - Technology

Using AI to unlock the genetic secrets of food

Illustration: Sarah Grillo/Axios

Illustration: Sarah Grillo/Axios

A startup is employing machine learning to identify what it calls the "dark matter of nutrition."

Why it matters: More than 99% of phytonutrients — the natural chemicals produced by plants — are unknown to science. If we can illuminate that dark matter, we can identify and cultivate compounds in foods for specific health value.

How it works: The startup Brightseed uses a proprietary AI platform called Forager to predict the likelihood that plants will have useful natural compounds and the likelihood that those phytonutrients will have specific health benefits.

  • The platform is trained on a vast library of biomedical and plant research. That allows the AI to make connections between plant ingredients and health effects far faster than any human scientist could alone.
  • "It effectively works as a Google search engine for these compounds and what they can do," says Jim Flatt, Brightseed's CEO. "Once we've found those compounds, we can develop products and services around them."

Details: The Forager system can screen by specific chemical compound, or by health benefit, searching for ingredients that might affect cholesterol or cognitive function.

  • Earlier this year Brightseed announced a partnership with Danone North America to use the Forager system on soy.
  • It's a sign that even one of the most heavily used plants on Earth may have additional nutritional secrets that AI can help tease out.

What they're saying: "AI allows us to tackle things that would have taken far too long in the past computationally," says Flatt.

The bottom line: As exciting as the possibility of using biotechnology to synthesize entirely new compounds is, we've only begun to understand what already exists in the world — and AI can accelerate those discoveries.

Go deeper