This consumer microbiome startup is betting profits on an AI moonshot

Seed Health has been in the business of microbiome scientific breakthroughs since its founding in 2015, but its biggest success to-date may have been becoming profitable as a bioscience startup. The popularity of the company’s original product, DS-01, a daily probiotic and prebiotic supplement and its pediatric equivalent — sold direct to consumers in a 30-day supply of sleek green containers through a subscription model — are now allowing the company to invest in frontier science related to probiotics and both human and environmental health.

“We’re one of the few biotech companies that can say they’re profitable, and one of the awesome aspects of profitability is that you can reinvest in future innovation and frontier science,” said Ara Katz, Seed Health co-founder, and who was named to the inaugural CNBC Changemakers list earlier this year.

On Thursday, Seed Health launched CODA, a computational biology platform funded by its consumer business profitability. “I’ve always kind of thought about constructing a sustainable business model that would allow us to continue to pursue frontier science,” Katz said.

The company’s SeedLabs division also works on environmental applications for bacteria and the microbiome, such as probiotics for coral and honey bees, bacteria to decompose plastic, and the use of volcanic bacteria in carbon capture.

Offering consumer product subscriptions in one-, three-, or six-month increments has helped to put the company in a position to bet on riskier scientific discoveries. 

“Part of the reason that we have a subscription model, primarily is because that allows you so much more flexibility and understanding like revenue, and really making sure that we can make these investments for the longer term,” Katz said. “The through-line for us is always thinking about frontier science that we can accelerate the translation of into real world impact,” she added.

At its core, CODA is a computational tool which utilizes AI to process massive amounts of phenotypic and genomic data from the Human Phenotype Project, a massive collection of human data points ranging across decades.

Advised by top researchers including Eran Segal of the Weizmann Institute of Science and Eric Topol from Scripps Research Institute, the Human Phenotype Project has a global sample of over 100,000 participants cataloged annually over 25 years that includes multiple biological measures, from genome, proteome, transcriptome to the microbiome. 

Katz’s co-founder and co-CEO at Seed Health, Raja Dhir, said CODA and the accompanying data set will help to standardize microbiome science methods, which has long been an issue in the field. Previous studies have drawn conclusions based on smaller sample sizes, using sampling, analytical, and storage methods which are not standardized across the industry. 

“What is the healthy control? You can’t just take a random person and assume that they’re healthy or not healthy. But when you have 10,000 people, and you have all this data on them… let’s contrast our least healthy people to our most healthy people and develop way more tools,” said Dhir, who oversees Seed Health’s environmental research and has expertise in the translation of scientific research for product innovation. 

As an example, Dhir pointed to a study of 500 people that could draw a conclusion that presence of a certain bacteria predicts, for example, weight gain, but another study of a different 500 people by a different organization or institution could draw a completely different conclusion. CODA’s large sample size and numerous data points offer the potential for a standardization that has not previously been achieved, according to Dhir. 

“We saw that the whole field of probiotics, and certainly so much of obsession with gut health, was not actually reflecting the scientific approach that we really thought, that we really wanted to take and pioneer,” added Katz.

Independent experts in the field agree that the science needs to improve. “I see a lot of probiotics there, everybody’s jumping on the probiotic train, like you’ll see probiotic drinks, probiotic food, and I’m like, really? Does it really work? And that, to me, as a scientist, and somebody who wants to improve health is frustrating. So I think increasing scientific rigor, tests, doing these clinical trials, I think is going to be really important,” said Dr. Arpana Gupta, an associate professor at UCLA and co-director of the Goodman-Luskin Microbiome Center. “There’s definitely promise.”

CODA’s first applications are in metabolic health, brain health, longevity, and menopause, research areas chosen because they have already been identified as areas of human health where early CODA data displayed the strongest evidence. 

“Taking those findings [from CODA], and eventually translating them and bringing them to the billions of people that could benefit from that, I think that would be the biggest contribution and that excites me, because then, without CODA, all of these findings would end up in, maybe very nice papers, but just papers. And this way they can be eventually translated into people,” said Segal, who has studied metabolic health and link between the microbiome and body composition.

Segal said the Human Phenotype project has included the collection of dietary logs and medical histories, as well as glucose monitoring records and DEXA (bone densitometry) scans, much of it covering over two decades and offering insight into the changing, and aging, human microbiome. 

“It’s a tremendous step forward in not just the standardization of data, but also the different types of data. That’s why it’s called deep phenotyping, because phenotyping means all these things come together to make the phenotype of a person,” Dhir said. “That’s what CODA unlocks. … There’s things that before were just drowning in noise, in bioinformatics data noise that now come out in such clear signals.”

Seed Health has been working on several efforts around pioneering microbiome science for human and planetary health, and many in the field believe the approach is destined to have wider applications.

“I think the top areas right now are in cancer,” said Dr. Joseph Petrosino, professor at Baylor College of Medicine and Director of the Center for Metagenomics and Microbiome Research, and a member of Seed Health’s scientific advisory board. “The ability to use the microbiome both as a diagnostic as well as a potential therapeutic to help with the responses to various cancer treatments, as well as to avoid some of the side effects of those treatments,” Petrosino said.