Nuritas 如何運用 AI 在實驗室尋找有益健康的新肽
How One Company Brings AI Into the Lab to Find New Peptides for Health
都柏林生物科技公司 Nuritas 運用 AI 模型篩選天然肽,尋找可能有益健康的候選肽,再由科學團隊在實驗室測試。公司以自有系統和專有數據訓練模型,按特定生物目標搜尋,並表示傳統方法要花 3,000 萬年才能研發出 PeptiStrong。展開臨牀測試和確認安全、功效及商業可行性,是產品上市前的重要環節。
It had been since high school that I’d donned a white coat and looked at cells under a microscope. But at Nuritas’s labs in the heart of Dublin, Ireland, I did just that to learn more about the intersection of AI and peptides.
Nuritas, a Dublin-based biotechnology company, brings together AI models, clinically proven peptide ingredients and cutting-edge science to accelerate the discovery of peptides that can be used for products that boost human health. Founded in 2014 by Nora Khaldi, a mathematician with a doctoral degree in molecular evolution and bioinformatics, the Nuritas team consists of scientists, AI specialists and industry experts to create health products for things like skincare, nutrition, muscle health, better sleep and aging.
Last month, I toured Nuritas, a beautiful converted bank that overlooks the city with views of Mansion House, the oldest Georgian house in Ireland and the Guinness Storehouse at St. James Gate. I spoke with Khaldi and two of the team’s top scientists to learn more about their patented products, peptide discovery and how AI systems fit into the process.
More from CNET
Peptides have been all the rage in the health sector, driven by the popularity of GLP-1 medications like Ozempic and Wegovy. These drugs use synthetic peptides inspired by those naturally found in the saliva and venom of the Gila monster, a lizard native to parts of North America, for weight management, cardiovascular risk reduction and treatment of Type 2 diabetes. GLP-1s are increasingly popular, as a survey last year found one in eight US adults (12%) were reportedly taking these drugs.
Since then, peptides have been a major focus for drug developers.
At the most basic level, peptides are just short chains of amino acids that can act as biological signals. For instance, insulin is a natural peptide that regulates blood sugar, but peptides can be used for healing, hormone regulation and cell signaling. More than 7,000 peptides naturally exist in the human body. But since amino acids can also be found in plants, animals and microbes and chain together in a multitude of permutations, the number of peptide sequences that can exist in nature is practically limitless. Potentially trillions of combinations await discovery, and some could offer profound health benefits to humanity.
That’s the core of Nuritas’ work: Finding these unique natural bioactive peptides and unlocking whether they can support our health.
Nuritas has looked to nature to derive peptides, focusing on food sources like milk, rice, fava beans and other plants, as well as algae and fungi. The trick, though, is that sifting through trillions of combinations at scale can take an exorbitant amount of time with human researchers alone.
That’s where the AI models come in.
Khaldi put the timeline into perspective, saying that creating one of their products, PeptiStrong, a peptide discovered in fava beans, would have taken 30 million years with traditional methods.
“I think for the first time ever in history, we are now really capable of deciphering what nature has developed over hundreds of millions of years through AI,” Khaldi said. “That means we can access molecules that really optimize our health in a way we couldn’t before.”
Nuritas’ AI models can sift through those trillions of combinations and provide candidates for the Nuritas team to test in the labs and collect data for further research. Brian Keogh, the company’s chief science officer, and Robert Radford, head of discovery, took me through the laboratories, describing how the company uses its own AI systems trained on proprietary data and on specific biological targets, such as inflammation or cortisol reduction, rather than relying on off-the-shelf LLMs. The team says they spent close to eight years in the research and development stage to build the data on cellular responses, animal models and human clinical trials.
The Nuritas team told me how they start with a biological need and how peptides act as a signaling mechanism for how cells communicate with each other. The goal is to optimize that specific signal, narrow the types of peptides they’re looking for using AI, and then explore nature for those peptides
“We use nature as our chemical lab because nature has evolved over hundreds of millions of years and has had the chance to optimize solutions,” Khaldi said. “It’s just that we as humans have not been able to access those molecules.”
There are three ways of producing a peptide: synthetically, by hydrolysis and through bio-fermentation.
Hydrolysis works by activating peptides (short chains) that are dormant while embedded in larger protein chains. The scientists use enzymes to cut them into smaller fragments, thereby activating peptides within the source materials. Bio-fermentation clones the peptide in yeast or bacteria.
The team uses different methods based on the focus area: Synthetic and bio-fermentation processes are used in cosmetics, while hydrolysis is used in the food sector.
This means the peptides, as impressive as they may be in the lab, must be commercially viable. Before any product hits the market, it must be thoroughly researched and supported by clinical testing to ensure it’s safe to use in food, beneficial to human bodies, and, well, tastes good if it’s going to be in a food or drink. Khaldi said the ingredient must be stable in formulation. If a peptide ingredient is put into an acidic beverage or if it’s cooked, baked or digested, then there are many other enzymes that can break the peptide’s activity.
Toward the end of my tour of Nuritas, I was able to taste-test a product that uses Nuritas’s PeptiStrong, a muscle health supplement. I was given a sample of a ginger and turmeric shot, a concentrated drink about the size of one or two sips made from freshly pressed ginger root, turmeric root and citrus juice, from the brand Good Shots. I’ve had numerous ginger and turmeric shots before, liking the purported immune support and anti-inflammatory properties, but never loving the bitter aftertaste. But unlike shots I’ve had in the past, this featured a daily dose (2.4 grams) of PeptiStrong, and the usual sharp kick that makes going back for another sip difficult was noticeably absent.
“Everybody’s talking about peptides at the moment because it’s a hot topic, and they’re all talking about injectables,” Radford said. “But this is a dietary peptide, a food supplement that switches on the pathways that control muscle protein synthesis.”

Beyond PeptiStrong’s ability to reduce muscle protein breakdown, increase protein synthesis and aid recovery, Nuritas has developed several other products, including PeptiYouth, an anti-aging skin care peptide, and PeptiSleep, which is intended to reduce cortisol levels to improve sleep quality. These peptide ingredients are found in over 100 consumer products on the market, from tablets to powders to creams.
Khaldi pointed out that there is certainly hype around peptides, especially GLP-1s, but after the dust settles, only the products that work and are backed by science and clinical studies will win out. “I think ultimately what’s going to count is: Does this thing work? Where’s the data?” she said.
Clinical results and proven science are key to Nuritas. The team explained that products like PeptiStrong went through three trials, starting with 30 to 60 participants and then moving into larger trials with 100-plus participants. Because it’s a natural food supplement rather than a pharmaceutical, a product like PeptiStrong is not subject to US Food and Drug Administration approval in the same way a drug would be. Instead, Nuritas is bound by the FDA’s generally recognized as safe (GRAS) standard, a designation they have received.
Looking ahead, Nuritas says it’s interested in developing a range of products across the health spectrum, from mental health and cognitive decline to metabolism, inflammation, gut health and cosmetics. With each new product, the Nuritas mission remains the same: leverage peptides to optimize human health prior to disease.
“My interest has always been: how do we use AI to understand human biology and try to optimize it before we get sick, so we stay healthier for a lot longer?” Khaldi said.
Editor’s note: Macy Meyer’s travel costs for the International AI Summit were covered by Enterprise Ireland. The judgments and opinions of CNET are our own.
來源:CNET · cnet.com