Newsletter Signup - Under Article / In Page
"*" indicates required fields
As artificial intelligence (AI) becomes ever more integrated into the world of biotechnology, one of its latest applications is the discovery and design of peptides. This is a rapidly emerging field that moves beyond traditional methods by using machine learning and deep learning to accelerate the discovery, de novo design, and optimization of therapeutic peptides, as well as peptides for areas like food and cosmetics.
In this article, we take a closer look at this emerging field and whether the incorporation of AI could signal the beginning of a golden age for peptides.
Table of contents
What are peptides?
Peptides are short chains of amino acids, which are the building blocks of proteins. They act as messengers in the body, helping to regulate various biological processes such as healing, growing, and repairing cells, and are crucial for many bodily functions, including the formation of proteins like collagen and elastin in the skin.
“I call them the language of life because all interactions between different species, whether it’s plants, humans, bacteria, etc., are through peptide interaction,” said Nora Khaldi, founder and chief executive officer (CEO) of Nuritas, an Irish biotech focused on using AI to discover natural bioactive peptide ingredients.
Khaldi told Labiotech that, from a healthcare point of view, peptides have several compelling advantages over other drug types like small molecules. “They are highly efficacious, they are extremely safe, they are found in nature, and they can be made sustainably. Peptides are so effective because they interact with precise biological targets with remarkable specificity. Peptides don’t accumulate or create toxic metabolites the way some synthetic drugs can.”
How AI can be leveraged to discover and design peptides
Traditionally, biochemical and medicinal chemistry approaches have been used for peptide drug discovery; however, these have relied on experimental and empirical methods, which often entail a significant amount of time, resources, and expertise – particularly when you consider that one of the primary challenges surrounding traditional peptide drug discovery is the vast chemical space that needs to be explored, as the number of possible peptide sequences is exponentially large, making it impractical to test all potential candidates experimentally.
This is where AI comes in extremely handy, as the idea of incorporating the technology into any area of biotech is to speed up certain processes. AI-based peptide design generally involves the use of algorithms that can generate and evaluate large numbers of peptide sequences based on desired properties, such as target affinity, selectivity, and bioavailability. Ultimately, these algorithms can explore large chemical spaces more comprehensively and efficiently than traditional methods.
Nuritas’ AI approach
To provide a specific example of how exactly AI can be used for peptide design, Khaldi explained how Nuritas’ technology works: “Our AI-driven discovery process is quite sophisticated. It begins with our algorithms analyzing massive databases of protein sequences from various natural sources. The AI Magnifier [Nuritas’ technology] has been trained to recognize patterns that link specific peptide sequences with biological activities. It also considers other multiple factors simultaneously, such as the final 3D structure the peptide will adopt, how the peptide will interact with specific biological targets, as well as stability and bioavailability factors. These factors play a part in determining whether a peptide will be effective.”
Once the AI Magnifier predicts promising candidates, Nuritas then validates them in its laboratories. “The beauty of this approach is that we’re guided by our predictive AI rather than trial and error. Traditional discovery might screen thousands of compounds hoping to find one that works. Our AI Magnifier narrows the field dramatically, pointing us directly to the peptides most likely to succeed.”
Khaldi added that, critically, the company’s technology is continuously learning, as every validation experiment feeds back into the algorithms, in turn making predictions more accurate over time. “These efficiencies also help us minimize wasted resources, reduce unnecessary experimentation, and accelerate the path to sustainable, health-promoting ingredients.”
Duke University develops AI platform PepPrCLIP to accelerate peptide discovery
Another example of a recently developed AI platform that designs peptides comes from Duke University. Inspired by OpenAI’s image generation model, their new algorithm rapidly identifies promising peptides for experimental testing, offering a potential breakthrough in treating a wide range of diseases.
Instead of attempting to map the 3D structures of peptides, the Duke researchers drew inspiration from generative large language models. Their solution, PepPrCLIP, combines two key components: PepPr, a generative algorithm trained on a vast library of natural protein sequences that designs new ‘guide’ proteins with specific characteristics, and CLIP, adapted from OpenAI’s image-caption matching algorithm, which screens these peptides, identifying those that best match their target proteins based solely on the target’s sequence.
“OpenAI’s CLIP algorithm connects language with an image. If you have text that says ‘dog,’ you should get an image of a dog,” explained the team leader, Pranam Chatterjee, in a press release. “Instead of language and image, we trained it to match peptides and proteins. PepPr makes the peptides, and our adapted CLIP algorithm will screen those peptides and tell us which ones will make a good match.”
In tests against RFDiffusion, an existing peptide generation platform that relies on 3D protein structure, PepPrCLIP proved faster and generated peptides that were consistently a better match for their target proteins. The platform’s effectiveness was also further validated through collaborations with researchers at Duke University Medical School, Cornell University, and Sanford Burnham Prebys Medical Discovery Institute.
When asked whether there are any challenges when it comes to data, in terms of availability and quality, for training AI models in peptide discovery, Khaldi explained: “Data is absolutely both an opportunity and a challenge in our field. We’ve had to be incredibly strategic about data generation. We’ve invested significantly in creating our own proprietary datasets. Every peptide we test, every biological assay we run, adds to our knowledge base and makes our algorithms more powerful. Further, quality data is paramount. Poor quality data will train poor quality models and generate poor quality products.”
New technologies related to AI and peptides
- Deep Learning Model Discovers Antibiotic Drugs in Extinct Organisms Effective Against Drug-Resistant Superbugs – University of Pennsylvania
- Research programme: peptide therapeutics – ProteinQure
Applications of AI-designed peptides
AI-designed peptides have applications in medicine for developing antimicrobial, antiviral, and anticancer therapies, with the potential to treat conditions like antibiotic resistance, cancer, and other diseases. They are also used in the cosmetics industry for anti-aging and skin regeneration products, and in functional foods for their antioxidant and immunomodulatory properties.
Indeed, food and nutrition are where Khaldi believes AI-designed peptides currently hold the most commercial promise. “We’re on the cusp of a revolution in personalized nutrition, where AI-discovered peptides can be incorporated into foods to provide specific health benefits ranging from supporting sleep to enhancing muscle recovery. What’s particularly exciting is that many of these peptides can be derived from plant-based sources, creating healthier food ingredients while reducing our environmental impact.”
For example, said Khaldi, Nuritas has commercialized two plant peptide ingredients: PeptiStrong, to boost the protein’s effect, and PeptiSleep, for deeper and better sleep quality; both of these are used in formats such as powders, beverages, bars, etc, which “customers are able to enjoy while getting much-needed health benefits.”
AI-designed peptides are also beginning to make headway as therapeutics. In May, biotech company ProteinQure closed an $11 million series A financing round to support the initiation of the first clinical trial for its first-in-class peptide-drug conjugate, PQ203, for triple-negative cancer; the company said that it is advancing what it believes to be “the first AI-designed peptide therapeutic into the clinic.” The first patient was dosed in a phase 1 study of the drug candidate in September.
Another recent event in the therapeutics area that shows how the use of AI and peptides is really taking off was the merger of two AI-focused drug discovery companies, Tandem AI and Pepetual Medicines, to integrate their platforms for small molecule and peptide discovery. The move reflects an effort to expand the use of AI beyond small molecules and into the increasingly popular drug class of peptides.
An emerging field: The beginning of a golden age for peptides?
“We’re standing at the beginning of what I believe will be a golden age for peptides…” said Khaldi. “AI is the key that will unlock it. In the next few years, I expect we’ll see a dramatic increase in the number of AI-discovered peptides reaching the market. The technology is maturing rapidly, and the early successes we are having are validating the approach. This will attract more investment and innovation into the field.”
She noted that there is also a broader shift happening in terms of how industries think about ingredients, with consumers, regulators, and companies all demanding better health solutions, products, and ingredients. AI-designed peptides could potentially fill these demands by delivering “clinically effective, scientifically validated ingredients that are also environmentally responsible.”
As AI is incorporated more and more into drug discovery, the market size is set to grow substantially – estimated at $6.31 billion in 2024, it is projected to increase to $16.52 billion by 2034, growing at a compound annual growth rate (CAGR) of 10.10% from 2025 to 2034.
And, as peptide drugs become increasingly popular with biotech and pharma companies, it is likely that the use of AI to discover and design them will also become the norm.
“Ultimately, I envision a future where AI-designed peptides are a routine part of our lives. I see a world where we use nature’s own peptides, intelligently discovered and optimized, to improve billions of lives,” stressed Khaldi.
Advancements in antimalarial drug discovery and development
