For many women, changes in hair can feel deeply personal. A widening part, a thinner ponytail, or more visible scalp can affect not only how hair looks, but also how confident someone feels day to day. As some would put it, hair… is everything.
That’s why our discovery process started long before a serum was bottled. It began with a question: Can we use modern computational discovery to accelerate the search for better molecules intelligently, and precisely?
The short answer is yes. We built a proprietary artificial intelligence (AI)-powered virtual screening pipeline designed specifically for hair follicle biology to answer this question. This is part 1 of a 4-part series on NOVOGRO™. To jump ahead, please explore the following links:
From Trial-and-Error to Smarter Discovery
Traditional molecule discovery was generally slow. Scientists screened molecules, meaning they tested them for a desired effect, one at a time, making the discovery process slow and labor-intensive. They learned rather slowly, but gradually, which kind of molecules looked promising and which did not. This process worked, but it took a long time, used significant resources, and oftentimes missed valuable candidates hidden in a much larger subset of ingredients.
In recent years, AI-based methods for molecule discovery have drawn much attention and made huge progress. In fact, the 2024 Nobel prize in chemistry was awarded to Google’s and University of Washington's scientists who were pioneers in artificial intelligence and protein design, two important advancements that made AI for molecule discovery possible.[1]
To discover NOVOGRO™, our scientists used advanced technology with a focused, highly targeted approach.
We created a proprietary discovery pipeline that combines multiple AI methods, each evaluating molecules from a different perspective. Each AI method has its own assumptions and potential blind spots, just as a microscope, camera, and measuring tool each reveal different details about a particular object but may miss something the others can detect. This means that a given method may favor certain molecules in some contexts, but not others.
Because the models predict which molecules interact with a given target, relying on only one method can introduce bias. Therefore, by comparing results across multiple methods and prioritizing molecules that consistently rank highly, we reduce method-specific bias and gain greater confidence in the candidates selected for laboratory testing.
By combining several approaches, our pipeline helped us narrow a very large search into a focused set of candidate molecules with the strongest potential.
Screening Millions of Molecules in Hours
With this powerful pipeline, RE:YOU scientists began with a clear focus: understanding hair follicle biology to identify root causes of hair thinning where new molecules could make a meaningful difference. Because hair thinning rarely results from addressing a single biological mechanism, the team focused on multiple complementary pathways - one supporting the hair follicle itself and the other supporting the environment around it. Together, these two complementary mechanisms became the foundation of RE:YOU’s dual-path discovery strategy.
The first direction focused on dermal papilla cell (DPC) health. DPCs sit at the base of the hair follicle and act as an important signaling center that helps regulate hair growth, hair shape, size, and follicle behavior[2] [3]. The second direction focused on supporting the scalp environment that the hair follicles live in via the PHD2/HIF-1α pathway. When activated, this pathway can support the growth of blood vessels that deliver oxygen and nutrients to the scalp, nourishing the surrounding environment for active hair follicles[4][5].
Our AI searched approximately 20 million molecules and predicted ~100 promising NOVOGRO™ candidates for each pathway. We also used well-studied molecules as benchmarks, helping us identify new candidates that perform even better. For example, we compared our candidates with well-known hair-growth ingredients such as minoxidil and PP405. This helped us look for molecules that could deliver stronger results while avoiding some of the drawbacks associated with existing options. By combining large-scale screening with direct comparisons to known molecules, we selected the most promising candidates for further laboratory testing. For full details regarding our pipeline, please view our manuscript on BIORXIV.
From Prediction to Testing and Validation
A molecule that looks good on a computer still has to prove itself in the real world, on real human hair follicle cells. That’s why the next step was experimental testing. We tested the top candidates to evaluate whether the predictions translated into measurable biological activity. This stage is critical. Computational discovery can help guide the search, but lab testing is what turns a prediction into evidence.
For more information about our experimental testing, click here: How we tested and validated NOVOGRO™.
At RE:YOU, we never settle for molecules that are “good enough.” Finding a promising molecule is only the beginning. We continue refining and optimizing them to unlock their full potential. It should be optimized with the intended mechanisms in mind, while also being suitable for a topical serum
- Dr. Yang Li, PhD, Lead Computational Scientist on the NOVOGRO™ research team.
After experimental testing and validation, our scientists used computational medicinal chemistry to improve the original candidates. In simple terms, this means we studied how small changes to molecular structure could improve performance. We looked for ways to make the molecules more soluble, more stable, and more active[6].
This is an intentional design process from which we developed three new molecules that form the actives for RE:YOU’s Dual Path Hair Revival Serum:
- NOVOGRO™-623 and NOVOGRO™-624, designed to support hair follicle health[6].
- NOVOGRO™-273, designed to nourish the scalp environment around the hair follicles[6].
Each molecule came from a process that combined large-scale computational screening, consensus-based selection, experimental testing and validation, and careful & intentional optimization.
Why This Matters for Women’s Hair Thinning
The entire category of female hair thinning has seen surprisingly little ingredient innovation, with many products relying on the same familiar approaches despite how much our understanding of hair biology has advanced. RE:YOU was created to move the category forward—using AI to explore new ingredients designed around the many biological factors that influence hair and scalp health, rather than simply repackaging what already exists.
Focusing on the multiple complementary directions that are closely related to hair biology, RE:YOU scientists built the state of art AI screening platform for female hair thinning [6]. With our proprietary platform, RE:YOU scientists searched through millions and millions of the molecules, found the most promising candidates, and optimized them to perform the best. The result is a new generation of molecules for hair science: designed to support healthier-looking, fuller-feeling hair through thoughtful discovery, not trial and error.
Because when it comes to your hair, you deserve more than hope in a bottle. You deserve science designed with care.
Frequently Asked Questions
What is Artificial Intelligence (AI)?
In molecule discovery, AI helps scientists quickly study millions of possible molecules. It can find patterns, predict which molecules may work best, and identify promising candidates for laboratory testing. This saves time and helps researchers focus on the molecules most likely to be effective and safe.
What is virtual screening?
Virtual screening is a way for scientists to evaluate large numbers of molecules using computer-based methods before testing them in the lab. It helps narrow millions of possibilities into a smaller group of promising candidates.
How did RE:YOU use AI?
RE:YOU created a proprietary virtual screening pipeline that combined several AI methods. Each method evaluated molecules from a different perspective, and our scientists prioritized candidates that consistently performed well across multiple methods.
Why use several AI methods instead of one?
Every method has its own strengths and potential blind spots. Comparing results across several methods helps reduce reliance on any single prediction and gives scientists greater confidence in the molecules selected for further testing.
How many molecules did RE:YOU screen?
RE:YOU’s virtual screening pipeline evaluated millions of molecules.
What are the major directions for the AI and virtual screening?
Dermal papilla cells (DPCs) are located at the base of the hair follicle and help coordinate signals involved in the hair cycle and follicle activity. Supporting the health of these cells is one of the key focuses behind NOVOGRO™-623 and NOVOGRO™-624. The PHD2/HIF-1α pathway is associated with the body’s response to oxygen and nutrient needs. It may help support signals related to vessel growth, nutrient delivery, and a healthier environment for the scalp. This biological direction guided the development of NOVOGRO™-273*.
Were the molecules selected only by AI?
No. Computational screening was only the first step. The selected candidates were tested experimentally to determine whether their predicted potential translated into measurable benefits.
What are NOVOGRO™-623, NOVOGRO™-624, and NOVOGRO™-273?
They are molecules discovered by RE:YOU scientists to help with hair care. NOVOGRO™-623 and NOVOGRO™-624 were developed with a focus on dermal papilla cell health. NOVOGRO™-273 was developed around PHD2 inhibition and HIF-1α activation, with the goal of supporting vessel growth, nutrient supply, and a healthy follicle environment*.
Does AI replace laboratory research?
No. AI helps scientists search more broadly and prioritize where to focus, but laboratory testing remains essential. RE:YOU’s discovery process combines computational screening, human scientific judgment, experimental validation, and molecule optimization.
References
- Zhang, K. et al. “Artificial intelligence in drug development.” Nature Medicine, 2025.
- Zhang, H. L. et al. “Dermal Papilla Cells: From Basic Research to Translational Applications.” 2024.
- Driskell, R. R. et al. “Hair follicle dermal papilla cells at a glance.” 2011.
- Seo, J. et al. “Hypoxia inducible factor-1α promotes trichogenic gene expression…” 2023.
- Yum, S. et al. “Minoxidil Induction of VEGF Is Mediated by Inhibition of HIF-Prolyl Hydroxylase.” 2017.
- *Qu, Z. et al. "AI-enabled discovery of small molecules targeting complementary pathways for hair follicle rejuvenation." 2026.
