Health claims on nutraceuticals can be limited due to the need for scientific evidence, which traditional products often lack. Here, we introduce a new approach to nutraceutical formulation that marries biology with engineering to address specific health issues – and the results are looking good.
The nutraceutical industry has long worked from a familiar logic: find an ingredient with a promising safety profile and some supporting research, run a trial, and if results look favourable then bring it to market. The ingredient’s efficacy is often assumed rather than established. The starting point is the ingredient, not the biological problem it is meant to address.
While this approach has produced useful products, it has also created a category wherein efficacy is hard to predict, evidence is often thin and formulations are built around population averages that may not reflect the biology of the people actually using them. Growing understanding of the gut microbiome is now facilitating a different – and arguably more appropriate – starting point for nutraceutical development.
The starting point is the ingredient, not the biological problem it is meant to address.”
The problem with population averages
The gut microbiome is not uniform. Composition varies between individuals, shaped by diet, geography, genetics and environment. A 2025 study in Nature Microbiology found high interpersonal and geographical variability in how gut bacteria metabolise hundreds of plant compounds. The same input, processed by different microbiomes, can produce very different outputs.
Most nutraceutical formulations are not built with this in mind. They draw on broad population data and the working assumption, rarely made explicit, is that what works for the average will work for most people. Yet in areas where the gut microbiome plays a direct role in health outcomes, that assumption is getting harder to defend.
The microbiome has now been implicated in metabolic health, immune regulation, neurological function, skin conditions and ocular health. Across all of these functions, the question of whether a formulation actually targets the relevant biology, rather than a general population average, is one that must be addressed.
A different starting point
The alternative approach is to begin with the biology – not the ingredient.
Instead of starting with a compound and testing whether it produces a measurable effect, a target-first approach begins with a specific biological target: a metabolic pathway, immune mechanism or microbial function that is disrupted in a given health condition.The ingredient becomes a means to a defined biological end and not the default starting point.
For the gut microbiome, this means identifying which microbial pathways are disrupted, which metabolites are insufficient and which food-grade inputs can support their production. The formulation is then built around what the biology actually needs, rather than what has historically been associated with a general health outcome.
It sounds like a small shift. In practice, it changes almost everything about how a formulation is developed, what it is designed to do and how its effects can be clinically measured and validated.
Computational modelling in practice
This is where biology and engineering meet. And that intersection is changing what formulation can look like.
Genome-scale metabolic modelling makes it possible to simulate how gut bacteria behave across thousands of real microbiome profiles, rather than relying on broad population data. The models identify which metabolic pathways are disrupted in a given condition and which food-grade compounds can be converted by gut bacteria into the molecules the body is not producing in sufficient amounts.
From there, an AI retrosynthesis engine works backwards from the biological target to find the smallest set of food-grade precursors that can do the job. Every ingredient ends up with a defined biological role, established before any clinical study begins.
At Enbiosis, this is the foundation of how formulations are developed:
- · A digital twin of the gut microbiome is built for a specific health condition
- · Simulations run across thousands of diverse microbiome profiles
- · The result is a formulation shaped around the biology of that condition – not a population average.
Clinical validation
Computational modelling changes how a formulation is designed but it does not replace the need for clinical evidence.
The first formulation developed through this engine targets dry eye – a condition where gut microbiome composition has been increasingly implicated in tear film dysfunction.
A prospective pilot study of a food-grade oral formulation targeting the gut–eye axis demonstrated meaningful changes across multiple clinical endpoints including tear production and patient-reported symptom scores. Participants took the formulation once daily for eight weeks, with effects maintained through a follow-up assessment at week sixteen. A randomised controlled trial is currently underway. (Article submitted for peer review)
Beyond dry eye
This new approach to nutraceutical formulation has seen promising results with its initial application to dry eye, but it is not specific to one condition.
The same computational framework applies wherever the gut microbiome plays a meaningful role in health outcomes. Active development is underway across metabolic health, where gut composition has been linked to insulin resistance and glycaemic regulation; skin conditions such as psoriasis, where gut–skin axis research is gaining traction; and neurological health, where connections between microbial composition and cognitive function are increasingly well characterised.
The logic does not change. A digital twin is built for the condition; simulations run across diverse microbiome profiles; the output is a formulation shaped around that specific biology – not a general wellness claim.
This is how an engine-based approach to microbiome formulation looks in practice. Not one product, but a systematic method for developing formulations across health areas, using the same biological and computational foundation.
Where you start determines what you can honestly say about what a product does and for whom.”
What this means for the industry
The questions being asked of nutraceutical companies are changing. Regulatory expectations are rising, retailers are more demanding, and the partners who matter are pushing harder on mechanism and evidence.
A formulation built around a defined biological target, validated computationally and tested clinically is able to answer those questions. Where you start determines what you can honestly say about what a product does and for whom.
Ingredient-based formulation is not going away. But the bar is moving. For companies working in health areas where the gut microbiome plays a meaningful role, the question is no longer whether a microbiome-targeted approach is scientifically credible; it is whether the formulation is actually designed to reflect that biology.
Conclusion
Nutraceutical formulation is not standing still. The science around the gut microbiome is moving fast and the gap between what biology can now tell us and what most formulations are actually designed around is getting harder to ignore.
A target-first approach, grounded in computational modelling and supported by clinical investigation, is no longer a mere theoretical proposition. The evidence is building. The tools are there. And the first products designed this way are beginning to reach the market.
For technical leaders in food and beverage, the practical question is not whether microbiome science matters to formulation. It clearly does. The question is whether the formulation process is set up to use what that science can now offer.
Disclosure: The authors are employed by Enbiosis Biotechnology.
No comments yet