Nature’s Most Promising Bioactives to Speed Up Evidence-Based Drug Discovery

By Deborah Borfitz 

August 19, 2025 | Natural products and medicines mimicking substances found in nature have been used clinically for well over 100 years and could help remedy what ails today’s alarmingly long and relatively unproductive drug discovery process. Since bioactive substances from natural resources have a track record of use against disease, they are a compelling starting point for evidence-based discovery of novel human therapies, according to Michael Krogh Jensen, Ph.D., CEO of Biomia, a Copenhagen-based synthetic biology startup that specializes in the discovery and manufacturing of plant-inspired human therapeutics, assisted by neural networks and machine learning.  

As it is, more than 40% of clinical trials fail due to a lack of therapeutic effect, Jensen says. By starting from natural products used in traditional medicine and applying expertise in synthetic biology and AI-driven drug discovery, Biomia biomanufactures novel molecules “inspired by nature” that are not feasibly manufactured at scale,” says Jensen. Biomia is attempting to develop these medicinal candidates to address unmet medical needs for central nervous system (CNS) disorders including mental health, pain, and addiction. 

Biomia’s current area of interest is monoterpene indole alkaloids (MIAs), a large and chemically diverse group of plant-specific compounds known for their diverse biological activities. Examples include stemmadenine, found in plant species abundant throughout tropical and subtropical regions of the world; yohimbine, from the bark of a West African tree; and ibogaine, traditionally pulled from a rainforest shrub native to central western Africa. Instead of sourcing them from their natural plant origins, with extraction yields well below 0.001%, Biomia biomanufactures MIAs and derivatives with improved pharmacological properties in the lab and at partner biomanufacturing sites. 

MIA-containing substances from plants have for centuries been dosed orally as part of traditional medicine, Jensen says, but the extracts were not “cleaned up in any sort of clinical sense” and therefore contain many different bioactive compounds. “What we are homing in on are the causal ingredients for therapeutic benefit, optimized for their pharmacological properties, and delivered as safe and efficacious new drug candidates.” 

Jensen will showcase the company’s approach during a presentation at the upcoming Discovery on Target conference in Boston. 

“Biomia envisions a new golden age of drug discovery and development, where AI and biotechnology enhance nature’s evolutionary chemistry to accelerate the design, perfection, and scaling of life-changing therapies for the millions of patients suffering from CNS disorders,” he says. “By uniting discovery and manufacturing under one platform, Biomia aims to not only vastly expand the utilizable chemical space suitable for human healthcare but also pioneer the scalable sourcing of a new generation of safe, effective, and affordable therapies.” 

Two-Pillared Drug Discovery Engine

Biomia’s drug discovery engine has two interconnected, AI-assisted foundational pillars, says Jensen. One focused on generating training datasets linking molecular fingerprints of chemistry with critical pharmacological properties for the exploration of novel, safe, and efficacious drug leads. The other is biomanufacturing using modeling, synthetic biology, and fermentation for optimizing yeast to produce new medicinal candidates.  

For biomanufacturing purposes, the company starts by copying the “MIA assembly lines” found in plants (e.g., pieces of DNA that encode enzymes turning amino acids and sugars into molecules) into yeast. The protective mechanisms found in plants protect against things like caterpillar attack, pests, and UV radiation—meaning they aren’t necessarily optimized for human healthcare, he continues. But genes can be taken from plants and other organisms and put into the genome of baker’s yeast to produce both approved medicines and novel drug candidates, akin to the process of brewing beer where sugar is being fermented into ethanol and carbon dioxide. 

“Yeast is the chassis from which we are manufacturing natural products, but our engineered yeasts also have the ability to augment and improve nature-inspired small molecules with optimized pharmacological properties,” says Jensen. The company is effectively reprogramming yeast cells to become microscopic medicinal cell factories, not distantly related to yeasts that have been used for decades to make commercial therapies, including insulin and more recently Ozempic. 

“It’s a biotechnological workhorse for a number of small molecules and biologics,” Jensen says. In the U.S., for example, Antheia, a biotech based in California, is using synthetic biology and yeast fermentation to brew medicinal candidates as key starting materials for pain management using “assembly lines” inspired from the opium poppy plant. 

Biomia uses AI logic to ensure the highest quantities of certain "privileged chemical scaffolds”—structural subunits that are more likely to exist in successful drug candidates—are produced, says Jensen, by modelling and exploring the correlation between the design of the DNA-encoded assembly lines inserted into the genome of yeast and the production profiles from yeast fermentations. Work in this realm began six years before the company’s 2022 founding as a spin-out from the Technical University of Denmark. 

“These models are helping us predict the design of engineered assembly lines that enable higher production of medicinal candidates of interest or limit the production of therapeutically irrelevant intermediates,” Jensen explains. The best-performing medicinal cell factories are cultivated, using scalable fermentation processes, in bioreactors or fermenters. 

The second, complementary pillar of Biomia’s drug discovery engine is the use of AI models for drug design, fueled by datasets on bioactivities of small molecules, both those found in nature and those already in clinical use, says Jensen. The AI models train on the datasets to learn the relationship between chemistry and function and suggest modifications to nature-inspired medicinal candidates predicted to yield further improved pharmacological properties.  

“Everything we are doing around testing bioactivity and other critical pharmacological properties [e.g., pharmacokinetic and bioavailability] we link to the molecular structure,” Jensen says. In this way, Biomia has created tools for predicting certain types of pharmacological properties based on modifications of the chemical structure of the small molecules being made and is currently exploring the chemical space around bioactive small molecules to further optimize the therapeutic potential of its pipeline lead assets. 

Translational Efficacy 

For Biomia’s mental health program, privileged scaffolds of a MIA called alstonine serve as a starting point. Alstonine is the most abundant MIA found in a plant extract with therapeutic effect in humans for treating psychosis, says Jensen. “However, due to the minute concentrations of alstonine found in plants, and a chemical complexity prohibitive of scaled chemical synthesis, alstonine and many other MIAs have remained a missed opportunity in drug discovery.”  

Alstonine has not been tested properly “for precisely the reasons Biomia exists,” according to Chief Scientific Officer Frederick Westhorpe. Without the company’s biomanufacturing method, it wasn’t possible to produce the compound at scale for drug discovery and development. 

In rodent models of schizophrenia, company scientists have demonstrated a reduction in symptoms following treatment with alstonine. “However, as these natural compounds are not designed by the plants to be efficacious in human disease, we’re looking at optimizing them in terms of making them the best they can be to treat mental health disorders and enhance cognitive health in patients,” Westhorpe says. “The final drug product we hope to produce will not be alstonine, but something inspired by alstonine that, for example, has the right interactions in the brain or gets into the brain better.” 

Biomia's platform can “access therapeutically relevant scaffolds with human validation for human diseases” beyond schizophrenia, he adds. Starting from a different MIA, translational efficacy has also been demonstrated in rodent models of pain relief (e.g., acute and post-surgical pain). “Optimized lead molecules have proven superior to the natural product starting point, underscoring the missed opportunity in natural products drug discovery and validating Biomia’s drug discovery engine in the sourcing and discovery of novel medicinal candidates with translational efficacy.” 

Modification Know-How

Biomia is leveraging data science work done by Jensen’s group prior to his leaving Technical University of Denmark but looks to build out the company’s computational team or partner. “It is rooted in the DNA of Biomia to be data-driven ... whether we are talking about bioprospecting for novel enzymes, mining clinical data packages, chemistry, or how to optimize manufacturing capabilities,” says Jensen. 

The company’s “Genentech moment” came at the end of 2022 when scientists engineered yeast cells to synthetically produce vinblastine, a complex small molecule used to treat childhood leukemia, he reports. The process involved 31 enzymatic reactions and adding approximately 100,000 DNA bases into the yeast genome and knocking out a lot of other genes as well. At the time, “it was the most extensive assembly line of small molecule chemistry that had ever been made in baker’s yeast,” and the title likely still stands. 

The company raised about $10 million in soft funding prior to launch, and in September 2023 received its first $3 million pre-seed round of venture capital, topped off with a grant from the European Research Council, Jensen says. Biomia is currently in its second fundraising round targeting the nomination of the company’s first development candidate and early, investigational-new-drug-enabling studies. 

Although still in pre-commercial mode, the short-term business strategy is to forge partnerships with medium to large pharma companies, preferably with established CNS teams, to help “drive forward some of the assets that we have already derisked,” he says. Longer term, Biomia could help optimize some of the assets of pharma companies and co-develop those compounds through early- to late-stage clinical trials. 

Bigger picture, it’ll be important to reflect on the “broken logic” currently at play in drug discovery and development globally, says Jensen. “We are still looking at timelines of 10 years, budgets on the upside of $2 billion, and failure rates that are hovering above 90% in the clinical stages. That needs to change, and I think starting from molecules that have therapeutic efficacy in humans for complex indications such as CNS puts Biomia, and a handful of other companies, in a favorable position for bringing forward novel, safe, and efficacious drug candidates for the benefit of patients and caregivers.”