Most anti-infective drug classes in clinical use today trace back to a natural product.
Photograph · Unsplash
Science
Our work rests on a claim that is not controversial in natural-product chemistry and is still underused in practice: the molecules that living organisms make to act on other living organisms are, on average, far better starting points than molecules designed from first principles.
Most anti-infective drug classes in clinical use today trace back to a natural product.
01 — The chemical argument
Evolution has already run the screen.
A synthetic compound library is large but chemically narrow. It is built from the reactions that are convenient to run at scale, which means it clusters in a small, flat region of possible chemical space — heavy on aromatic rings and amide bonds, light on stereocentres, rings fused in unusual ways, and the three-dimensional complexity that lets a molecule bind something difficult.
Compounds made by living organisms sit somewhere else entirely. They carry more stereochemistry, more sp3 character, more structural novelty per molecule. More importantly, they were not made at random. A compound an organism produces to defend itself has been under selective pressure for a very long time, against real targets, in the presence of real competitors. It arrives pre-optimised for a job.
That is why the record is so lopsided. The overwhelming majority of antibacterial and antifungal drug classes in clinical use derive from natural products or are semi-synthetic modifications of them. Decades of purely synthetic screening did not repeat that result.
02 — Resistance biology
Most modern anti-infectives inhibit one enzyme. That precision is a virtue for safety and a liability for durability: a single point mutation in the binding site, or the acquisition of one efflux pump or one hydrolytic enzyme, can eliminate activity outright. Because the selective pressure acts on one locus, resistance emerges quickly and transfers easily between organisms.
Compounds that act on structure rather than a single catalytic site behave differently. Where activity depends on the physical organisation of a membrane or a cell envelope, escape requires the organism to remodel something fundamental to its own viability. The mutational cost is high and the fitness penalty is often severe, which is why resistance to this class of chemistry has historically been slow to appear and slow to spread.
This is the general property we look for, and it is one of the reasons naturally occurring defensive chemistry is interesting to us. Organisms that have been producing antimicrobial compounds for millions of years, in environments dense with competing microbes, have not been outrun by resistance in the way our clinical antibiotics have. Something about the chemistry they arrived at is harder to escape.
Understanding why a compound is difficult to resist is as much a part of our work as establishing that it is active. Activity without a mechanism is a result. Activity with a mechanism is an asset.
03 — What changed
Natural-product screening was not abandoned because it stopped producing hits. It was abandoned because the work after the hit was punishing. Separating a genuinely novel compound from one already described required painstaking manual comparison, and teams repeatedly spent months isolating something the literature already contained.
Determining a structure was slow and instrument-hungry. Extracts were chemically complex, activity was often distributed across several components, and throughput was limited by hand. None of this was a failure of the biology. It was a failure of measurement economics: the cost of knowing what you actually had was higher than the value of finding out.
High-resolution mass spectrometry, coupled to chromatographic separation, resolves the components of a complex extract and returns accurate mass and fragmentation data in a single run. What once took months of dedicated effort per compound now takes a well-designed sequence of experiments.
Public spectral and sequence databases make it possible to recognise known chemistry early, before committing resources to it — the specific failure that made the original approach uneconomic. Computational tools rank candidates on physicochemical grounds before anything is cultured, and standardised assay panels test activity against clinically relevant organisms at real throughput.
The chemistry is the same chemistry. The organisms are the same organisms. What has changed is the price of reading the result — and that single change is what makes this a viable research programme today rather than a historical footnote.
04 — Position
We are not betting on a new technique. We are running a proven one at a cost that was never available to the people who set it aside.
This distinction matters to how we describe ourselves. We are not claiming a novel discovery platform or a proprietary methodology that has never been tried. The methodology is well established, extensively published, and taught. Our position is that it was retired for economic reasons which no longer hold, and that the organisms and geographies it was never systematically applied to remain largely unexamined.
The scientific risk in our work is therefore not whether the method works. It is the ordinary risk of any discovery programme: that a given source yields nothing novel, that an active compound proves unstable or non-selective, that a promising result does not reproduce. Those risks are managed by running many parallel lines of enquiry and by killing weak ones early.
This page describes our reasoning, not our compounds. Sources, extraction conditions, structures, sequences, assay results and programme status are not published here.
That is not vagueness — it is standard practice. Early-stage chemistry is protected by remaining unpublished until it is protected by filing. We are glad to go considerably deeper with research collaborators and prospective partners under an appropriate agreement.