Blog

By Jean-Baptiste Doucet  |  April 2026

Most of my week looks like batch planning, client timelines, and coordinating between our teams and the developers we work with. That is the project manager’s world at a CDMO like Biose, and I am generally comfortable in it.

Copenhagen was a good reason to step out of it.

The 2026 Microbiome Times Partnering Forum Europe brought together two full days of presentations spanning Harvard research labs, clinical-stage start-ups, investor perspectives, and a few genuine surprises – including a session on gut bacteria that absorb PFAS, and another on soil consortia improving crop yields. I went as a listener – curious, not expert. Here is what stood out.

From 100 subjects to 100,000

The opening panel set the tone. Four voices – Denise Kelly of Seventure Partners, Curtis Huttenhower from Harvard, Jeroen Raes from KU Leuven, and Seth Bordenstein from Penn State – were each asked to say where they think the microbiome field actually stands today.

Raes put it in numbers: from MetaHIT cohorts of around 100 subjects fifteen years ago to studies now reaching 100,000 participants. The field has moved from a community under construction to one entering a serious validation phase – clinical and preclinical data converging, AI and machine learning increasingly used to turn associations into something closer to causality.

Bordenstein added a dimension worth noting: the next wave goes beyond human health. A One Health framework – where host, microbiome and environment are studied as a system, not separate chapters. Two examples he gave: oral microbiome transplants being explored to prevent dental caries, and Wolbachia-based approaches to disrupt mosquito-borne viral transmission. Neither felt far away.

The question is no longer just ‘what?’ – it is ‘how?’

One of the clearest threads across the two days was the shift from composition to function. Huttenhower’s standalone session opened with a figure worth sitting with: only around 25% of microbial genes are well annotated. The rest – a functional dark matter inside every gut microbiome – remains largely uncharacterised.

His team’s Human Microbiome Bioactives Resource (HMBR) is being built to address this. One application stood out: understanding why some IBD patients do not respond to 5-ASA, one of the most commonly prescribed treatments for ulcerative colitis. The answer, in many cases, is a microbial enzyme inactivating the drug before it can act. Identifying the organisms responsible opens the door to microbiome-aware prescribing.

Andrew Bartko from UC San Diego’s Center for Microbiome Innovation took a related angle. Around 25% of tested drugs are modified by gut microbial communities – via dihydroxylation, demethylation, hydrolysis, and other pathways. Donepezil in Alzheimer’s. 5-ASA in IBD. Knowing which organisms interfere, and how, could convert non-responders into responders without changing the molecule.

The clinical data is starting to arrive

If there was one session that stopped the room, it was Nikole Kimes presenting results from Siolta Therapeutics‘ ADORED study. Their oral LBP candidate, STMC-103H, was given to at-risk infants over 336 days to prevent the atopic march – the sequential development of atopic dermatitis, food allergy and asthma. Among infants who completed one year of treatment: a 64% reduction in the risk of physician-diagnosed atopic dermatitis (23.3% vs. 43.1% on placebo) and a 77% reduction in the risk of physician-diagnosed food allergy (4.7% vs. 16.7%). Good tolerability, no major safety signal. Numbers worth remembering.

The session that followed felt closer to home. Sam Possemiers of MRM Health and Adrien Nivoliez, our CEO, took the stage together to present MH002 – and the eight years of industrial co-development that sit behind it.

MH002 is a live consortium of six commensal strains being developed for mild-to-moderate ulcerative colitis. Phase 2a results in 45 patients (randomised, double-blind, placebo-controlled, 8 weeks): clinical remission of 14% vs. 7% on placebo at week 8, and 18% vs. 0% in the per-protocol population. CRP normalisation in patients elevated at baseline: 60% vs. 25%. Faecal calprotectin normalisation: 36% vs. 15%. No antibiotic preconditioning required.

What makes this story worth telling – from where I sit at Biose – is not only the clinical signal, but the manufacturing story behind it. Producing six anaerobic strains as a single drug substance in co-culture, maintaining viability, purity and stability as you scale, is genuinely hard. Adrien walked through eight years of progressive development: fermentation to 2,000 litres at late GMP, specific analytical methods to identify and quantify each of the six strains separately, and a innovative delivery methods to protect the product and target the intestine. Phase IIb and then Phase III are next, with scale projected toward 5,000 litres. The industrial partnership starts on day one – not when the clinical trial is already running.

The gut-brain axis gets serious

The second afternoon gave considerable space to the gut-brain axis, and the perspective that stayed with me most came from Amy Kruse, Chief Investment Officer at Satori Neuro. Her message was straightforward: the space is gaining credibility not because the idea is new, but because the evidence is becoming cleaner. Biomarkers, mechanisms, a credible regulatory path – when those three are aligned, the field starts looking investable. The era of broad microbiome enthusiasm is over; what investors now want is specificity.

Bloom Science‘s Christopher Reyes backed it up with data. Their oral LBP candidate BL-001 is designed to reproduce the effects of a ketogenic diet through the microbiome, with applications in Dravet syndrome, other developmental and epileptic encephalopathies, and obesity. Phase 1 results in 32 adults showed good tolerability and, in overweight subjects, a placebo-adjusted average weight reduction of around 2.3% at day 28, with dose-dependent effects on metabolic markers consistent with the ketogenic hypothesis. Early data – but with a mechanism behind it.

Valerie Taylor of Taylored Biotherapeutics added another angle: FMT trials in psychiatric disorders – bipolar disorder, OCD, depression – as a tool to identify which microbiomes are associated with treatment response, before moving toward a more defined microbial product. Around 80% response reported with FMT in bipolar disorder in their trial data. Eighteen candidate strains are now being evaluated for the next development stage. Still early. Still striking.

The wider picture

Some of the sessions that stayed with me fell outside the more familiar therapeutic tracks – and they are worth noting.
Yolanda Sanz from CSIC presented work linking a specific strain, Phascolarctobacterium faecium DSM 32890, to a favourable metabolic profile in human cohorts – less detected in obese subjects, with preclinical data showing reduced weight gain, adiposity and glucose intolerance in a high-fat diet mouse model. A small human study in overweight and obese subjects showed signals consistent with the preclinical data, though still very early stage. The next-generation probiotic space is moving toward strain-specific, mechanism-driven candidates – this felt like a good example of what that looks like in practice.

Kevin Horgan of Melius Microbiomics presented a different kind of next-generation approach: genetically engineered microorganisms, using E. coli Nissle 1917 as a chassis. Their platform, BioPersist, transfers the ttr operon from Salmonella Typhimurium to improve colonisation in an inflammatory environment. Lead programme MMB-003 is in development for ulcerative colitis. The engineering is layered – better colonisation, targeted colonic delivery, and specific metabolic pathway modulation. Whether you find GEMMs straightforward or complex as a regulatory proposition, the science is serious.

Stephen Robinson from the Quadram Institute introduced Bioscopic, a UK spin-out developing therapeutic candidates from infant microbiota bacteria – specifically Bifidobacterium-derived compounds, not live bacteria themselves. Their lead, BIOS-001, is in preclinical development for immuno-oncology, showing around 50% reduction in tumour burden in breast cancer mouse models via dendritic cell activation and CD8 T-cell engagement. The active molecule appears to be a surface sugar from Bifidobacterium pseudocatenulatum. Still preclinical, but a genuinely different angle on what microbiome-derived therapeutics can look like.

Peter Holme Jensen of Cambiotics presented something that felt more consumer health than pharma, but which raised questions worth taking seriously: gut bacteria capable of accumulating PFAS – the ‘forever chemicals’ present at measurable levels in most human populations. Their publicly reported strains show bioaccumulation of 25-74% of PFAS within minutes. A first human trial is planned for 2026. Whether or not this becomes a mainstream therapeutic direction, the underlying question – what else are commensal bacteria doing to environmental contaminants in our gut? – is a legitimate one.

Shelly Buffington from Baylor College of Medicine Ventures presented work tracking the infant microbiome across the first six months of life and linking specific microbial and metabolite profiles to neurodevelopmental outcomes at age two. In a cohort of 35 infants with 123 faecal samples and Bayley scores at two years, her team identified a predictive metabolite and then designed a microbial intervention to increase it – achieving more than a three-fold increase in the target metabolite and observing behavioural improvement in animal models. A reminder that the earliest developmental windows may be where microbiome intervention has the most to offer.
On the agritech side, Thomas Veje Flintegaard from Collaborome presented soil bacterial consortia being tested as biopesticides against fungal crop disease – with yields of up to 25% higher under fungal pressure in field trials. Not the core focus of most people in the room, but a useful reminder of where the broader microbiome toolkit is heading.

Two days in Copenhagen. Thirty-plus sessions. The common thread was not a single breakthrough, but a field steadily building the evidence base it needs to move from promising to proven – across therapeutics, prevention, nutrition, and now further afield than most people in the room had perhaps expected.

From our side at Biose, the manufacturing takeaway is the one we hear most consistently from our partners: the path from a promising strain to a stable, scalable drug substance is longer and more complex than it looks from the outside. The Biose-MRM Health story is one illustration of what that journey looks like when it goes well.

If you were at the Forum and want to compare notes, or if you are working on an LBP programme and want to discuss what the science means for your manufacturing strategy, feel free to reach out.

The 2026 Microbiome Times Partnering Forum Europe was held in Copenhagen. For more information: microbiomeforum.com