Mycelium-Bound Roads: Reshape Biotech's Automated Imaging Accelerates Biotech Innovation
- Marc Violo
- 13 hours ago
- 5 min read
Article in partnership with Reshape Biotech, a Danish company that develops advanced lab automation robots and a cloud-based AI platform to accelerate biological research and development.
Most readers here will know mycelium as a capable, adaptable builder of structure. In forest soils it colonises particulate matter, threads itself into gaps, and stabilises the whole. Visibuilt’s work takes that familiar behaviour and applies it to a long‑standing industrial substrate: mineral aggregates in asphalt.
The motivation is straightforward. Conventional bitumen is fossil‑derived and typically processed and paved at temperatures well above 150 °C. Those conditions are energy and emissions intensive, and a replacement that can be produced and applied without high heat would be a material step forward, provided it meets mechanical and durability requirements on the road.
Growing A Natural Binder
Visibuilt’s binder, Visibit, is grown by fermentation, uses agricultural by‑products as primary carbon source with naturally occurring (non‑GMO) fungal strains. In the lab, mycelium is encouraged to colonise through and around aggregate‑like substrates so that the resulting composite behaves as a stone‑rich, mycelium‑bound material.
Early prototypes did what early prototypes commonly do: some held, some failed, and then the programme iterated. Strength targets were reached first; attention has shifted to long‑term performance under moisture, freeze–thaw cycling and traffic loads, the questions any pavement engineer or materials scientist will ask before specifying a new binder.
Oleksii Rybrov, Head of Science at Visibuilt, as shared that field‑scale pilots have already been conducted using industrial asphalt equipment, notably without the heating step that conventional mixes require. In practice that means the same plant, trucks and pavers can be used while omitting the burner; a test section has carried traffic for months without visible cracking during initial observation windows, and outdoor durability studies are ongoing.
From Manual Observation to Continuous Data
For fungal scientists, another interesting part of this story may be how the team is running its experiments. Screening a fungus–substrate–process space this large by hand is slow and, more importantly, subjective. Plate checks “by eye” make it difficult to compare assays across weeks, let alone relate early growth patterns to later mechanical outcomes. Visibuilt addressed this by instrumenting the biology. Automated time‑lapse imaging and analysis (via Reshape’s systems) turned plates from occasional snapshots into continuous data streams.
Colonisation can be quantified as percentage surface coverage and intensity over time; growth curves can be compared between strains, substrates and inoculation strategies; and replicates become the default rather than the exception. Caroline Thyssen Petersen, a research scientist who has been with the company since its inception, noted that moving from ruler‑on‑a‑plate estimates to automated imaging made it feasible to screen “a lot of different strains” across “a lot of different substrates” with proper replication and consistent readouts, and that having all assays stored as organised time‑lapses made re‑analysis practical as the programme evolved.
Solving the Visual Problem
The visual problem Visibuilt faces is not a textbook agar plate. Substrates are often dark, rocky and heterogeneous; the fungi are frequently white; and the desired outcome is uniform colonisation across an irregular surface. Off‑the‑shelf analysis models do not always cope with that contrast.
In joint sessions with Reshape’s applications team, analysis pipelines were adjusted for this domain: a “microbiology” detector that unifies small‑ and large‑colony detection so scientists needn’t guess which model to select at set‑up; colour‑based segmentation to separate mycelium from substrate on dark backgrounds; and downstream, local (offline) HTML tools for visualising and post‑processing results so that sensitive data never leaves the lab’s network.
As goals changed, switching organisms, testing different morphologies, or redefining what “success” on a plate means, models were retrained or even applied retroactively to existing time‑lapses, extracting new measurements from previously run experiments. That ability to reuse data matters when the biology is moving quickly and sample throughput is precious.
Not every question can be answered at the surface. During technical reviews the Visibuilt team asked about seeing below the top layer - whether changes in density or internal hyphal architecture could be tracked inside a small composite coupon.
Today’s imaging is two‑dimensional and surface‑based; alternative modalities would be needed to probe the interior, and the teams treated that as roadmap input rather than a current feature.
There were also practical discussions about sample height and optics. Some taller prototypes sat too close to the camera and introduced a fish‑eye effect; the solution was to keep assays within the instrument’s designed focus envelope by adjusting plate formats. These constraints are mundane but important, because consistent image geometry is what makes longitudinal analysis reliable.
Building a Disciplined Workflow
What emerges from this lab practice is a disciplined, repeatable workflow. Inoculation strategies vary, from central points to distributed “all‑over” seeding, depending on the question being asked. Colonisation is quantified by surface coverage and whiteness intensity; uniformity can be scored; and time‑series behaviour can be mapped to later mechanical testing on stone‑bound coupons.
Because every image is retained, new metrics can be applied as the team learns what correlates with strength and durability in the end product. It is a familiar pattern to anyone running applied mycology: turn qualitative observations into quantitative traces, then use those traces to decide which candidates deserve scale‑up.
The Path to Pilot Scale
Scale‑up itself has followed a deliberate path. After bench-top assays and small composite tests, a ~300 kg pilot batch was processed at an asphalt factory using standard equipment, without heating, and laid on a test section. The production line did not require modification; the trial behaved like a conventional job with the single exception that the burner stayed off.
Those results, combined with the strength achieved in the lab, now motivate the longer outdoor study programme focused on water, temperature cycling and fatigue. In parallel, the fermentation process is being standardised so that the binder’s properties are consistent between runs - a necessary step for any material that hopes to meet specification.
Starting from the Base Layer
It is worth noting the scope of the first applications. Visibuilt’s immediate focus is the base layer within the pavement structure, rather than the top course. The base is structurally important but faces fewer regulatory and performance constraints than the wearing surface, which must also satisfy texture, skid resistance, marking adhesion and other top‑layer requirements. Working from the base up allows the team to accumulate the durability and statistical evidence that standards bodies and asset owners will eventually require for wider adoption.
There is ongoing research about whether the binder should remain biologically active in the final composite, “self‑healing” is a tempting phrase in our field, or be inactivated for control and predictability. Each path has advantages; Visibuilt has treated this as an empirical question tied to performance and process compatibility rather than a point of doctrine.
Method Over Marketing
For fungal R&D specialists, the take‑home is less about slogans and more about method. A fermentation‑made binder can, in principle, replace a significant source of heat and fossil input in road building. Demonstrating that requires the usual fungal R&D levers - strain selection, substrate formulation, moisture and aeration control - combined with quantitative, instrumented screening and a careful bridge to field conditions.
Visibiult’s programme shows how those pieces can be assembled: iterate on plates you can measure, build small composites you can break, and only then run factory‑scale material through the equipment that will eventually have to use it. The fact that this workflow is now producing road‑worthy sections is encouraging, but the most valuable output for the community may be the map from Petri dish to pavement, and the role that well‑designed imaging and analysis plays in shortening that path.
Reshape Biotech is a Danish company that develops advanced lab automation robots and a cloud-based AI platform to accelerate biological research and development. Their technology automates repetitive tasks in labs, such as microbial growth monitoring, to increase experiment throughput, enable global collaboration, and provide deeper insights through AI-powered data analysis.
Learn more on: www.reshapebiotech.com
