Why Aster?
Building the data infrastructure for the post-animal era of biology.
For nearly two decades, biofabrication has been treated as artisanal craft when it has always been a manufacturing problem. Researchers have produced extraordinary individual results that don't add up to a discipline. Print parameters live in lab notebooks. QC is done by eye. Datasets are fragmented and small. The connection between fabrication and outcome is rebuilt from scratch in every project.
Every transformative biotech category has been unlocked by data infrastructure. Genomics by GenBank. Structural biology by the Protein Data Bank. Drug discovery by ChEMBL. The pattern is consistent. Fragmented work produces breakthroughs that don't compound, until shared infrastructure turns isolated experiments into a connected scientific record.
Biofabrication doesn't have its equivalent yet. Aster Biofabrication is building it.
What biofabrication is, and why it's stuck
Biofabrication is the production of living biological constructs through controlled processes. It includes 3D bioprinting of cell-laden hydrogels, organoids and spheroids, organ-on-chip systems, and tissue-engineered scaffolds seeded with living cells. By any reasonable definition, this is manufacturing. The field has not yet treated it that way.
Biofabrication exists because human hands have reached their limits. Building vascularized constructs, layered epithelia, and the complex architectures that genuinely model human tissue requires precision and reproducibility that manual technique cannot provide. Biofabrication is, in this sense, a cousin to high-throughput screening: both bring controlled, automated processes to where skilled hands run out of room. HTS does this for compound testing. Biofabrication does it for tissue construction.
The field grew out of tissue engineering, which Howard Green and Robert Langer established as a discipline in the late 1970s and early 1980s. After nearly half a century, tissue engineering has produced a handful of FDA-approved products against hundreds of billions of dollars in research investment. The bottleneck has not been a lack of tools to understand biology. It has been the inability to produce tissue models that are reproducible and multifaceted enough to genuinely study human physiology. That is a manufacturing problem before it is a biological one.
Biofabrication has a manufacturing problem
This section's title is more literal than it sounds. Biofabrication is a manufacturing discipline. The field has been allowed to operate as if it were not, and the consequences are concrete. Producing one construct is a science exercise. Producing a thousand consistent constructs on a documented schedule is a manufacturing exercise. The field today is good at the first and bad at the second.
Three failures compound. Throughput: a pharma toxicology campaign needs thousands of constructs on a predictable timeline, but biofabrication labs produce them in tens to hundreds, mostly by hand (minus the biofabricator of course), with each batch optimized fresh. Quality control: when the output is large, manual scoring under a microscope becomes the bottleneck. Provenance: manufacturing demands every output be traceable to the parameters that produced it, but biofabrication today rarely captures those parameters at all.
These failures share a root cause. Researchers stitch together generic slicers built for plastic 3D printing, electronic notebooks meant to document experiments rather than capture biofabrication-specific parameters, visual checks of printed constructs without the analytical tools to determine when and why something went wrong, and materials management that does not track the bioinks themselves. How old is the alginate in the cabinet? Which lot? Stored how? Mixed with what? The data never connects.
Why now?
Three forces are converging. First, regulatory tailwind. The FDA Modernization Act 2.0, signed in late 2022, removed the federal mandate that drugs be tested in animals before human trials. The FDA's 2025 roadmap to reduce animal testing in monoclonal antibody development is the first concrete implementation. The European Union has prioritized New Approach Methodologies for over a decade through EURL ECVAM and Directive 2010/63/EU, and U.S. policy is now catching up. The pressure on drug developers to validate non-animal models has shifted from theoretical to operational.
There is a useful historical parallel. High-throughput screening emerged in the late 1980s as the response to the 1962 Kefauver-Harris Amendment, which required drugs to prove efficacy rather than just safety. By 1989 Pfizer had industrialized compound screening, and the infrastructure built for that transition reshaped pharmaceutical R&D for the next thirty years. HTS was the moment chemistry-based drug discovery accepted that it was, at its core, a manufacturing activity. New Approach Methodologies are now driving the same demand on biofabrication. The infrastructure to support it doesn't exist yet, which is precisely the opportunity.
Second, biofabrication has matured past the hardware bottleneck. Commercial bioprinters are commoditizing, with multiple vendors offering comparable specs at falling prices. Stable cell lines and organoid protocols are widely available. Bioinks remain a heterogeneous landscape where commercial products serve as starting points for further modification, and those modifications are rarely captured. The hardware works. The integration is what doesn't.
Third, the historical pattern is accelerating. The Protein Data Bank, founded in 1971, waited nearly fifty years for compute to catch up before AlphaFold made it foundational. GenBank, founded in 1982, waited about thirty years for next-generation sequencing and cloud compute to make modern genomics tractable. ChEMBL, founded in 2009, waited closer to ten years before machine learning could exploit it. The arc from infrastructure-built to discipline-productive is compressing because the substrate is finally ready. The bottleneck has stopped being compute. The bottleneck is whether anyone bothers to build the data layer.
What we are building
Capitula is the platform we're building to solve this. Five integrated modules cover the experimental lifecycle from planning to prediction. The connective tissue between them is the entire point. A researcher's bench workflow becomes a production-line process without changing what the researcher does. The infrastructure does the translation.
Why Aster (the name)
The name is deliberate. Asteraceae, the aster family, is one of the most diverse and ecologically successful plant families on Earth. Daisies, sunflowers, marigolds, calendula, cosmos, goldenrod, asters: thousands of species, each adapted to a different niche, together forming the backbone of pollinator ecosystems across nearly every climate. No single species does the work alone. The strength of the family is in its diversity.
The same is true for the in vitro models that will replace animal testing. There is no single platform that solves everything. Organoids, organ-on-chip systems, bioprinted constructs, spheroids, scaffolds, microphysiological systems: each plays a role in a vibrant ecosystem of complementary approaches. The future of biology after animal models is not a winner-takes-all platform. It is a diverse, interconnected set of methods supported by infrastructure that makes them all stronger.
Aster's modules borrow their names from the family. Calendula, Cosmos, Helianthus, and Solidago are all members of Asteraceae. Not as decoration, but as a reminder of what the platform is for. The aster family supports its pollinators because it supports diversity. We are building Capitula to do the same for the in vitro ecosystem.
What this requires
Infrastructure built specifically for biofabrication and not adapted from traditional 3D printing.
Standardization at the data layer, so work in one lab can be read, reproduced, and built on by another.
Things that should already exist (but don’t)
Each of these requires structured data at capture time.
Each one is what we are building Capitula to enable.
A platform that recommends the right combination of in vitro models for each question, with organoids for emergent behavior, organ-on-chip for tissue mechanics, and biofabricated constructs for more complex questions, and runs them at scale with documented variation.
A bioprinting workflow that gets to publication-quality reproducibility on the first attempt rather than the fifth.
A working protocol that another lab can replicate without re-optimizing it, because every re-optimization quietly changes the model being studied.
A predictive model that flags which print parameters will fail before a week of cell culture is wasted finding out.
A regulatory dossier where every claim traces back to versioned source data and a captured pipeline.
Infrastructure that closes the loop from experiment to improvement, where every run automatically informs the next instead of contributing to a fragmented archive that no one can learn from.