Digital twin for biotech research teams
Choose better experiments before you enter the lab.
Bring a biological question and the limited data you already have. The ALTDATA Digital Twin compares conditions in software, shows where the result is uncertain and helps you narrow the list worth testing for real.
What you bring
A research question and the evidence you already have.
A biotech team rarely runs out of ideas. It runs out of time, samples and budget to test every condition. The result is a small dataset and a large decision resting on it.
The wider problem is measurable. 90% of drugs that clear preclinical testing still fail in the clinic, and the correlation between in vivo results and trial outcomes runs under 8% (van Rijt et al., Cancers, 2023).
We don’t claim to move those numbers. We work earlier in the process, where teams are deciding which conditions deserve scarce experimental resources.
To start, we need two things:
- The decision. The conditions, compounds or settings you need to compare.
- The evidence. The biological measurements you already trust, even when the dataset is limited.
You do not need a perfect dataset to begin. You need a clear question and an honest account of the evidence.
What you get
From a broad question to a smaller lab plan.
We define the conditions to compare, use the Digital Twin to run them virtually and return the predicted outcomes with an uncertainty signal attached. The result is not a replacement for the lab. It is a clearer view of which experiments still deserve one.
The engine combines a physics simulation with a fast AI model. Physics provides the grounding; AI makes it practical to explore many conditions. Real biological measurements keep the system tied to evidence rather than patterns alone.
Scope today: the engine has been validated for 3-D biological growth and structure over time, on one cell line under 1G conditions. Additional biological models and microgravity validation remain development work.
More conditions explored
Compare possibilities in software before using samples, equipment time or lab budget.
Uncertainty attached
See which results are dependable and which ones still need a real-world check.
A more focused lab plan
Take a smaller, better-supported set of conditions into physical testing.
A demanding application
An Australian microgravity digital twin project.
Microgravity research leaves little room for trial and error. Flight missions are rare, expensive and high-risk, so a team may have one opportunity to ask the right question.
To put a number on “rare”: across all of 2024 the ISS National Lab supported 110 payloads across seven missions. Seven windows in a year, for every discipline competing for them. If your experiment misses one, you wait.
The Gravity Knob lets a research team explore how changes in gravity’s settling force may affect the structure of a 3-D culture. Its purpose is to test more possibilities on the ground and arrive at a flight experiment with a sharper plan.
What has and has not been validated: the settling mechanism has been checked against real 1G suspension-culture data. It has not yet been validated against matched data from true microgravity. ALTDATA is the digital twin partner in a ResearchSat-led Australian microgravity project, funded through Round 3 of the SA Space Collaboration and Innovation Fund. Its planned ISS mission is intended to gather that evidence. Until then, the Gravity Knob is a planning tool, not a substitute for flight data.
Validation
What has been checked, and against what.
The current engine has been tested in two different ways. One checks whether the fast AI model reproduces the physics simulation. The other checks whether a model trained on its synthetic output still performs against real biological data.
- Checked against reality. Real biological measurements calibrate the system, so it is not judged against simulated data alone.
- Two published benchmarks. The fast model reproduces the physics simulation at ~88% fidelity. On a train-on-synthetic, test-on-real benchmark, it reaches 96% of the accuracy of a model trained directly on the real growth data.
- Physical limits built in. Rules constrain results to the biological limits represented by the model.
- Uncertainty stays visible. Each result shows where a real check is still needed. True microgravity is the clearest current gap.
- Local use is possible. The model is lean enough to run on your own hardware, so sensitive inputs do not need to leave your organisation.
Where it fits today, and where it may go
The validated case today is 3-D biological growth over time for one cell line under 1G conditions. Everything else below is development work or a direction the method is designed to support.
Validated today
Explore 3-D biological growth and structure over time for the current validated cell line under 1G conditions.
Additional cancer models
Extend the engine to further cell lines and cancer models, with each one requiring its own validation.
Microgravity research
Use the 1G-validated settling mechanism to plan future flight research while matched microgravity validation remains ahead.
Broader research use
Apply the same method to bioprocess, cell-culture and condition-screening questions once the relevant biological model has been validated.
Frequently asked questions
What do we need to start?
A clear research question, the conditions you want to compare and the biological measurements you already trust. The point is to work carefully with limited evidence, not wait for a perfect dataset.
What does the Digital Twin give us?
A set of virtual experiment results, an uncertainty signal for each result and a smaller group of conditions worth checking in the lab. The exact output depends on the research question and the current validated scope.
Can a virtual experiment replace wet-lab work?
No, and it isn't meant to. It helps move early exploration into software, then directs limited lab time toward the conditions that still need a real check. The real experiment remains the final test.
What is the Gravity Knob?
It lets researchers explore how changes in gravity's settling force may affect a 3-D culture. The mechanism has been checked against real 1G data, not true microgravity data, so it is a planning tool for future flight research rather than a replacement for a flight experiment.
How do you know the twin is right?
We check the fast AI model against the physics simulation, then check the system against real biological measurements. Each result also carries an uncertainty signal, so the model shows where a laboratory check is still needed.
Is our data safe? Can it run on our own hardware?
The current model is lean enough to run locally on your own hardware, so sensitive inputs do not need to leave your organisation.
Which research areas is it built for, and how do we start?
The validated case today covers 3-D biological growth over time for one cell line under 1G conditions. Additional biological models and microgravity validation are development work. Start by bringing us the decision you need to make and the evidence you already have. Discuss your research problem →
The bottom line
Use the lab to confirm, not to search blindly.
The ALTDATA Digital Twin helps biotech teams explore more possibilities in software and reserve physical experiments for the conditions that still need real evidence.
The real experiment remains the final test. The twin helps you choose a better one.
Bring us the experiment you're deciding whether to run.
We'll start with the question, the data you have and the current scope of the engine. If it fits, we'll show you what a useful first piece of work looks like.