01 / The possibility
Biology is becoming programmable.
Explore the models, companies and technologies changing what we can build with biology.
02 / Understand
Big breakthroughs. Made understandable.
Start with the idea, not the jargon. DigitalBio connects the science to the questions that matter: what changed, what it enables, and where the limits are.
A concept, unpacked
What is protein structure prediction?
A protein is a chain of amino acids. Its three-dimensional shape helps determine how it works. Structure prediction estimates that shape from the sequence.
A predicted shape is a hypothesis. It does not, by itself, prove that a protein works or that a drug will be effective.
Explore a real example: AlphaFold3 ↗Source: Abramson et al., Nature (2024) ↗03 / Discover
Find the technology behind the breakthrough.
NextIn Bio turns a biological task into a starting point. Explore relevant Model Book records, then follow the evidence. A curated selection, never a claim to cover everything.
Explore the Model Book →NextIn Bio / task matches
Predict protein structure
6 matching records · shown in catalogue order
Structure Prediction · Complex Prediction · Ligand Docking
Consider: Production drug discovery pipelines requiring IP freedomStructure Prediction · Complex Prediction
Consider: When absolute peak accuracy on complexes matters more than accessibilityStructure Prediction · Molecular Docking · Complex Prediction
Consider: Pure protein structure prediction — Boltz-1 is lighter and nearly as goodSee all task matches →04 / Compare
What’s possible is different from what’s practical.
A promising capability is only the beginning. Put the evidence, license and hardware requirements side by side before choosing what to evaluate.
Compare these models →NextIn Bio / decision workspace
| Decision lens | AlphaFold3 ↗Google DeepMind | Boltz-1 ↗MIT |
|---|---|---|
| Capabilities | Structure Prediction · Complex Prediction · Ligand Docking | Structure Prediction · Complex Prediction |
| Evidence | CASP15: 0.85 GDT-TS (avg) | CASP15: 0.82 GDT-TS (avg) |
| Licensing | restricted | open |
| Deployment | API-only — no local deployment possible | 1x A100 40GB for most predictions |
| Tradeoffs | Production drug discovery pipelines requiring IP freedom | When absolute peak accuracy on complexes matters more than accessibility |
| Sources & review | AlphaFold3 Paper (Nature 2024) ↗AlphaFold Server ↗CASP15 Results ↗Record reviewed | Boltz-1 GitHub ↗Record reviewed |
06 / Build
Turn what you discover into something real.
Cells & Compute is where scientists, engineers and operators come together around real problems. Bring a question. Find a collaborator. Start building.
Find your collaborators ↗Cells & Compute
Different disciplines.
Shared curiosity.
Science meets engineering meets the work of making it happen.
Meet Cells & Compute ↗Explore the DigitalBio Lab →Keep your curiosity close
The next idea starts here.
Accessible science, useful context, and a little room to think.