Protein Structure · MIT / Microsoft Research
AlphaFlow / BioEmu
Generates protein conformational ensembles directly from sequence for dynamic and disordered structures.
What AlphaFlow / BioEmu is
AlphaFlow and BioEmu are protein structure modelling methods developed by MIT and Microsoft Research. Rather than predicting a single static conformation, they treat protein structure as a dynamic distribution and sample multiple conformational states directly from sequence.
These methods are designed to characterise flexible biomolecules, making them applicable to intrinsically disordered regions, cryptic binding pockets, and allosteric mechanisms.
How it works
AlphaFlow integrates flow matching into the AlphaFold2 framework to sample structural distributions. BioEmu uses diffusion generative modelling to sample the conformational landscape directly from the target protein sequence.
Key facts
- Version
- 1.0
- Domain
- protein-structure
- Tasks
- conformational_ensemble
- Modalities
- protein_sequence
- GPU memory
- 40 GB
What you can use it for
- Intrinsically disordered region analysis
- Generates structural ensembles to represent the flexible conformations of intrinsically disordered proteins and regions.
- Cryptic pocket detection
- Samples diverse structural conformers to reveal transient or cryptic binding pockets across the ensemble.
- Allosteric mechanism investigation
- Models conformational heterogeneity to inspect alternate functional states and allosteric structural changes.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Not runnable yet
This model isn't available to run right now.Run this model
AlphaFlow / BioEmu
Strengths and limits
Good at
- Generates multi-state conformational ensembles rather than single static structures.
- Produces all-atom PDB structures alongside ensemble metrics such as RMSD, radius of gyration, and cluster occupancy.
- Samples conformational landscapes directly from primary amino acid sequences.
Watch out for
- Requires 40 GB of GPU memory to operate.
- Input is restricted to protein sequence and a target sample count.
Inputs and outputs
You provide
Protein sequence
Single-letter amino acid code. Whitespace and FASTA headers are stripped automatically.
Ensemble sizeoptional
Number of conformations to sample.
You get back
All-atom coordinates in PDB format, viewable and downloadable.
Per-conformer RMSD, radius of gyration and cluster occupancy.
Licence
Commercial use permitted · MIT
The MIT licence allows use in commercial research and products.
Questions
Can I use AlphaFlow / BioEmu commercially?
Commercial use permitted. The MIT licence allows use in commercial research and products.
How do I run AlphaFlow / BioEmu without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on 4–8× A100/H100, typically in 30min, and a run costs about 11,200 credits. Visitors without an account get 1,000 credits a month.
What inputs are required to run the model?
The model requires an amino acid sequence and a designated sample size specifying the number of conformers to generate.
What data formats are provided in the output?
The output includes viewable and downloadable all-atom coordinates in PDB format, as well as a table containing per-conformer RMSD, radius of gyration, and cluster occupancy.
What computational framework does AlphaFlow use compared to BioEmu?
AlphaFlow adds flow matching to AlphaFold2, whereas BioEmu applies diffusion modelling to sample conformational states directly from sequence.
What hardware is necessary to run this tool?
The system requires 40 GB of GPU memory.
Written from AlphaFlow / BioEmu's own documentation. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.