THE SCIENCE OF SCALE-UP

From batch log to optimized protocol in hours.

Upload your fermentation batch data. Fermvyne runs constraint-based flux balance simulations across your target scale points. You get a ready-to-use protocol for each vessel size — before you commit vessel time.

Run a Simulation The Science
STEP 01

Upload your strain & fermentation data.

Import a CSV from your batch historian or connect via API. Fermvyne accepts the standard bioprocess measurement set: OD600, dissolved oxygen (%), pH, glucose concentration, CO₂ evolution rate, and measured metabolite titers.

The system automatically maps column headers using common naming conventions from DeltaV, Siemens PCS 7, and standard CSV exports from bioreactor software (Sartorius MFCS, Eppendorf BioCommand). You can review and correct the mapping before running.

OD600 DO% pH Glucose g/L CO₂ evolution Titer g/L Feed rate L/h Agitation rpm
STEP 02

Flux balance simulation across scale points.

Fermvyne constructs a constraint-based metabolic model from your strain's observed phenotype — the exchange fluxes (uptake and secretion rates) inferred from your batch data. This becomes the stoichiometric foundation of the simulation.

The model is then solved under the mass-balance constraints of each target vessel: different kLa values, mixing times, and substrate gradient profiles. Outputs include a metabolic bottleneck map showing which pathways are flux-limited at each scale, a DO depletion risk score (0–100), and a predicted fed-batch titer curve.

FBA / constraint-based kLa estimation Bottleneck map DO risk score Titer curve
STEP 03

Protocol output, ready for the process binder.

Fermvyne generates a protocol card for each target vessel size in your simulation job. Each protocol contains: the feed schedule (feed rate vs time or vs DO setpoint), agitation ramp (rpm vs time), and DO setpoint cascade for cascade control loops.

Export as PDF for your process binder or as JSON for programmatic loading into your DCS/SCADA historian. The Team and Enterprise API allows automated protocol retrieval to feed directly into batch record systems.

VALIDATION

Internal accuracy benchmarks.

We report internal validation numbers honestly. These reflect agreement between Fermvyne's predictions and observed outcomes in our internal test dataset — not a certified third-party audit.

~88%
Predicted vs observed titer agreement within ±15% at pilot scale

Internal validation across 47 fed-batch runs in our development dataset (2L–500L, E. coli and P. pastoris). Results vary by organism, media complexity, and data quality of the input batch log. External validation with customer data is ongoing.

±8% Median DO% error at scale-up in internal test set
<2h Average simulation run time per scale-point batch
Scale-point comparisons per simulation job (Team tier)

Run a simulation on your strain data.

Request early access and we'll set up your first simulation — using your actual batch log, run against your target vessel size.