Pick the strain that survives scale-up, not just the bench.
Metabolic engineering produces multiple candidate strains. Fermvyne tells you which phenotypes have the flux distribution to sustain performance at scale — before you invest pilot resources in candidates that will overflow or become oxygen-limited under commercial conditions.
Request Early AccessUp to 5 strain variants per simulation job.
Upload bench run data from each strain candidate. Fermvyne infers the flux phenotype for each and simulates how each phenotype will perform at your target scale under standardized conditions — making strain comparison apples-to-apples rather than confounded by different bench run conditions.
What Fermvyne compares across strains
- Overflow onset at target scale density — which strains accumulate acetate or other overflow metabolites at high-OD fed-batch conditions
- Oxygen sensitivity — minimum kLa each strain needs to maintain aerobic growth at peak OUR demand
- Optimal induction OD (for recombinant strains) — the cell density where metabolic capacity for product synthesis is highest relative to growth maintenance cost
- Predicted volumetric productivity at target scale — not just bench-derived specific productivity extrapolated linearly
- Sensitivity to feed rate perturbations — which strains are robust to DCS control variability and which are brittle
Already have a genome-scale model?
If your team has built or adapted a genome-scale metabolic model (GEM) for your organism — in COBRApy, MATLAB's COBRA Toolbox, or exported as SBML — you can import it directly into Fermvyne.
Fermvyne overlays your model's predicted intracellular flux constraints onto its kLa and fed-batch feed optimization engine. Your model provides the metabolic phenotype; Fermvyne provides the scale-up physics. The combination produces more accurate scale-up predictions than either approach alone.
Ask About GSM ImportSupported formats
- SBML Level 3 (COBRApy / COBRA Toolbox export)
- JSON (COBRApy native)
- Exchange flux CSV (manually curated reaction bounds)
- Measured exchange flux CSV (from your experimental data)
Fermvyne's built-in FBA engine is used by default if no external model is imported. External models provide higher fidelity for organisms with well-characterized GEMs (e.g., iML1515 for E. coli, iMT1026 for P. pastoris).
Rank your strain candidates before the pilot.
Upload bench run data from up to 5 strain variants. Get a comparative scale-up performance prediction within two hours. Select your pilot candidate with data, not guesswork.