Which states belong in a CHO cell culture process model?

The model should represent viable and total cells, growth and death, product formation and degradation, glucose and glutamine, lactate and ammonium, dissolved oxygen and carbon dioxide, volume, osmolality, pH-related additions and the selected quality proxies, with sources and validity ranges for every parameter.

Can a CHO model predict manufacturing performance from literature data alone?

No. Literature values can establish a screening prior and plausible interval, but manufacturing prediction requires cell-line, medium, vessel, scale and operating-policy data plus independent validation against the intended context of use.

CHO cell culture process model

Connect CHO cell physiology to mAb scale-up, purification load and COGS.

Model viable-cell growth, death, productivity, substrate uptake, lactate and ammonium, oxygen transfer, feeds, osmolality, harvest quality, downstream cycles and facility capacity with explicit uncertainty and evidence status.

Review the 2,000 L benchmark See the full biopharma train

A CHO model is useful only when biological trajectories remain inside real vessel, gas, heat and downstream constraints.

InoculumCell growthFeed + gasesmAb productionMetabolitesHarvestDSP load

State model

Resolve the variables that drive productivity, viability and process robustness.

Use literature and vendor information to define screening priors, never to imply cell-line-specific predictive accuracy. Manufacturing use requires authorized development and plant evidence.

Cells and product

Track viable and total cell concentration, growth and death, integral viable-cell density, specific productivity, product formation and degradation, cell lysis and the quality proxies required by the context of use. Parameterize temperature shifts, feed phases and productivity changes independently where data supports them.

Substrates and metabolites

Represent glucose, glutamine or alternative nitrogen sources, amino-acid or aggregate nutrient groups, lactate, ammonium, dissolved oxygen and carbon dioxide, osmolality and pH-related additions. Add inhibition, overflow metabolism, lactate consumption or maintenance only when the data can identify those mechanisms.

Scale and facility coupling

Couple OUR to OTR, gas flow and oxygen enrichment; link feeds to working volume, mixing and local exposure; reconcile heat and cooling; then pass harvest titer, volume, solids and impurities into clarification, capture, viral safety, polishing, UF/DF, formulation, filling and finite-capacity scheduling.

Model layerMinimum inputsReview outputs

KineticsBatch-aligned cell, product, substrate, metabolite and event dataTrajectories, residuals, uncertainty and parameter identifiability

Mass transferGeometry, working volume, kLa basis, gas policy, OUR and broth propertiesOTR margin, gas demand, limiting period and extrapolation risk

Harvest + DSPVolume, titer, solids, impurity assays, step capacities and yieldsCycles, pool sizes, consumables, yield, queue and hold-time compliance

EconomicsMedia and feed costs, single-use demand, utilities, labor and equipment basisCOGS interval, annual output, dominant drivers and break-even targets

Do not report only final titerExport all time points, state variables, feeds, events, mass residuals, boundaries, uncertainty intervals and sensitivity runs.

Do not hide maturityLabel the model screening, calibrated, validated or not decision-ready for the named decision and operating envelope.