CC-Sim: Roadmap

Possible Future Updates

• Kinetic Cell Model
• CO2 Depletion in Gaseous Phase
• Lactic Acid production and consumption (cell model)
• Sheared / Lysed Cells (cell model) Done
• Perfusion / Chemostat Modules
• Sieving Done
• Better Scaleup Support
• Foam and Antifoam

Kinetic Cell Model

While CC-Sim comes with a macroscopic cell model that approximates how CHO cells would behave, it could be more accurate if the behavior were modeled from deeper mechanics. By incorporating key internal cell enzymatic reactions, a better model of how the cell would behave can be created.

Because Tensorflow has python support, and cell reactions can be calculated in a matrix format, there should be methods for efficient calculation of cell state using tensorflow.

CO2 Depletion Done

has nearly the same as oxygen; however, there is much lower relative mass transfer of between liquid and gaseous phases due to the high solubility of coupled with the low saturation of in the gas. While present correlations do not account for this, any simulation needs to accurately model depletion.

Presently, CC-Sim uses a lower value to approximate the correct mechanics.

Lactic Acid Production

Cells can use glucose for energy aerobically or anaerobically, with the latter producing lactic acid. However, since all reactions are really equilibria, the specificity of using the aerobic in anaerobic chain depends on concentrations of all reactants; low oxygen, high glucose, or low lactate concentrations will all tend to favor lactic acid production.

Modeling this correctly would require reaction kinetics, and would be included in the kinetic model. However, in the short term, lactic acid production can be included in the macroscopic model.

It is important to include lactic acid production to have an appropriate pH profile, as lactic acid is a primary influencer of pH (along with levels).

Shear / Cell Lysing Done

The current macroscopic model does not include cell shear or lysing (although the environment does output mean and maximum shear). Further research needs to be done to model this, although it probably isn’t important unless modeling perfusion (where higher shear rates are encountered).

Perfusion / Chemostat

The first version of CC-Sim operates purely in batch mode. Perfusion and Chemostat could be included once the mass balance equations have been solved, default controls and media have been set up, and a sieving correlation developed.

Sieving Done

Sieving has plagued perfusion culture for decades, and while many techniques have been utilized to reduce sieving, the exact mechanism is as yet poorly understood. Any accurate modeling of sieving would require the cause to be known, rather than crude correlations made.

Scaleup Support

As a general rule of thumb, every order of magnitude increase in scale yields new problems in process development. Scale yields different pressures (and solubilities), levels of shear, mixing profiles, equipment mechanisms, and even media preparation techniques.

CC-Sim uses correlations and equipment from primarily reusable benchtop-scale bioreactors. Different bioreactors (SUBs, waves) along with different correlations need to be included to accurately model different scales and types of bioreactors.

Foam and Antifoam

Bioreactors produce foam. Antifoam is used to battle this, at the cost of aeration mass transfer and (sometimes) effects on culture health. This is fairly low priority.