Here a brief guide to the many examples of cultures and cultivation techniques. The guide is mainly organized around the culture model used. There is about one Jupyter notebook around an example per repository. Reference to the literature you find in the notebooks. I hope the variety of examples give an idea of of the versatility of using compiled Modelica code as FMU and Jupyter notebooks with small Python scripts to improve understanding of bioprocesses.
In some notebooks multiple simulations are done for sensitivity analysis, model calibration, or process optimization. Here are used extra Python packages SALib and SciPy. The key in these type of notebooks is that an application dependent evaluation function is defined in Python and a general flexible and open format is used and facilitated with the FMU-explore command line package. The optimization technique used is gradient free.
There are Python techniques to parallelise computations using multiple CPU-kernels but not shown here. In the near future the updated FMI-standard 3.0 for generation of FMU provides gradient information and can be used to make optimization more effective for a class of problems
We start with the text book model of a culture with only states for: substrate, biomass, and reactor volume. This model is used in many text books and even used for mammalian cultures to explain the main ideas around continuous and perfusion cultivation.
- BPL_TEST2_Batch - Batch cultivation
- BPL_TEST2_Fedbatch - Fedbatch cultivation
- BPL_TEST2_Chemostat - Continuous cultivation
- BPL_TEST2_Perfusion - Perfusion cultivation
Model calibration is important in practice and a basic example using Python scipy optimization routine is given below. A similar structure of the notebook is found in other "calibration-notebooks" on this site.
- BPL_TEST2_Batch_calibration - Batch cultivation calibration
Design space calculation is an important kind of application and a basic example is given below. Here noise is added to the measurement signal in the later part of the notebook.
- BPL_TEST2_Batch_design_space - Batch cultivation and design space characterization
The next step in culture model complexity is to include byproduct formation. This is important for S cerevisiase and E coli and also for CHO-cultures. We start with a yeast model and it can be run both as as batch and fedbatch here. The model also includes aeration, e.g. both liquid and gasphase included. The model was originally developed for continuous cultivation.
- BPL_YEAST_AIR_Fedbatch - Fedbatch cultivation of yeast in an aerated reactor with DO-control
The yeast model can be structured in two parts and here the details around the respiratory bottleneck is separated and different modifications of the bottlenecks are investigated. This is studied in continuous cultivation and the phenomenon of multiple steady states is elucidated with these model modifications.
- BPL_YEAST_O2LIM_Chemostat - Continuous cultivation - not yet public
The yeast respiratory bottleneck model can be described in terms of constraint-based modelling. This is illustrated in the next example. The notebook include connection to a public genome based model.
- BPL_YEAST_COB - Yeast and constraint-based modelling
Yeast is industrially culitivated in large reactors where gradients in substrate concentration play a role for lowering yield. Reactor size typically 100
- BPL_YEAST_Fedbatch_hotspot - Yeast and fedbatch two-reactor setup - not yet public
The E coli model below is similar to the the yeast model but also include maintenance metabolism. Although this bacteria is in many ways simpler than yeast the metabolism and growth can be more complex in reality. The model was originally developed for fedbatch cultivation.
- BPL_ECOLI_AIR_Fedbatch - Fedbatch cultivation of E coli in an aerated reactor - not yet public
The CHO-model below is also inspired by the bottleneck model of yeast above. In this model we describe both byproducts lactate and ammonia and the model has two different bottlenecks. The maintenance metabolism is also included as well as certain inhibition effects. The model was originally developed for fedbatch cultivation with continuous feed.
- BPL_CHO_Fedbatch - Fedbatch cultivation of CHO-culture
- BPL_CHO_Fedbatch_optimization - Optimization of fedbatch cultivation of CHO-culture - not yet public
- BPL_CHO_Fedbatch_bolus - Bolus feeding of fedbatch cultivation - not yet public
The same CHO-model can also be used to describe continuous perfusion cultivation.
- BPL_CHO_Perfusion - Perfusion cultivation of CHO-cells - not yet public
- BPL_CHO_Perfusion_cspr - Perfusion cultivation of CHO-cells illustrating the CSPR concept
Ion exchange chromatograpy (IEC) is an important process unit to separate the product of interest from similar molecules. Here a simplified model is used to illustrate the main principles of operation.
- BPL_IEC_validation - Ion Exchange Chromatograpy introduction
- BPL_IEC_operation - Ion Exchange Chromatograpy operation
- BPL_IEC_uv_pooling - Ion Exchange Chromatograpy using UV-pooling - not yet public
The Bioprocess Library is well suited to study control and operation of a bioreactor and we also include a filter and an example of downstream chromatograhy. It is also possible to simulate a small sequence of unit operations like culture expansion and culture harvest. Such operation is here modelled mainly by time schedules of signals for valves and pumps and simpler control logic. There is a possibility for more comprehensive sequence control in Modelica and may be explored in the future.