66

Biological Examples for Boolean Modeling

Basics:

In a review paper, we systematically compared different approaches to Boolean modeling and

dynamic modeling, e.g. SQUAD, ODEFY, and CellNetAnalyzer (Schlatter et al. 2012). Another

good starting publication is Di Cara et al. (2007) on SQUAD. Our software Jimena is a nice further

development (Karl and Dandekar 2013a). Jimena also offers to distinguish between direct and

dynamic network control quantitatively and qualitatively in networks (Karl and Dandekar 2015).

Specific models for different cells and processes:

• heart: Brietz et al. (2016a) and Breitenbach et al. (2019a, b),

• liver: Philippi et al. (2009),

• immune cells: Czakai et al. (2016),

• tumours: Stratmann et al. (2014), Göttlich et al. (2016a), Baur et al. (2020), and Kunz

et al. (2020),

• plants (hormones and infections): Naseem et  al. (2012, 2013a, b), and Kunz

et al. (2017),

• bacteria: Audretsch et al. (2013),

• platelets: Mischnik et al. (2013a, b).

Extension of such semi-quantitative models to fully dynamic models:

Two papers on dynamic modeling via platelets are helpful here for comparison:

Mischnik et al. (2014) describe the function of the signal molecule Src, but now with differential

equations and estimates of the velocities of all processes (“kinetic parameters”). It is crucial to

switch between active and inactive platelets. In the process, the mathematical description was also

verified in detail experimentally.

Wangorsch et al. (2011) again describe the function of inhibitory cyclic nucleotides in the platelet

using differential equations that take into account the different rates of the processes involved and the

absolute signal strength. In particular, I can cause the platelet to become inactive by increasing the level

of cAMP. This can be used medically, for example, to prevent a new blood clot in the case of strokes.

The behaviour for different active substances and their combination is described in detail in the paper.

In both works, this was used to accurately estimate the kinetic parameters through experimental

data and then develop corresponding optimal fitting differential equations (ODEs). One can also calcu­

late in general what the optimal pharmacological intervention should be (Breitenbach et al. 2019a, b).

In addition to this selection of one’s own work on the topic, there are of course also large reposi­

tories of models, so that one can compare models from many authors or search for the optimal one

for a question, which one can then possibly adapt to one’s own question, for example:

https://systems-­biology.org/resources/model-­repositories/ (from the journal “Systems Biology

and Applications”).

celldesigner.org/models.html (from the software CellDesigner, very nicely linked to the Panther

Pathway database).

https://www.ebi.ac.uk/biomodels/ (The “Biomodels Database” of the EBI, with many mathemat­

ical, pharmacological and physiological dynamic models collected from the literature).

The examples above show that semiquantitative models can be used to cover the entire

range of systems biology regulation and biological signalling networks. The particular

5  Systems Biology Helps to Discover Causes of Disease