62

The diagram once again illustrates the system behaviour of these signalling cascades.

On the one hand, bioinformatic simulation enables us to understand how a cell with its

signalling cascades reacts in a healthy or pathological way to external signals. On the other

hand, it is also possible to test individual strategies in detail and determine which signals

are stronger or weaker and what then prevails in each case. Of course, one could also find

this out directly with many experiments, but this is much more time-consuming and also

requires many, many experiments.

The model description we have just used is “semi-quantitative”, i.e. we explain exactly

which is stronger or weaker, which signal is important first, second and last. But we are not

yet exactly quantitative, so that one already has exact quantities/concentrations. Of course,

there are such exact quantitative models in bioinformatics. However, the disadvantage of

such models is that they need much more additional information, in particular how fast the

most important processes change with time, or how strong the signals are at the beginning

and at least at four further time points. Then I can calculate with which function I describe

the change with time, i.e. I can set up the so-called differential equation of this property.

This makes sense, for example, if I want to thin the blood and I do not want to make the

platelets too weak or too strong for this purpose. That’s why we set up a fairly accurate

model with differential equations for this (and collected a lot of experimental data before­

hand). In many other cases, however, one does not have the time to measure everything so

precisely experimentally, and a semi-quantitative model is then already very good for

describing the corresponding system effects, for example when we want to protect plants

against pests or heat stress, to give a completely different example. To do this, we then

looked more closely at the effect of plant hormones (“cytokinins”), with which you would

spray the plant in the event of a bacterial infestation, for example, and then you have a

completely biological and readily degradable pest control agent (Naseem et al. 2012a). In

order to find the right cytokinin, the complex further effects were simulated more precisely

in a systems biology model, as shown above for the heart (cytokinins also control many

other processes, for example in the growth of plants).

In summary, then, systems biology descriptions are an important area of bioinformatics

today for better understanding systems behavior and signal processing in cells and organ­

isms. Often, relatively few data are sufficient for this purpose, because even a rather small

semi-quantitative model answers the questions about the best or most interesting system

effect, as in the case of heart failure, blood thinning, plant pests or, another exciting topic,

for example, cancer and cancer drugs (antibodies, cytostatics). These drugs need to be

optimally combined and correctly dosed – ideally even individually and patient-specific.

This is precisely where bioinformatics can calculate the best strategy for the patient.

There is often also more informatic preliminary work to the semiquantitative model.

This applies to biological systems and their system effects, which are hidden in large

amounts of data (e.g. gene expression data, genome sequences, metabolites, pharmaceuti­

cal levels, etc.) and where the decisive system components must first be filtered out using

statistics or even complex sequence analysis programs. This is also an important and

5  Systems Biology Helps to Discover Causes of Disease