88

cascade (see Chap. 5), all other signalling pathways are also active. The cell works with

biochemical reactions and not like a digital silicon computer. Therefore, signals can only

reach their destination if they are amplified in a cascade. Nice examples are the blood

clotting cascade, so that the broken vessel is guaranteed to be closed again safely and

quickly, and also the opposite blood clot dissolving cascade (plasminogen cascade). In the

blood, for example, there is then also the complement cascade for the immune system and

so on. So in general, biology has to come up with a lot of things to cope with the noise.

One possibility to reach highest sensitivity is given for example by the photoreceptors of

our eye, where three inhibitory mechanisms all together return to the resting state and the

initial situation is a hyperpolarization.

A computer or even you yourself with the next transfer with IBAN number use check

bits to be sure that nothing has been changed by mistake. This mechanism also exists. First

of all, all kinds of sequence signals are used for this purpose, which you can find out with

the ELM server, for example, and which ensure in a relatively error-tolerant way that every

protein gets to the right place. However, the stability signals and signals that ensure that a

“wrong” protein, for example one that is too short, is rapidly degraded (so-called “non­

sense mediated decay”, NMD, for stopping too early in the case of mRNA from eukary­

otes) are also a kind of check bit for proteins. Similar check bits exist for RNA, such as

various methylguanosine caps that mark different types of RNA as mature and regulate the

nuclear or cytoplasmic transport of that RNA and its proteins. Another strategy to better

understand the notoriously complex codes in biological systems is simplification (techni­

cal term: dimensionality reduction). The aim is to transform and visualise high-­dimensional

data in a new coordinate system (usually 2D). For this purpose, methods of multivariate

statistics such as PCA (Principal Component Analysis; for examples in R see our web

application [Fuchs et al. 2020] or https://rpubs.com/amos593/419546) are applied (explor­

ative data analysis). Through dimensionality reduction, one wants to get an overview of

the data and reduce its complexity by decomposing it into principal components. Through

this structuring one wants to extract relevant variables (features) and groups, for example

for the construction of predictive models (Chap. 14), but also to make visible possible

batch effects in the data that may need to be corrected (especially in omics analyses). For

example, the pattern of gene expression is determined by the interaction of many 1000

genes. To get an overview of the most important components involved, PCA can be used

to calculate the two main components of the differences between datasets, giving a quick

overview of which combination of important genes decisively determines the differences.

The method is applicable to all complex datasets, e.g. cardiac fibrosis (Fuchs et al. 2020),

but also in ecology, for example to quickly screen bacterial communities (Kim et al. 2020).

One can also look at the challenges of reliable signal transmission and coding in the cell

in a mathematically exact way for signal cascades and the phosphatases that switch off the

signal and thus better understand how these cellular signals are formed and transmitted

(Heinrich et al. 2002). Phosphatases are important for the regulation of signal amplitude,

7  How to Better Understand Signal Cascades and Measure the Encoded Information