293

states using the reduced-order binary decision diagram (ROBDD) algorithm. A steady

state is a network state to which the network returns, i.e. a stable state that is reached

again even after changes or disturbances and does not change. Especially helpful is the

perturbation function in SQUAD, with which one can write one’s own protocol and

define exactly which activation a certain node has at a certain point in time, in order to

map or predict the simulation e.g. according to the experimental data and the mutation

background (administration of a drug, knockouts, activation of receptors). For step c),

there is a good tutorial for SQUAD and an example network (T-helper cell network)

that you can practice with to get started. In addition, you can practice a bioinformatics

in silico simulation on your own by watching our online tutorial (https://www.ncbi.

nlm.nih.gov/pubmed/27077967). Here you will be shown all the necessary steps and

can “recreate” it yourself (scripts for simulation can be found there as well). An alter­

native is our own software Jimena, which also has a nice online tutorial (https://www.

bioinfo.biozentrum.uni-­wuerzburg.de/computing/jimena_c/).

How Do I Perform Metabolic Modeling of Metabolic Pathways/Fluxes?

It should be noted that one needs as input file for the elementary mode analysis a list of all

enzymes (reversible or irreversible should be decided according to the physiological con­

ditions) and a list of all enzyme substrates. Then the given algorithms can calculate all

modes effortlessly. But unfortunately, an enzyme can have more substrates than known in

the KEGG database (https://www.genome.jp/kegg/). So, in addition, one has to consider

biochemical knowledge, literature and databases like the BRENDA database (https://

www.brenda-­enzymes.de), which collects very many substrates for an enzyme, along with

information about Michaelis–Menten constant and biochemistry. Finally, metabolic

enzymes without substrate or under special conditions (e.g. without iron) can suddenly

acquire new regulatory functions.

It is interesting to note that dynamic modelling using gene expression data is only an

approximation of the true fluxes, but in practice such gene expression data are much more

likely to be available than the laborious determination of metabolite concentrations.

Dynamic modelling can then also look at true concentrations and kinetics for metabolites,

for example using the software PLAS (Power Law Analysis Software – modelled with

power functions; https://enzymology.fc.ul.pt/software/plas/). In addition, for the calcula­

tion of metabolic pathways/fluxes (elementary mode analysis and flux mode calculation)

there are our developed programs Metatool (calculation of all possible metabolic path­

ways; the Metatool input files have to be edited exactly, otherwise the simple program

crashes. It is recommended to start with a simple example, see online tutorial, and then

adapt the example file step by step) and YANAsquare (calculation possible for certain situ­

ations, e.g. exponential growth with glucose as nutrient source or without oxygen: which

pathways are then active and how strongly, see exercise tasks for elementary mode analy­

sis). As a first introduction and good basis for metabolic analysis, the online tutorials for

Metatool (https://www.bioinfo.biozentrum.uni-­wuerzburg.de/computing/metatool_4_5/;

19.4  Cellular Communication, Signalling Cascades, Metabolism, Shannon Entropy