226

humans, and perhaps want to better cure their diseases or simply better understand and

recognize how they are built and what separates humans from animals (anthropology).

Consider, both our now very good anti-viral drugs for HIV infection and our modern tar­

geted therapies for cancer (Duell et al. 2017) require bioinformatics computations to a

very significant degree. For example, the Antiretroviral Therapy Cohort Collaboration

(2008) showed that even with HIV disease, one has a near normal life expectancy with

early therapy. The new approaches found for this, as well as the countless molecular ther­

apy successes in the last two decades, would not have been possible without the support of

molecular experiments by bioinformatics. Intriguingly, the article by Lengauer et  al.

(2014) describes how bioinformatics can help develop optimal individualized therapy

against HIV. Similarly, Stratmann et al. (2014), Göttlich et al. (2016) and Baur et al. (2020)

step by step improve a targeted cancer therapy using bioinformatics and cell culture exper­

iments. The same is true for attempts to better understand the human brain. Here, com­

puter models are important and are currently also being massively funded as an EU lead

project (“Blue Brain” project of the EU). Perhaps a better strategy is to simply listen

carefully to the brain and not immediately think of new computer architectures. This is

precisely the goal of the US government’s Brain Activity project, which is even three times

more heavily funded than the EU project.

16.1

Solving Problems Using Bioinformatics

A common thread in all the great challenges of bioinformatics is climbing to a new level

of language. Whether it is understanding the genetic (protein prediction) and genomic

(gene prediction) code and correctly predicting proteins from foreign genomes or translat­

ing the sequence of a protein into three-dimensional protein structures, one is always

climbing a new language level. Of course, this is even more true when doing systems biol­

ogy, i.e., approaching the very essence of biological regulation in a deeper way and under­

standing forward and feedback loops, recognizing stable system states and can be said in

the same way for ecosystem modeling (Kriegler et al. 2009). Thus, an important starting

point for bioinformatics is first of all interest in the biological problem one wants to

explore. Once one has delved a bit deeper into the problem, it is a matter of finding the

right language to now build a suitable model for this phenomenon. This makes a great deal

clear from the outset: we do not have the truth. It could well be that with a different lan­

guage, with new software or even just a different perspective on the biological question,

completely different insights will be possible than with the first approach just chosen. It is

equally clear that only close collaboration with experimental biologists can help to figure

out the best models. “True”, i.e., internally consistent and correct, should be as consistent

as possible in any model. But which model I then actually use is determined solely by the

16  Bioinformatics Connects Life with the Universe and All the Rest