349

Question 14.8

Data in biology and medicine are usually high-dimensional, i.e. they contain different

variables (features), correlations, confounders, batch effects and multicollinearity. For

this, machine learning methods in bioinformatics are helpful to structure the data and

extract relevant features, but also to develop classification models (predictive models).

PCA tries to decompose high dimensional data into principal components and reduce their

complexity (dimension reduction), but also to detect group differences. Cluster analyses

try to classify data into groups (clusters) with similar feature structures (characteristics),

e.g. healthy group (normal blood pressure) and diseased group (high blood pressure).

Regression analyses attempt to find correlations and relationships between a dependent

(“response variable”) and independent (“predictor variable”) variable, e.g. probability of

developing high blood pressure (and subsequently dying of heart failure) if one is over­

weight. It is important to also look again at the underlying algorithms and statistical

parameters to assess model goodness of fit. Further details and information can be found

in the papers Worster et al. (2007), Schneider et al. (2010), Singh and Mukhopadhyay

(2011) and Zwiener et al. (2011).

20.15 How Is Our Own Extremely Powerful Brain Constructed?

Question 15.1

For this purpose, please refer to the website https://www.neuron.yale.edu/neuron/ (tuto­

rial: https://www.neuron.yale.edu/neuron/docs available).

Question 15.2

For more information, please visit the website https://www.openworm.org/index.html.

Question 15.3

For more information, please visit the website https://www.humanconnectomeproject.org/.

Question 15.4

To do this, simply search the Internet, for example, with size constancy in the brain,

and inform.

Question 15.5

To do this, simply search the Internet and inform yourself (there are also nice Youtube

videos about this).

20.15  How Is Our Own Extremely Powerful Brain Constructed?