256

Hardware  All the physical aspects of a computer (chassis, transistors, input and output

devices such as printers, memory disks, etc.).

Hidden Markov model  Stochastic model that calculates certain transition probabilities

from an observable system state on the basis of a Markov chain and allows statements

about the hidden states.

Homology Protein Structure Prediction  Method for predicting protein structure, using

known protein structures as templates.

Hypothesis-Free  On the positive side, this is research that looks at the results without

bias and develops appropriate conclusions. On the other hand, it usually means expensive

omics experiments. For these, it is not a good idea to just produce a lot of data without an

initial scientific hypothesis and hope that the statistician will then find something “signifi­

cant”. The following paper explains this problem very well: Ioannidis JPA (2005) Why

most published research findings are false. PLoS Med 2(8): e124. 

In Silico Analysis  with the computer, computer-assisted analysis. Literally: "in silicon"

(Latin).

Internet Protocol (IP for Short)  Used to exchange data packets from computer to com­

puter, so that this is done as quickly as possible, making use of all free space and also

coping with a major computer failure.

IP  Abbreviation for the Internet Protocol (see there).  Knowledge-based 

Bioinformatics work, such as the creation of a network, based on literature and

expert knowledge. 

Last Universal Common Ancestor (LUCA) or LCA  Last common ancestor of all life,

which can be inferred bioinformatically via protein sequences and protein families (lived

about 2.5 billion years ago, 1000–1500 basic protein genes, the COGs [clusters of ortholo­

gous genes; bacterial sequence families to genes with the same function] give an approxi­

mate impression of these oldest genes).

Local Alignment  Two sequences are compared by contrasting which amino acid resi­

dues are modified and which are conserved (see also global alignment). In local alignment,

one looks for only one piece, especially very well, to catch a local piece that is particularly

similar, especially the domain that has the highest sequence similarity, i.e. a characteristic

domain of the protein. But then please remember that after that you should also examine

the other pieces in the protein and assign the function. For local alignment there is an exact

method, the Smith and Waterman search. This is slow, but accurate. Various heuristic

methods (not quite exact, but much faster with only a few errors) use this method, where

after a (FASTA; fast alignment) or a double (BLAST) fast index search then an exact

Smith and Waterman alignment is done.

18  Glossary