260

thesis, the freshly synthesized molecules are separated again by heating each 1000 base

pairs, and the two primers are then allowed to reattach to the new strands, thus obtaining

more and more identical DNA strands as long as there are enough primer molecules and

polymerase for the PCR. 

Polynomial Complexity  Problems that are not very computationally intensive and whose

mathematical description is done with a polynomial (computation time depends on the

length) (see also P-problems). 

Positive/Negative Feedback Loops  Feedback loops in networks that reinforce every­

thing and stabilize the system (positive feedback loops) or dampen excessive regulation

(negative feedback loops).

P-problems  Problems that do grow polynomial as the sequence or number of units

increases for which I make my prediction, but it happens exactly and clearly. So there is a

safe solution strategy with clear computational timing. Examples are database searches or

RNA folding. Both grow quadratically with sequence length for many related algorithms,

where for databases it is the product of search sequence and total database length (see also

polynomial complexity). 

Precision  (specificity) Correct-negative rate, i.e. the proportion of correctly negatively

grouped hits out of the total of correctly negative hits, e.g. predicted as non-interacting

interaction partners that also show no interaction experimentally, or people grouped as

healthy who are actually healthy.

Primary Database  Databases that contain only the basic data and information, such as

protein sequences.

Primary Metabolites  See secondary metabolism.

Primary Metabolism  see Secondary metabolism.

Principal Component Analysis  (PCA). 

Principal Component Analysis (PCA)  Method of multivariate statistics. It aims to

transform high-dimensional data into a new coordinate system (usually 2D) in order to

reduce complexity in the data and extract relevant variables (principal components). 

Programs  Tools (software) used to examine and analyze data sets or data from experi­

ments. Programs first consist of a declaration part, which defines the variables and data

fields used, followed by the calculation part. Typically, the computational part consists of

a read/input part, a main loop, and an output part. The main loop reads the input data and

performs the calculations, often accesses subroutines itself, writes results to the output

part, and monitors the sequence of the program through logical queries until the program

finally processes everything and stops. In practice, stopping or halting a program is not

always easy to predict.

18  Glossary