... Denise¹

This work was partially supported by the French IMPG Bioinformatics program, the CNRS Specific Action "Modélisation et algorithmique des structures secondaires d'ARN" and the French education ministry founded action ACI IMPBio.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... grammars'^1.1

i.e. If a property is captured by a Prosite-based model, it is always possible to build a context-free grammar-based model capturing the same property.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... model^2.1

Which may seem a little lazy as the model is simple, but the size of a Markovian model grows exponentially with respect to its order, and rewriting manually a Markovian model of order 6 may bore the most wilful scientist.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... engines^3.1

Google computes its PageRank, which is a score for the relevance of a page, by modelling the behavior of a randomly clicking net-surfer using a Markovian model $\ldots$

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... weakly^3.2

When total control over the occurrences of $k+1$-length words(and shorter...) is required, one should consider using shuffling algorithms

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... phase^3.3

We recall that the phase equals to $n \mod Phases$, where $n$ is the number of bases previously generated.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... code^5.1

See http://www.chem.qmul.ac.uk/iubmb/misc/naseq.htmlhttp://www.chem.qmul.ac.uk/iubmb/misc/naseq.html for details

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.