 |
|
Version 4.0.0
|
The inheritance links between classes are essential to understand the design of seq++ and the behaviour of each object. Please browse the top task bar to discover the Class Hierarchy.
Basicaly seq++ focuses on both aspects:
- the key of the seq++ flexibility is the possibility to customize any alphabet thanks to the Translator class. Therefore a seq++-based software can support alphabets of DNA, codons, proteins, secondary structure, hydrophobicity class and so on.
- the SequenceSet and Sequence classes are dedicated to the storage and the manipulation of external sequences provided by the user. This is done keeping in mind the purpose of modeling sequences with Markov models: at each position of the sequences the associated markovian-predictor (word, see PrimaryCount) is coded and stored.
- Markovian modelling (estimation, simulation...)
- the PhasedMarkov object represents the general Markov model: the order and the number of phases have to be selected. Markov focuses on non-phased models.
- PhasedPMarkov and PMarkov are dedicated to Parcimonious Markov models. Note that the definition of a set of partitions of the alphabet (see Partition) can be required.
- PhasedMTDMarkov and MTDMarkov are dedicated to Mixture Transition Distribution Markov models. Note that the estimation is performed by a E.M. algorithm so that E.M. parameters have to be set up.
- PhasedVLMarkov and VLMarkov are dedicated to Variable Length Markov models. Beware that these models require a cutoff value which can be estimated by optimizing the BIC criteria.
Please watch the References page for more details on the models.
Generated on Sun Apr 3 14:16:10 2005 for seqpp by
1.3.9.1