|
|
GeneMark
|
Universal gene prediction program – predicts protein-coding regions, in genomes
of any type.
|
|
MkMat
|
Model building tool for GeneMark. Makes matrices from sample protein-coding and
non-coding sequences selected by the user.
|
|
Prokaryotic GeneMark.hmm
|
Predicts genes in prokaryotic genomes. Leaves less room for expert than GeneMark
|
|
GeneMarkS
|
Self-training version of GeneMark.hmm for prokaryotic genomes.
|
|
Eukaryotic GeneMark.hmm
|
Gene prediction program – predicts protein-coding regions in eukaryotic genomes.
(exon-intron structures)
|
|
Custom models
|
We can build models for a new eukaryotic genome, you name it.
|
|
Remark: The difference between GeneMark and GeneMark.hmm program can be described
in terms of the theory of Hidden Markov models (HMM)
- the GeneMark algorithm can be approximated by the Forward-Backward algorithm computing
local a posteriori probability of protein coding function.
- the GeneMark.hmm program implements Viterbi algorithm for semi Markov HMM to determine
a global parse of a DNA sequence into protein coding and non-coding regions.