Documentation

A datatype representing Infernal covariance models. This is a new representation that is incompatible with the one once found in Biobase. The most important difference is that lookups are mapped onto efficient data structures, currently PrimitiveArray.

1: Each State of a covariance model has up to 6 transition scores, hence we need s*6 cells for transitions.
2: Each State of a covariance has up to 16 emission scores, so we have s*16 cells for emissions, with unused cells set to a really high score.

On top of these basic structures, we then place additional high-level constructs.

3: paths are allowed transitions. This can safe a check, if the transition is encoded with a forbidden score.
4: localBegin and localEnd are local entry and exit strategies. A localBegin is a transition score to certain states, all such transitions are in begins. A localEnd is a transition score to a local end state.

NOTE that trustedCutoff > gathering > noiseCutoff

TODO as with other projects, we should not use Double's but Score and Probability newtypes.

Constructors

CM

Fields

name :: ModelIdentification: name of model as in tRNA
accession :: ModelAccession: RFxxxxx identification
trustedCutoff :: BitScore: lowest score of true member
gathering :: BitScore: all scores at or above gathering score are in the full alignment
noiseCutoff :: Maybe BitScore: highest score NOT included as member
transition :: PrimArray (Int, Int) Double
emission :: PrimArray (Int, Int) Double
paths :: Vector (Vector Double)
localBegin :: Vector Double
begins :: Vector Int
localEnd :: Vector Double
nodes :: Vector (Vector Int)

Instances

Show CM

type ID2CM = Map ModelIdentification CM Source

Map of model names to individual CMs.

type AC2CM = Map ModelAccession CM Source

Map of model accession numbers to individual CMs.