dmmEM-methods: dmmEM method

Description Usage Arguments Details Value Methods Author(s) See Also Examples

Description

This function trains the Dirichlet multinomial mixture models parameters for a set of profile matrices.

Usage

1
  dmmEM(x, K=6, alg=c("C", "R"))

Arguments

x

x can be a matrix, PFMatrixList or JASPAR2014 to be trained.

K

The maximal number of components to test in the mixture model when alg is "C". Then an optimal number of components between 1 and K will be chosen based on the fitness of the model.

The fixed number of components to use when alg is "R". The default is 6.

alg

The algorithm to use. "C" uses the implementation from DirichletMultinomial package which has more advanced feature and performance. "R" uses our own implemention in R.

Details

When using the implementation from DirichletMultinomial package, the final number of components can be 1:K. An internal selection will be made based on the maximum likelihood.

When using the implementation of R, the number of component is fixed to K.

Value

A list of trainned alpha0, pmix and likelihood during the training.

Methods

signature(x = "ANY")
signature(x = "matrix")
signature(x = "PFMatrixList")

Author(s)

Ge Tan

See Also

rPWMDmm

Examples

1
2
3
4
5
6
  
    data(MA0003.2)
    data(MA0004.1)
    pfmList <- PFMatrixList(pfm1=MA0003.2, pfm2=MA0004.1, use.names=TRUE)
    dmmParameters <- dmmEM(pfmList, K=6, alg="C")
  

ge11232002/TFBSTools documentation built on Sept. 12, 2021, 12:07 p.m.