mixNBHMM: Three-state HMM with mixture model for differential peak...

Description Usage Arguments Value Author(s) References Examples

View source: R/mixNBHMM.R

Description

This function fits a three-state HMM with mixture model of Negative Binomials for differential peak detection across conditions with replicates.

Usage

1
mixNBHMM(object, control)

Arguments

object

a mixNBHMMDataSet

control

list of control arguments from controlEM()

Value

The mixNBHMMDataSet with the following list as metadata:

pi

Vector of estimated initial probabilities of the HMM

gamma

Vector of estimated transition probabilities of the HMM

psi

Vector of estimated means and dispersion parameters pertaining to the emission distributions

prob

Data table of window-based (rows) posterior probabilities of the HMM emission distributions (columns)

mixProb

Data table of window-based (rows) posterior probabilities of the mixture model from each mixture component (columns)

viterbi

Vector of Viterbi sequence of hidden states

logF

Data table of window-based (rows) log forward probabilities from the emission distributions (columns)

logB

Data table of window-based (rows) log backward probabilities from the emission distributions (columns)

logLik

Data table of window-based (rows) log probabilities from the emission distributions (columns)

parHist

Data table of parameter estimates (column) from every EM iteration (rows)

IMPORTANT: the output mixNBHMMDataSet has conditions and replicates reordered. Make sure to check colData() of your output.

Author(s)

Pedro L. Baldoni, pedrobaldoni@gmail.com

References

https://github.com/plbaldoni/mixNBHMM

Examples

1
2
3
data(ENCODE)
ENCODE <- createOffset(object = ENCODE,type = 'loess',span = 1)
## Not run: ENCODE <- mixNBHMM(object = ENCODE,control = controlEM())

plbaldoni/mixNBHMM documentation built on Dec. 24, 2019, 1:31 p.m.