hummingbirdEMinternal: EM Algorithm (internal)

In hummingbird: Bayesian Hidden Markov Model for the detection of differentially methylated regions

Description

Expectation-Maximization Algorithm for Fitting the Hidden Markov Model. This function reads in methylated and unmethylated read count data, transforms it into logarithm bin-wise data, sets up initial values and implements the EM algorithm to estimate HMM parameters and find the best sequence of hidden states based on model fitting.

Usage

hummingbirdEMinternal(normM, normUM, abnormM, abnormUM, pos, binSize)

Arguments

normM	A matrix containing the methylated read count data of the normal group. Each column of a matrix represents a replicate and each row represents a CpG position.
normUM	A matrix containing the unmethylated read count data of the normal group. Each column of a matrix represents a replicate and each row represents a CpG position.
abnormM	A matrix containing the methylated read count data of the abnormal group. Each column of a matrix represents a replicate and each row represents a CpG position.
abnormUM	A matrix containing the unmethylated read count data of the abnormal group. Each column of a matrix represents a replicate and each row represents a CpG position.
pos	The CpG position information.
binSize	The size of a bin.

Details

Users do not need to call this function directly. This is a low-level function used by the higher-level function in the hummingbird package, the hummingbirdEM.

Value

obs	For each bin: The predicted direction of methylation change ("0" means a predicted normal bin; "1" means a predicted hypermethylated bin; "-1" means a predicted hypomethylated bin). The distance between the current bin and the bin ahead of it, the start and end positions of each bin.
normAbnorm	The average methylation rate of normal and abnormal groups.

See Also

Users may call the hummingbirdEM function.

Examples

library(GenomicRanges)
library(SummarizedExperiment)

# Load sample dataset containing input data 
data(exampleHummingbird)

# Run the EM (internal) function
hmmbirdEMinternalOutput <- hummingbirdEMinternal(
    normM = assays(exampleSEControl)[["normM"]],
    normUM = assays(exampleSEControl)[["normUM"]],
    abnormM = assays(exampleSECase)[["abnormM"]],
    abnormUM = assays(exampleSECase)[["abnormUM"]],
    pos = pos, binSize = 40)

hummingbird documentation built on April 18, 2021, 6 p.m.