EMclust: EM (Expectation maximization) function

In daewoooo/SaaRclust: R Package to cluster long sequencing reads into chromosomes to facilitate de novo genome assembly.

Description

This function performs basic steps of EM algorithm.

Usage

EMclust(counts.l, theta.param = NULL, pi.param = NULL,
  num.iter = 100, alpha = 0.1, logL.th = 1, log.scale = FALSE)

Arguments

counts.l	A list of plus and minus alignments per genomic region.
theta.param	A list of estimated cell types for each single cell. (rows=clusters, cols=strand states)
pi.param	A vector of estimated sizes of each cluster based on initial hard clustering.
num.iter	Set number of iteration to EM algorithm.
alpha	Estimated level of background in Strand-seq reads.
logL.th	Set the difference between objective function from the current and the previous interation for EM algorithm to converge.

Value

A list of various exported results [pVal, pVal.logL, log.l, theta.param, pi.param].

Author(s)

David Porubsky, Maryam Ghareghani

daewoooo/SaaRclust documentation built on May 28, 2019, 7:50 p.m.