compGoM: compGoM: compare GoM model fits across K or across different...

In CountClust: Clustering and Visualizing RNA-Seq Expression Data using Grade of Membership Models

Description

This function takes the FitGoM/maptpx fitted model and computes log likelihood, BIC and null model loglikelihood for the fitted GoM models.

Usage

compGoM(data, model)

Arguments

data	matrix on which GoM model is fitted (samples along rows, genes along columns)
model	FitGoM ormaptpx::topics function output (either a class topics or a list of class topics).

Value

compGoM_models a vector list that returns the BIC and loglikelihood values for each of the fitted models in model.

Examples

read.data <- function() {
  x <- tempfile()
  download.file(paste0("https://cdn.rawgit.com/kkdey/",
                         "singleCellRNASeqMouseDeng2014",
                         "/master/data/Deng2014MouseEsc.rda"),
                destfile = x, quiet = TRUE)
  z <- get(load((x)))
  return(z)
  }
Deng2014MouseESC <-read.data()

# Extract observed counts
deng.counts <- Biobase::exprs(Deng2014MouseESC)

# Import GoM fitting results
data("MouseDeng2014.FitGoM")
names(MouseDeng2014.FitGoM)

compGoM(data = t(deng.counts),
           model = MouseDeng2014.FitGoM)
compGoM(data = t(deng.counts),
           model = MouseDeng2014.FitGoM$clust_3)

CountClust documentation built on Nov. 8, 2020, 5:01 p.m.