CellScore: CellScore evaluates the identity of cells undergoing cell...
In CellScore: Tool for Evaluation of Cell Identity from Transcription Profiles

Description Usage Arguments Value See Also Examples

This function will calculate the CellScore (summary score) for a cell that is undergoing a transition in cell identity from a starting cell type to a target cell type.

1	CellScore(eset, cell.change, scores.onoff, scores.cosine)

`eset`	an ExpressionSet containing data matrices of normalized expression data, present/absent calls, a gene annotation data frame and a phenotype data frame.
`cell.change`	a data frame containing three columns, one for the start (donor) test and target cell type. Each row of the data. frame describes one transition from the start to a target cell type.
`scores.onoff`	a data.frame of OnOff Scores for all samples in the expression matrix as generated by the function OnOff().
`scores.cosine`	a numeric matrix of cosine similarity between general groups, subgroups and individual samples as calculated by the function CosineSimScore().

The function returns a data frame with 29 columns and M*N rows, where M is the number of unqiue start and target cell types pairs listed in the cell.change argument, while N is the number of all samples in the input dataset eset. The columns include sample phenotype features and all score (components), including the on/off score, cosine similarity and CellScore.

CosineSimScore, OnOff for detials on specfic score calculations, and hgu133plus2CellScore for details on the specific expressionSet object that shoud be provided as an input.

## Load the expression set for the standard cell types
library(Biobase)
library(hgu133plus2CellScore) # eset.std

## Locate the external data files in the CellScore package
rdata.path <- system.file("extdata", "eset48.RData", package = "CellScore")
tsvdata.path <- system.file("extdata", "cell_change_test.tsv",
                             package = "CellScore")

if (file.exists(rdata.path) && file.exists(tsvdata.path)) {

   ## Load the expression set with normalized expressions of 48 test samples
   load(rdata.path)

   ## Import the cell change info for the loaded test samples
   cell.change <- read.delim(file= tsvdata.path, sep="\t",
                             header=TRUE, stringsAsFactors=FALSE)

   ## Combine the standards and the test data
   eset <- combine(eset.std, eset48)

   ## Generate cosine similarity for the combined data
   ## NOTE: May take 1-2 minutes on the full eset object
   ## so we subset it for 4 cell types
   pdata <- pData(eset)
   sel.samples <- pdata$general_cell_type %in% c("ESC", "EC", "FIB", "KER")
   eset.sub <- eset[, sel.samples]
   cs <- CosineSimScore(eset.sub, cell.change, iqr.cutoff=0.1)

   ## Generate the on/off scores for the combined data
   individ.OnOff <- OnOff(eset.sub, cell.change, out.put="individual")

   ## Generate the CellScore values for all samples
   cellscore <- CellScore(eset.sub, cell.change, individ.OnOff$scores,
                          cs$cosine.samples)
}