CellScore: CellScore evaluates the identity of cells undergoing cell...

In nmah/CellScore: Tool for Evaluation of Cell Identity from Transcription Profiles

CellScore evaluates the identity of cells undergoing cell type transition

Description

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. 'transitions' is mandatory, and either 'data' or all three of 'scores.onoff', 'scores.cosine' and 'pdata' are as well. If you provide 'data', then 'scores.onoff', 'scores.cosine' and 'pdata' will override the respective calculations only.

Usage

CellScore(
  transitions,
  data = NULL,
  scores.onoff = NULL,
  scores.cosine = NULL,
  pdata = NULL
)

Arguments

transitions	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.
data	a SummarizedExperiment or ExpressionSet containing data matrices of normalized expression data, present/absent calls, a gene annotation data frame and a phenotype data frame.
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().
pdata	a data frame with samples as rows, variables as columns.

Value

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.

See Also

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

Examples

## 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(cell.change, data = eset.sub, scores.onoff = individ.OnOff$scores,
                          scores.cosine = cs$cosine.samples)
}

nmah/CellScore documentation built on May 4, 2023, 2:52 p.m.