c3_test: c3_test

In VCCRI/C3: An R package for cross-species compendium-based cell-type identification for transcriptomic profile

Description

This function performs the XGSA Fisher's exact tests between the marker genes of the query data and each cell type or tissue in the marker gene compendium.

Usage

c3_test(processed.queryData, marker.gene.compendium, min = 1, max = 5000,
  top.expressed.genes = NULL)

Arguments

processed.queryData	A list containing the average gene expression values for the cell types and/or tissues with ensembl gene IDs and the species name, as returned by 'preprocess_querydata'.
marker.gene.compendium	A list containing the data of marker gene compendium, as returned by 'make_marker_gene_compendium'.
min	The minimum number of marker genes required in a marker gene set to be tested. Setting this higher may reduce false positives. Default is 0.
max	The maximum number of genes allowed in a marker gene set to be tested. Default is '4000'.
top.expressed.genes	The number of highest expressed genes from each cell/tissue that will be considered for marker gene status. Default is NULL, i.e., set the same number internally as the compendium.

Details

This function performs the XGSA Fisher's exact tests between the marker genes of the query data and each cell type or tissue in the marker gene compendium.

Value

This function returns a list containing the XGSA p-values and the corresponding overlapping genes from the query data and the compendium. Each element of the list corresponds to one query sample vs. one compendium cell type.

Examples

## Here we will use "human.encode.data" and "human.encode.data.descriptor" from C3 repository to make the compendium.
## These data sets are loaded automatically with the package.
## The gene expression data sets are in list format. Here every list element contains transcriptomic profile data of a cell or tissue type.
human.compendium<-make_gene_expression_compendium(gene.expression.data=human.encode.data, experiment.descriptor=human.encode.data.descriptor, expression.data.format="list")

## Then we will add "hawse.human.lens.data" to the compendium. This data is also loaded with the package.
## This data set has 2 different cell types - lens epithelial cell (LEC) and lens fiber cell (LFC). Each cell type has 3 biological replicates.
Hawse.data.descriptor<-c("Hawse_LEC","Hawse_LEC","Hawse_LEC","Hawse_LFC","Hawse_LFC","Hawse_LFC")
Hawse.human.compendium<-add_sample_into_compendium(compendium.data=human.compendium, sample.data=hawse.human.lens.data, species = "hsapiens", data.format = "matrix", geneID = "external_gene_name", experiment.descriptor = Hawse.data.descriptor)

## Next we will make the marker gene compendium from the Hawse.human.compendium
human.marker.gene.compendium<-make_marker_gene_compendium(compendium.data =Hawse.human.compendium, specific.cutoff = 0.05, top.expressed.genes = 500)

## Now we will pre-process the "hoang.mouse.lens.data" (query data). This data is also loaded with the package.
## This data set also contains 2 separate cell type: LEC and LFC. Each cell contains average expression value of 3 replicates.
Hoang.processed.query.data<-preprocess_querydata(cell.tissue.data = hoang.mouse.lens.data, species = "mmusculus", data.format = "matrix", geneID = "external_gene_name")

## Finally we will perform the test of the query data with the human.marker.gene.compendium
Hoang.data.test.result<-c3_test(processed.queryData = Hoang.processed.query.data, marker.gene.compendium = human.marker.gene.compendium)
head(sort(Hoang.data.test.result$Hoang_LEC$pvalue))
head(sort(Hoang.data.test.result$Hoang_LFC$pvalue))

VCCRI/C3 documentation built on May 14, 2019, 8:41 a.m.