getUNDmatrix: getUNDmatrix

In bicbioeng/BrainSABER: Brain Span Atlas in Biobase Expressionset R toolset

Description

This function returns a matrix showing whether gene expression values in dataSet are up-regulated, down-regulated, or normal. method = "discrete" will function on any CellScabbard object, while method = "log2FC" requires a trimmed data set as returned by getTrimmedExternalSet and a matching subset of AIBSARNA as returned by getRelevantGenes. Results are stored in the 'UNDmatrices' slot of the dataSet if it's a CellScabbard object.

Usage

getUNDmatrix(
  dataSet,
  relevantGenes = NULL,
  method = c("discrete", "log2FC"),
  up_threshold = 0.5,
  down_threshold = -0.5,
  matrix_type = c("num", "char")
)

Arguments

dataSet	a CellScabbard or SummarizedExperiment object
relevantGenes	(optional) a SummarizedExperiment and subset of AIBSARNA
method	"discrete" applies thresholds directly to expression data. "log2FC" applies thresholds to the log2 fold-change between the expression data of each sample from dataSet and relevantGenes.
up_threshold	a numerical value defining the lower bound (inclusive) by which to consider a gene up-regulated, defaults to 0.5
down_threshold	a numerical value defining the upper bound (inclusive) by which to consider a gene down-regulated, defaults to -0.5
matrix_type	either "num" for a numerical matrix with -1 indicating down-regulation, 1 indicating up-regulation, and 0 indicating normal, or "char" for a character matrix with "D" indicating down-regulation, "U" indicating up-regulation, and "N" indicating normal

Value

a list containing as many numerical or character matrices as samples in dataSet, with each matrix having as many rows as genes in dataSet and as many columns as samples in relevantGenes

Examples

AIBSARNA <- buildAIBSARNA(mini = TRUE)
# Example 1 - using CellScabbard class
# get a random sample of 3 genes
totalGenes <- nrow(AIBSARNA)
gene_idx <- sample.int(totalGenes, 3)
sample_idx <- c(1,3,5)
# Subset AIBSARNA
exprs <- assay(AIBSARNA)[gene_idx, sample_idx]
fd <- rowData(AIBSARNA)[gene_idx, ]
pd <- colData(AIBSARNA)[sample_idx, ]
# build a trimmed data set
myGenes <- CellScabbard(exprsData = exprs, phenoData = pd, featureData = fd,
                        AIBSARNA = AIBSARNA, autoTrim = TRUE)
UNDmatrices(myGenes) <- getUNDmatrix(myGenes, method = "discrete",
                                     up_threshold = 3,
down_threshold = 1, matrix_type = "char")
UNDmatrices(myGenes)
UNDmatrices(myGenes) <- getUNDmatrix(myGenes, method = "log2FC",
                                     up_threshold = 3,
down_threshold = 1, matrix_type = "num")
UNDmatrices(myGenes)

# Example 2 - manual gene selection and relevant gene extraction
myGenes <- c(4.484885, 0.121902, 0.510035)
names(myGenes) <- c("TSPAN6", "DPM1", "C1orf112")
myGeneSet <- getRelevantGenes(myGenes, AIBSARNA = AIBSARNA,
    AIBSARNAid = "gene_symbol")
myTrimmedGeneSet <- getTrimmedExternalSet(myGeneSet,
    dataSetId = "gene_symbol", AIBSARNA, AIBSARNAid = "gene_symbol")
myUNDnumericalMatrix <- getUNDmatrix(myTrimmedGeneSet, method = "discrete",
    up_threshold = 3, down_threshold = 1, matrix_type = "num")
myUNDcharacterMatrix <- getUNDmatrix(myTrimmedGeneSet, myGeneSet,
                                     method = "log2FC",
    up_threshold = 3, down_threshold = 1, matrix_type = "char")

bicbioeng/BrainSABER documentation built on Oct. 10, 2021, 6:38 a.m.