filterdata: Data filtering

Description

This function allows filtering of genes and cells to be used in the RaceID3 analysis. It also can perform batch effect correction using an internal method or a recently published alternative mnnCorrect from the batchelor package.

Usage

filterdata(
  object,
  mintotal = 3000,
  minexpr = 5,
  minnumber = 5,
  LBatch = NULL,
  knn = 10,
  CGenes = NULL,
  FGenes = NULL,
  ccor = 0.4,
  bmode = "RaceID",
  verbose = TRUE
)

Value

An SCseq class object with filtered and normalized expression data.

Arguments

object: SCseq class object.
mintotal: minimum total transcript number required. Cells with less than mintotal transcripts are filtered out. Default is 3000.
minexpr: minimum required transcript count of a gene in at least minnumber cells. All other genes are filtered out. Default is 5.
minnumber: See minexpr. Default is 5.
LBatch: List of experimental batches used for batch effect correction. Each list element contains a vector with cell names (i.e. column names of the input expression data) falling into this batch. Default is NULL, i.e. no batch correction.
knn: Number of nearest neighbors used to infer corresponding cell types in different batches. Defult is 10.
CGenes: List of gene names. All genes with correlated expression to any of the genes in CGenes are filtered out for cell type inference. Default is NULL.
FGenes: List of gene names to be filtered out for cell type inference. Default is NULL.
ccor: Correlation coefficient used as a trehshold for determining genes correlated to genes in CGenes. Only genes correlating less than ccor to all genes in CGenes are retained for analysis. Default is 0.4.
bmode: Method used for batch effect correction. Any of "RaceID","mnnCorrect". If mnnCorrect from the batchelor package is desired, this package needs to be installed from bioconductor. Default is "RaceID".
verbose: logical. If FALSE then status output messages are disabled. Default is TRUE.

Examples

Run this code

sc <- SCseq(intestinalDataSmall)
sc <- filterdata(sc)

Run the code above in your browser using DataLab