SeqGate-package: Filtering of Lowly Expressed Features

In SeqGate: Filtering of Lowly Expressed Features

Description

SeqGate is a method to filter lowly expressed features (e.g. genes).

Details

From a matrix of counts where lines correspond to features and columns to biological samples, provided as a SummarizedExperiment object, a threshold is computed and applied in order to filter lowly expressed features. The threshold is computed based on the distribution of counts measured along with zeros within replicates of the same condition. The objective of SeqGate is to rationalize the filtering step by using the information of replicate samples. The computed threshold corresponds to the count value below which we can not be sure that the count can be considered different from zero.
The filtering is made by calling the applySeqGate() function.

Author(s)

Christelle Reynès [e-mail],
Stéphanie Rialle [e-mail],
Maintainer: Stéphanie Rialle <[e-mail]>

References

Rialle, R. et al. (2020): SeqGate: a bioconductor package to perform data-driven filtering of RNAseq datasets manuscript in preparation

Examples

    # Loading of input data frame
    data(data_MiTF_1000genes)
    # Annotating conditions
    cond<-c("A","A","B","B","A","B")
    # Setting the SummarizedExperiment input
    rowData <- DataFrame(row.names=rownames(data_MiTF_1000genes))
    colData <- DataFrame(Conditions=cond)
    counts_strub <- SummarizedExperiment(
                        assays=list(counts=data_MiTF_1000genes),
                        rowData=rowData,
                        colData=colData)
    # Applying SeqGate
    counts_strub <- applySeqGate(counts_strub,"counts","Conditions")
    # Getting the matrix of kept genes after filtering
    keptGenes <- assay(counts_strub[rowData(counts_strub)$onFilter == TRUE,])

SeqGate documentation built on Jan. 24, 2021, 2:03 a.m.