SeqGate-package {SeqGate} | R Documentation |
SeqGate is a method to filter lowly expressed features (e.g. genes).
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.
Christelle Reynès christelle.reynes@igf.cnrs.fr,
Stéphanie Rialle stephanie.rialle@mgx.cnrs.fr,
Maintainer: NA
Rialle, R. et al. (2020): SeqGate: a bioconductor package to perform data-driven filtering of RNAseq datasets manuscript in preparation
# 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,])