Principal component analysis plot

Principal component analysis plot

plotPCA(object, ...)

# S4 method for bcbioRNASeq
plotPCA(object, normalized = c("vst", "rlog"),
  ...)

Arguments

object	Object.
normalized	character(1) or logical(1). Normalization method to apply: FALSE: Raw counts. When using a tximport-compatible caller, these are length scaled by default (see countsFromAbundance argument). When using a featureCounts-compatible caller, these are integer. tximport caller-specific normalizations: "tpm": Transcripts per million. Additional gene-level-specific normalizations: TRUE / "sf": Size factor (i.e. library size) normalized counts. See DESeq2::sizeFactors() for details. "fpkm": Fragments per kilobase per million mapped fragments. Requires fpkm = TRUE in bcbioRNASeq() call and gene annotations in rowRanges() with defined width. See DESeq2::fpkm() for details. "vst": Variance-stabilizing transformation (log2). Requires vst = TRUE to be set during bcbioRNASeq() call. See DESeq2::varianceStabilizingTransformation for more information. "rlog": Regularized log transformation (log2). Requires rlog = TRUE to be set during bcbioRNASeq() call. See DESeq2::rlog() for details. "tmm": Trimmed mean of M-values. Calculated on the fly. See edgeR::calcNormFactors() for details. "rle": Relative log expression transformation. Calculated on the fly. See relativeLogExpression() for details. Note that logical(1) support only applies to counts(). Other functions in the package require character(1) and use match.arg() internally.
...	Passthrough to SummarizedExperiment method defined in acidplots. See acidplots::plotPCA() for details.

Value

ggplot or DataFrame.

Note

Updated 2019-09-15.

Principal component analysis

PCA (Jolliffe, et al., 2002) is a multivariate technique that allows us to summarize the systematic patterns of variations in the data. PCA takes the expression levels for genes and transforms it in principal component space, reducing each sample into one point. Thereby, we can separate samples by expression variation, and identify potential sample outliers. The PCA plot is a way to look at how samples are clustering.

References

Jolliffe, et al., 2002.

See also

DESeq2::plotPCA().

We're using a modified version of the DESeqTransform method here.

methodFunction(
    f = "plotPCA",
    signature = "DESeqTransform",
    package = "DESeq2"
)

Examples

data(bcb)
plotPCA(bcb, label = FALSE)
#> Using vst counts.
#> Plotting PCA using 100 features.
plotPCA(bcb, label = TRUE)
#> Using vst counts.
#> Plotting PCA using 100 features.