Visualization Methods for Differential Expression Analysis

exploreDDS Convenience wrapper function to transform raw read counts using the DESeq2::DESeq2-package() package transformations methods. The input file has to contain all the genes, not just differentially expressed ones.

exploreDDSplot Scatterplot of transformed counts from two samples or grid of all samples.

hclustplot This function computes the sample-wise correlation coefficients using the stats::cor() function from the transformed expression values. After transformation to a distance matrix, hierarchical clustering is performed with the stats::hclust() function, and the result is plotted as a dendrogram.

heatMaplot This function performs hierarchical clustering on the transformed expression matrix. It uses, by default, a Pearson correlation-based distance measure and complete linkage for cluster join.

PCAplot This function plots a Principal Component Analysis (PCA) from transformed expression matrix. This plot shows samples variation based on the expression values and identifies batch effects.

MDSplot This function computes and plots multidimensional scaling analysis for dimension reduction of count expression matrix. Internally, it is applied the stats::dist() function to the transformed count matrix to get sample-to-sample distances.

GLMplot This function computes and plots generalized principal components analysis for dimension reduction of count expression matrix.

MAplot This function plots log2 fold changes (y-axis) versus the mean of normalized counts (on the x-axis). Statistically significant features are colored.

tSNEplot This function computes and plots t-Distributed Stochastic Neighbor embedding (t-SNE) analysis for unsupervised nonlinear dimensionality reduction of count expression matrix. Internally, it is applied the Rtsne::Rtsne() function, using the exact t-SNE computing with theta=0.0.

volcanoplot A simple function that shows statistical significance (p-value) versus magnitude of change (log2 fold change).

exploreDDS(countMatrix, targets, cmp = cmp[[1]], preFilter = NULL, transformationMethod = "raw", blind = TRUE)
exploreDDSplot(countMatrix, targets, cmp = cmp[[1]],preFilter = NULL, samples, blind = TRUE,  scattermatrix = FALSE, plotly = FALSE, savePlot = FALSE, filePlot = NULL)
hclustplot(exploredds, method = "spearman", plotly = FALSE, savePlot = FALSE, filePlot = NULL)
heatMaplot(exploredds, clust, DEGlist = NULL, plotly = FALSE, savePlot = FALSE, filePlot = NULL, ...)
PCAplot(exploredds, plotly = FALSE, savePlot = FALSE, filePlot = NULL)
MDSplot(exploredds, method = "spearman", plotly = FALSE, savePlot = FALSE, filePlot = NULL)
GLMplot(exploredds, plotly = FALSE, savePlot = FALSE, filePlot = NULL, ...)
MAplot(exploredds, lfcShrink = FALSE, padj.cutoff = 0.05, plotly = FALSE, savePlot = FALSE, filePlot = NULL)
tSNEplot(countMatrix, targets, plotly = FALSE, savePlot = FALSE, filePlot = NULL, ...)
volcanoplot(degseqDF, comparison = "M12-A12", filter = c(Fold = 2, FDR = 10), genes = "NULL", plotly = FALSE, savePlot = FALSE, filePlot = NULL)

Arguments

countMatrix	`date.frame` or `matrix` containing raw read counts.
targets	targets `data.frame`.
cmp	`character matrix` where comparisons are defined in two columns. This matrix should be generated with the `systemPipeR::readComp()` function from the targets file. Values used for comparisons need to match those in the `Factor` column of the targets file.
preFilter	allows removing rows in which there are very few reads. Accepts a numeric value with the minimum of total reads to keep. Default is `NULL`.
transformationMethod	a `character string` indicating which transformation method it will be used on the raw read counts. Supported methods include `rlog` and `vst` using the `DESeq2` package or default `raw` for no data transformation.
blind	logical, whether to blind the transformation to the experimental design (see varianceStabilizingTransformation), from `DESeq2::vst()` or `DESeq2::rlog()`.
samples	a `character vector` of two samples or `ALL` samples in the dataset. Could be specified the `SampleName` column name of the targets file or the respective numeric values. Also, if set as `ALL`, a correlation matrix it will be plot.
scattermatrix	if `samples` set as `ALL`, requires to assign `TRUE` to build a correlation matrix and plot the correlogram of all the samples.
plotly	logical: when `FALSE` (default), the `ggplot2` plot will be returned. `TRUE` returns the `plotly` version of the plot.
savePlot	logical: when `FALSE` (default), the plot will not be saved. If `TRUE` the plot will be saved, and requires the `filePlot` argument.
filePlot	file name where the plot will be saved. For more information, please consult the `ggplot2::ggsave()` function.
exploredds	object of class `DESeq2::DESeqTransform()`. `GLMplot` and `MAplot`, `exploredds` object of class `DESeq2::DESeqDataSet()`, generated from `exploreDDS` function.
method	a `character string` indicating which correlation coefficient is to be computed, based on the `stats::cor()` function. Options are: c("pearson" "kendall", "spearman").
clust	select the data to apply the distance matrix computation. If `samples` selected, it will be applied the `stats::dist()` function to the transformed count matrix to get sample-to-sample distances. If `ind`, it is necessary to provide the list of differentially expressed genes, for the `exploredds` subset.
DEGlist	List of up or down regulated gene/transcript identifiers meeting the chosen filter settings for all comparisons defined in data frames `pval` and `log2FC`.
lfcShrink	logiacal. If `TRUE` adds shrunken log2 fold changes (LFC) to the object.
padj.cutoff	filter cutoffs for the p-value adjusted.
degseqDF	object of class `data.frame` generated by `systemPipeR::run_edgeR()` or `systemPipeR::run_DESeq2()`.
comparison	`character vector` specifying the factor names for comparison.
filter	Named vector with filter cutoffs of format c(Fold=2, FDR=1) where Fold refers to the fold change cutoff (unlogged) and FDR to the p-value cutoff.
genes	`character vecto`r of genes names to show on the plot.
...	For `heatMaplot` additional parameters for the `pheatmap::pheatmap()` function. For `GLMplot` additional parameters for the `glmpca::glmpca()` function. For `tSNEplot` additional parameters for the `Rtsne::Rtsne()` function.

Value

exploreDDS function returns an object of class DESeq2::DESeqTransform(). heatMaplot function returns an object of pheatmap or plotly class. exploreDDSplot function returns an object of ggplot2 plot. hclustplot, PCAplot, MDSplot, GLMplot, MAplot, tSNEplot, and volcanoplot functions returns an object of ggplot or plotly class.

Details

Note that the recommendation is to use the resulting transformed values in the transformationMethod argument only for visualization and clustering, not for differential expression analysis which needs raw counts. Users are strongly encouraged to consult the DESeq2::DESeq2-package() vignette for more detailed information on this topic and how to properly run DESeq2 on data sets with more complex experimental designs.

References

For more details on DESeq2, please consult the following page: DESeq2. For more details on targets file definition, please consult the following page: systemPipeR.

Author

Daniela Cassol

Examples

## Targets file
targetspath <- system.file("extdata", "targets.txt", package="systemPipeR")
targets <- read.delim(targetspath, comment="#")
cmp <- systemPipeR::readComp(file=targetspath, format="matrix", delim="-")
## Count table file
countMatrixPath <- system.file("extdata", "countDFeByg.xls", package="systemPipeR")
countMatrix <- read.delim(countMatrixPath, row.names=1)

## Run exploredds
exploredds <- exploreDDS(countMatrix, targets, cmp=cmp[[1]], preFilter=NULL, transformationMethod="rlog")
#> Warning: some variables in design formula are characters, converting to factors

## Plot exploreDDSplot
exploreDDSplot(countMatrix, targets, cmp=cmp[[1]], preFilter=NULL, samples=c(3,4))

## Plot hclustplot
hclustplot(exploredds, method = "spearman")

## Plot heatMaplot Samples
heatMaplot(exploredds, clust="samples", plotly = TRUE)

## Individuals genes identified in DEG analysis
### DEG analysis with `systemPipeR`
degseqDF <- systemPipeR::run_DESeq2(countDF = countMatrix, targets = targets, cmp = cmp[[1]], independent = FALSE)
#> Warning: some variables in design formula are characters, converting to factors
DEG_list <- systemPipeR::filterDEGs(degDF = degseqDF, filter = c(Fold = 2, FDR = 10))
## Plot heatMaplot
heatMaplot(exploredds, clust="ind", DEGlist = unique(as.character(unlist(DEG_list[[1]]))))

## Plot PCAplot
PCAplot(exploredds, plotly = TRUE)

## Plot MDSplot
MDSplot(exploredds, plotly = FALSE)

## Plot GLMplot
exploredds_raw <- exploreDDS(countMatrix, targets, cmp=cmp[[1]], preFilter=NULL, transformationMethod="raw")
#> Warning: some variables in design formula are characters, converting to factors
GLMplot(exploredds_raw, plotly = FALSE)

## Plot MAplot
MAplot(exploredds_raw, plotly = TRUE)
#> estimating size factors
#> estimating dispersions
#> gene-wise dispersion estimates
#> mean-dispersion relationship
#> final dispersion estimates
#> fitting model and testing
#> [1] "removing NA from the results"

## Plot tSNEplot
tSNEplot(countMatrix, targets, perplexity = 5)

## Plot volcanoplot
volcanoplot(degseqDF, comparison = "M12-A12", filter = c(Fold = 1, FDR = 20), genes = "ATCG00280")