In UPSCb/RnaSeqTutorial: RNA-Seq Tutorial

Appendix A

This appendix list the functions used during the biological QA and the DE analysis. The function code is displayed first and then briefly explained.

readAbundance

showMethods(readAbundance,includeDefs = TRUE)

We read in the kallisto abundance files using the tximport library. This returns expression estimates for every transcripts for all samples. Next, from the data, we derived the mapping between transcripts and genes (i.e. which splicing isoforms are encoded a given gene). The P. trichocarpa gene are easy to extract from the transcript names, they simply have an extra dot followed by numbers following the gene identifier. Using this transcript to gene mapping we can summarise the expression estimate at the gene level. An alternative pattern for the file matching can be provided. Also an alternative type for the files can be given. It, however, needs to be one of kallisto (the default) and salmon. Note that there is no rationale other than the alphabetic order in selecting that default.

nonExpressed

showMethods(nonExpressed,includeDefs = TRUE)

This function just identifies the genes that have no expression; i.e. the sum of the expression of these genes across all samples is 0. It then simply calculate the proportion of the total genes and reports these values in a text message.

rawDataMeanPlot

showMethods(rawDataMeanPlot,includeDefs = TRUE)

We create a color palette and then plot the density curve of the average expression of every gene across all samples

rawDataSamplePlot

showMethods(rawDataSamplePlot,includeDefs = TRUE)

The function performs the same plotting as the method above, but instead of plotting the average, every samples is plotted individually on the same plot.

createDESeqDataSet

showMethods(createDESeqDataSet,includeDefs = TRUE)

This function instantiate a DESeqDataSeq object from the count table and the metadata. An alternative design can be provided.

reportSizeFactors

showMethods(reportSizeFactors,includeDefs = TRUE)

This functions estimate the size factors (the effective sequencing depth) and report them as a boxplot.

transform

showMethods(transform,includeDefs = TRUE)

This function performs the Variance Stabilising Transformation (VST) of the count data, not using the prior (the variable(s)) of the model (i.e. in blind mode).

validateVST

showMethods(validateVST,includeDefs = TRUE)

This function validates the VST by plotting the mean-variance relationship.

plotUnTransformed

showMethods(plotUnTransformed,includeDefs = TRUE)

This function plots the log2 of the raw and of the library-size-corrected data.

plotPca

showMethods(plotPca,includeDefs = TRUE)

This function runs a principal component (prcomp) analysis (PCA) on the samples ( hence the count matrix is transposed first). From the PCA results, the percentage of variance explained by each component is retrieved. Then the first two dimensions of the PCA are plotted.

UPSCb/RnaSeqTutorial documentation built on Nov. 24, 2020, 12:40 a.m.