runRomer: A function to run the romer function on a set of contrasts.

In affycoretools: Functions useful for those doing repetitive analyses with Affymetrix GeneChips

Description

This function automates both running romer on a set of contrasts as well as the creation of output HTML tables that can be used to explore the results. The basic idea here is that one might have used limma to fit a model and compute some contrasts, and then want to do a GSEA using romer.

Usage

runRomer(object, ...)

## S4 method for signature 'ExpressionSet'
runRomer(
  object,
  fit,
  setloc,
  annot = NULL,
  design = NULL,
  contrast = NULL,
  wts = NULL,
  save = TRUE,
  baseline.hmap = TRUE,
  affy = TRUE,
  ...
)

## S4 method for signature 'DGEList'
runRomer(
  object,
  fit,
  setloc,
  design = NULL,
  contrast = NULL,
  save = TRUE,
  baseline.hmap = TRUE,
  ...
)

## S4 method for signature 'EList'
runRomer(
  object,
  fit,
  setloc,
  design = NULL,
  contrast = NULL,
  save = TRUE,
  baseline.hmap = TRUE,
  ...
)

Arguments

object	An ExpressionSet, DGEList, or EList object
...	Used to pass arguments to lower-level functions. See outputRomer geneSetPage, dataAndHeatmapPage and gsHeatmap for available arguments.
fit	A fitted model from either limma (e.g., MArrayLM) or edgeR (e.g., DGEGLM)
setloc	A character vector giving the path for gene set RData files (see description for more information), or a named list (or list of lists), where the top-level names consist of gene set grouping names (like KeGG or GO), the next level names consist of gene set names (like NAKAMURA_CANCER_MICROENVIRONMENT_UP), and the list items themselves are gene symbols, matching the expected capitalization for the species being used (e.g., for human, they are ALL CAPS. For most other species only the First Letter Is Capitalized).
annot	Character. The name of the array annotation package. If NULL, the annotation data will be extracted from the fData slot (for ExpressionSets) or the genes list (for DGEList or EList objects).
design	A design matrix describing the model fit to the data. Ideally this should be a cell-means model (e.g., no intercept term), as the design and contrast matrices are used to infer which data to include in the output heatmaps. There is no guarantee that this will work correctly with a treatment-contrasts parameterization (e.g., a model with an intercept).
contrast	A contrast matrix describing the contrasts that were computed from the data. This contrast should have colnames, which will be used to create parts of the resulting directory structure.
wts	Optional weights vector - if array weights were used to fit the model, they should be supplied here as well.
save	Boolean. If true, after running the romer step, the results will be saved in a file 'romer.Rdata', which can be used as input for outputRomer to create HTML tables. Since romer can take a long time to run, it is advantageous to keep the default.
baseline.hmap	Boolean. If TRUE, then the resulting heatmaps will be centered by subtracting the mean of the baseline sample. As an example, in a contrast of treatment A - treatment B, the mean of the treatment B samples will be subtracted. The heatmap colors then represent the fold change between the A and B samples.
affy	Boolean; are these Affymetrix arrays? If TRUE, the output tables will contain links to the netaffx site.

Details

The romer expects as input a list or lists of gene symbols that represent individual gene sets. One example is the various gene sets from the Broad Institute that are available at http://bioinf.wehi.edu.au/software/MSigDB/, which are distributed as RData files. The default assumption for this function is that the end user will have downloaded these files, and the setloc argument simply tells runRomer where to find them.

Alternatively, user-based gene sets could be created (these should consist of lists of character vectors of gene symbols - see one of the Broad gene sets for an example).

This function will run romer using all the gene sets in the referenced directory, on all the contrasts supplied, and then output the results in a (default) 'genesets' subdirectory. There will be an HTML file in the working directory with a (default) filename of 'indexRomer.html' that will point to individual HTML files in the genesets subdirectory, which will point to individual files in subdirectories within the genesets subdirectory (named after the colnames of the contrast matrix).

Value

If save is TRUE, return a list that can be re-processed using outputRomer. this is useful in cases where you might need to re-run multiple times.

Nothing is returned. This function is called only for the side-effects of creating output HTML files in the working and sub-directories.

Methods (by class)

• ExpressionSet: Perform gene set analysis using microarray data.

• DGEList: Perform gene set analysis using RNA-Seq data processed using edgeR.

• EList: Perform gene set analysis using RNA-Seq data processed using voom.

Author(s)

James W. MacDonald <[e-mail]>

James W. MacDonald <[e-mail]>

affycoretools documentation built on Nov. 8, 2020, 6 p.m.