publish-methods: Methods for publishing a variety of data types in selected...
In ReportingTools: Tools for making reports in various formats

These are a series of methods for taking various data types and coercing them into selected output formats.

Data types can be output in several formats. Exactly what is done to coerce a data type into the given output format is described below.

HTMLReport or HTMLReportRef

data.frame:
publish(object, htmlReport, tableTitle = NULL, ...) The most basic object to publish to an HTMLReport is a data.frame. Most other methods involve coercing their objects to a data.frame and then calling publish on that data.frame. As such, this is where all styling for publishing tables can be centrally controlled. If tableTitle is specified, the title is printed above the table.

MArrayLM:
publish(object, htmlReport, eSet, factor, n = 1000, pvalueCutoff = 0.01, lfc = 0, coef = NULL, adjust.method='BH', make.plots = TRUE, ...) An MArrayLM object is coerced to a data.frame using a method similar to the topTable function from the limma package. The resulting table includes some selected feature data, and the log fold change and adjusted p-value from the linear model. The p-value adjustment method is set using the “adjust.method” argument. A glyph showing expression levels of each gene are also optionally plotted, based on the expression values from the ExpressionSet given in eSet, and the levels set in factor.

DESeqDataSet:
publish(object, publicationType, factor = NULL, n = 1000, pvalueCutoff = 0.01, lfc = 0, contrast = NULL, resultName = NULL, make.plots = TRUE, ..., name) To coerce a DESeqDataSet to a data.frame, we use the results function from the DESeq2 package, therefore DESeq, or something similar must be run prior to publish. contrast and resultName are passed to the results function as the contrast and name parameters, please consult the documentation for results to see how these are specified. If present, annotation.db is used to find gene-level annotations for the rows in the DESeqDataSet. Stripplots showing the expression values for each transcript are produced based on the normalized counts from the DESeqDataSet, grouped by the levels in factor.

DESeqResults:
publish(object, publicationType, DataSet = NULL, annotation.db = NULL, n = 500, pvalueCutoff = 0.01, lfc = 0, make.plots = TRUE, ..., name) To coerce a DESeqResults object to a data.frame, we filter the results DataFrame such that the absolute log fold changes is greater than lfc and the adjusted p-value is less than pvalueCutoff. If present, annotation.db is used to find gene-level annotations for the rows in the DESeqResults. If make.plots is TRUE, stripplots showing the expression values for each transcript are produced based on the normalized counts from the DataSet, which should be of class DESeqDataSet, grouped by the levels in factor.

DGEExact:
publish(object, htmlReport, countTable, conditions, annotation.db = "org.Hs.eg", n = 1000, pvalueCutoff = 0.01, lfc = 0, adjust.method = "BH", make.plots = TRUE, ...) DGEExact objects are coerced to a data.frame using the topTags function from the edgeR package and filtered base on pvalueCutoff and lfc. The resulting table includes feature data, derived from the annotation package defined in annotation.db, along with the fold change and adjusted p-value. For this to work, the feature names in the DGEExact object and the countTable have to be the primary identifier from the annotation packge. In most cases, the primary identifier will be Entrez Gene IDs. If no valid annotation package name is provided in annotation.db, then feature data will come from the “genes” slot in the DGEExact object. The counts stored in countTable should be in units of counts per million provided by the cpm function in the edgeR package. The p-value adjustment method can be set using the “adjust.method” argument.

DGELRT:
publish(object, htmlReport, countTable, conditions, annotation.db = "org.Hs.eg", n = 1000, pvalueCutoff = 0.01, lfc = 0, adjust.method = "BH", make.plots = TRUE, ...) DGELRT objects are coerced to a data.frame using the topTags function from the edgeR package and filtered base on pvalueCutoff and lfc. The resulting table includes feature data, derived from the annotation package defined in annotation.db, along with the fold change and adjusted p-value. For this to work, the feature names in the DGELRT object and the countTable have to be the primary identifier from the annotation packge. In most cases, the primary identifier will be Entrez Gene IDs. If no valid annotation package name is provided in annotation.db, then feature data will come from the “genes” slot in the DGEExact object. The counts stored in countTable should be in units of counts per million provided by the cpm function in the edgeR package. The p-value adjustment method can be set using the “adjust.method” argument.

HyperGResultBase, GOHyperGResult, PFAMHyperGResult:
publish(object, htmlReport, selectedIDs, annotation.db, pvalueCutoff = 0.01,categorySize=10, makePlot=FALSE,...) A HyperGResult object is coerced to a data.frame using a method similar to the summary function from the Category package. The resulting table includes for each classification the id, name, odds ratio of enrichment and p-value from the hypergeometric test. A glyph showing the level of overlap of each classification with the selected genes is also plotted. The number of genes found in this classification is listed in the table and links to another page with a table of the corresponding genes, symbols and names. An additional page listing the overlap genes is also linked to the main output. For a GOHyperGResult object, a plot depicting the relationship between the significant ontologies and their parents is also plotted if makePlot is TRUE. The selectedIDs are the Entrez ids of the genes of interest (i.e. from the gene universe); annotation.db is the species of the ids.

GeneSetCollection:
publish(object, htmlReport, annotation.db, setStats=NULL, setPValues=NULL, geneStats = NULL, ..., .setToHTML = NULL, .setToDF = NULL, .modifySetDF = NULL) A GeneSetCollection object is coerced to a data.frame. If setStats and/or setPValues are provided, they are included in the table of gene sets. The resulting table includes links to additional pages containing the ids, names and symbols of each gene in the corresponding set. To get the appropriate annotations for the individual gene sets, either annotation.db has to be the name of the appropriate annotation package, or the geneIdType of the individual gene sets has to have a non-empty annotation slot. The user can provide custom functions for coercing the enclosed GeneSets to their own preferred HTML representation by providing a function for .setToHTML. Likewise, a user-provided function can be passed in for .setToDF, to control the coercion of individual GeneSets to a data.frame. Additional modifications to the individual GeneSet data.frames can be provided as a list of functions in .modifySetDF.

GeneSet:
publish(object, htmlReport, annotation.db, geneStats = NULL, ...) A GeneSet object is coerced to a data.frame. If geneStats are provided, an extra column of values for each of the geneIds in the GeneSet is appended to the table. The data.frame will optionally include feature annotations, such as Entrez IDs, gene symbols and gene names, if either annotation.db is a valid annotation package name or if the annotation slot of the geneIdType of the GeneSet is a valid annotation package name.

trellis:
publish(object, publicationType, figureTitle = NULL, filename = NULL, png.height = 480, png.width = 480, pdf.height = 7, pdf.width = 7, br = TRUE, ...) Trellis objects, as created by lattice functions are published to an HTMLReport by first printing a pdf and png version of the object, and then including the png image in the HTML page, linking it to the pdf version of the plot.

Special cases:
There are a few special types that can be published to a HTMLReport. Publishing an HTMLReport to another HTMLReport will create a link from one report to the other, using the title of the published report as the link text.

DataPackage

There are two ways to publish R objects to a DataPackage: by using a character vector of object names, or by calling publish on the object directly, providing a name to use in saving the object. For most object types, basic documentation is created for the object in the data package, using the promptData function.

CSVFile

data.frames can be published directly to csv files. Other data types, such as MArrayLM, DGEExact, GOHyperGResult, PFAMHyperGResult can be output as CSVFile, by means of coercion to a data.frame first.

my.df <- data.frame()
## html.report <- HTMLReport(shortName = "my_html_file",
##     reportDirectory = "reportDirectory")
# publish(my.df, html.report)
## finish(html.report)