hyperG2annaffy: HTML tables from GOIDs

Description

Output HTML tables containing the 'enriched' genes for each GO term resulting from a call to hyperGtable.

Usage

hyperG2annaffy(probids, lib, eset, fit = NULL, subset = NULL, comp = 1, type = "MF", pvalue = 0.05, min.count = 10)

Arguments

probids

A vector of Affymetrix probe IDs

lib

An annotation package (e.g., hgu95av2)

eset

An ExpressionSet

fit

An lmFit object. Only necessary if statistics are desired in the resulting table. Defaults to NULL.

subset

A numeric vector used to select GO terms to output (see description for more information). Defaults to NULL

comp

Which contrast/parameter estimate should be used to extract the relevant statistics? Only used if fit is not NULL. See description for more information.

type

One of "MF", "CC", "BP", indicating molecular function, cellular component, or biological process, respectively.

pvalue

The significance level used to choose GO terms

min.count

The minimum number of a given GO term that must be on the chip in order to choose that GO term. This protects against very low p-values that result from the situation where there are very few genes with a given GO term on the chip, but one or two are found in the set of significant genes.

Value

This function is used only for the side effect of creating HTML tables.

Details

This function is used to create HTML tables based on the output of hyperGtable. The basic idea is as follows; as part of an analysis, say hyperGtable was used to create a table of 'enriched' GO terms. Unfortunately, the table only lists GO terms and the number of probesets that are annotated to those GO terms, and the client may be interested in knowing what probesets are enriched for each (or some) GO term.

The default behaviour is to output an HTML table for each GO term, containing the probesets that are annotated at that term (and that are in the set of significant genes). If only some of the GO terms are of interest, one may use the subset argument to select only particular rows. In addition, if the relevant t-statistics, p-values and fold changes are of interest, one can also use the fit argument to point to an lmFit object that contains these data, as well as the comp argument to indicate which parameter or contrast to use. Note that the comp argument defaults to 1, so the first parameter or contrast will be extracted by default.