collapse.topTable: Take a topTable with one row per probeset, and collapse it to...
In drmjc/microarrays: A collection of microarray and genomic helper code.

Description Usage Arguments Value Author(s)

Probes either map to Gene Symbols, or not & some Genes can have multiple probes mapping to them (due to array design, or isoforms). This method will choose the best probe per gene, by selecting that with the largest abs(t.stat), since that probe 'performed' the best, (can be either up or down-regulated). In my experience, for those genes with lots of probes, many do not hyb well (eg on the HGU133+2 array, due to poor probe design), and one will hyb better than the others & thus produce larger +/- t.stat. This is the best peforming probe & should be selected to represent that gene. Other alternatives include median probe, max.avg, ... (not yet implemented.) Probes that don't map to symbols can either be filtered out (only.genes=TRUE), or left in (only.genes=FALSE). Probes that don't map to gene symbols are identified via the probe2gene table & either have NA, NULL, or symbols.ignore.list in the 2nd column of the probe2gene. You can add custom values to the symbols.ignore.list, eg c("N/A", "missing", "blank", "neg.con", ...)

1
2
3

  collapse.topTable(tt, probe2gene, toupper = FALSE,
    symbols.ignore.list = c("---", ""), only.genes = FALSE,
    verbose = TRUE)

`tt`	a toptable, or a list of toptables
`probe2gene`	a 2 column data.frame, 1 row per probe, mapping probe ID's (column 1) to gene symbols (column 2)
`toupper`	convert the gene symbol in column 2 to UPPER case?
`symbols.ignore.list`	a vector of symbols that are used to identify probes with no gene. NB: the special values NA and NULL are always used, so no need to specify these.
`only.genes`	if TRUE, then filter out the probes without a gene symbol. if FALSE, then probes with or without gene symbols are exported.
`verbose`	Helpful messages?

a data.frame like tt, but with some new columns: "Gene.Symbol" in pos 2, "Nprobes" = the number of possible probes for the given gene, "Possible.Probes" = a comma separated character(1) of the possible probes for the given gene. nrows = number of unique genes + if(only.genes=FALSE) the number of probes with no gene. row ordering is given by the original tt's row ordering.

Mark Cowley, 2011-02-22

drmjc/microarrays documentation built on May 15, 2019, 2:26 p.m.