Description Usage Arguments Details Author(s) See Also
View source: R/export.gsea.rnk.topTable.R
GSEA's Max_probe algorithm
GSEA has 2 modes for collapsing probes to genes: “Max_probe” and
“Median_probe”, both
work well when the input data is a GCT file, consisting of expression levels.
However, I feel pretty strongly that GSEApreRanked needs a 3rd option which
chooses the best probe, based on the one that obtained the largest t-statistic
in either direction, a “Best_probe” if you will. If you provide this
function a probe2gene
map (from probe ids to gene symbols), then we will
automatically export just the best probes, ie 1 row per gene, still using
the probe ID as the identifier (so that we can still use the same chip file).
1 2 |
tt |
a limma |
file |
the output file <****.rnk> |
values |
choose the column name that contains the values. This must be in the colnames of the topTable. If P.Value is chosen, then the -log10 P will be calculated, thus small P-values become large +ve numbers. Default is “t”. |
probe2gene |
a 2 column matrix-like object with mappings from probes 2 genes. Preferably, this will be the chip file used for the GSEA. This allows the identification of the best performing probe per gene, rather than GSEA's "Max_Probe" method which chooses the probe with the larges value which is unsuitable for down-regulated genes where one poorly performing probe could be ~0. Leave as NULL to ignore this. |
convert2symbol |
logical: if |
verbose |
logical: verbose messages |
probe2gene
The justification for probe2gene
is described above. probe2gene
should be a
2 column matrix-like
object with mappings from probes 2 genes. Preferably,
this will be from the same chip file that will be used
for running GSEA.
Mark Cowley, 8/2/08
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.