This tool can process a gene expression matrix (in GCT or TXT format) ranked list (RNK format) and:

Converting and collapsing can be done either individually or both at the same time.

In case you are collapsing a ranked list (RNK format) to perform a "preRanked GSEA" that you later on want to analyze with EnrichmentMap and want to see an expression heatmap for the genesets, you need to generate an expression matrix that contains the expression values from the same probesets that were chosen to represent the gene in the ranked list. This can by selecting the ranked List (RNK) as the primary input file (-i) and the expression Matrix (GCT or TXT) as additional input Expression-table (-e). When using the GUI this can be done by selecting the mode "Ranked List with Expression Matrix".

In this use-case ID-conversion and collapsing have to be done in the same step. The DESCRIPTION column of the collapsed expression matrix will for every given gene then contain the Probeset-ID of the Probeset with the highest absolute Score in the RNK file and in brackets followed by a list of Probeset-IDs that where omitted due to lower absolute rank-scores.

The option 'Suppress gene "NULL"' (--null) will drop all Probeset ID's assigned to the Gene Symbol NULL, as this is used for probesets that are not linked to any Gene in several Chip-Annotation files available from the Broad Institute's FTP server. (These will be dropped by GSEA anyway)



Supported Operating Systems:

GUI Mode

Screenshot of GUI now has a Tk-based Graphical User Interface (GUI). To use the GUI, just start the program without any arguments. This can be done:

After starting the GUI:

Command Line Mode

If you are familiar with command line tools under Unix/Linux, -h gives you all the information you need (if not, see below):

$ -h
Usage: [options] -i input.gct -o output.gct [-c platform.chip] [--collapse]

This tool can process a gene expression matrix (in GCT or TXT format) or
ranked list (RNK format) and either replace the Identifier based on a Chip
Annotation file (e.g. AffyID -> Gene Symbol), or collapse the expression
values or rank-scores for Genes from more than one probe set. Both can be done
in one step by using both '-c platform.chip' and '--collapse' at the same
time. If a ranked list is to be collapsed, an additional expression matrix can
be supplied by the -e/-x parameters and will be filtered to contain the same
probe-sets as selected from the RNK file. If however the file supplied by -i
is not recognized as a RNK file, these options have no effect.  For detailed
descriptions of the file formats, please refer to:
Call without any parameters to select the files and options with a GUI
(Graphical User Interface)

  --version             show program's version number and exit
  -h, --help            show this help message and exit
  -i FILE, --input=FILE
                        input expression table or ranked list
  -o FILE, --output=FILE
                        output expression table or ranked list
  -c FILE, --chip=FILE  Chip File This implies that the Identifiers are to be
  -e FILE, --ei=FILE    (optional) additional input Expression-table, to be
                        restricted to the same probe-sets as the RNK file
  -x FILE, --xo=FILE    (optional) corresponding output file for -i/--ei
  --collapse            Collapse multiple probe sets for the same gene symbol
  --no-collapse         Don't collapse multiple probesets [default]
  --null                suppress Gene with Symbol NULL
  -g, --gui             Open a Window to choose the files and options.
  -q, --quiet           be quiet

On MacOS and Linux you need to make the program executable. Therefore:

On Windows:

Software/EnrichmentMap/CollapseExpressionMatrix (last edited 2010-04-21 16:37:06 by OliverStueker)

MoinMoin Appliance - Powered by TurnKey Linux