• Installation

        For small size dataset (up to 5000 genes/probesets/probes) simply use the web application.


        For large size dataset please use the standalone application. Installation for the standalone application:

1. Download and install the pre-requirements:

 > wget 
 > R CMD INSTALL corpcor_1.6.2.tar.gz

 > wget
 > R CMD INSTALL qvalue_1.28.0.tar.gz

 > wget
 > R CMD INSTALL sva_3.1.2.tar.gz

2. Download the application:

 > wget 
 > tar -zxvf SVAw_standalone.tar.gz


  • Data Format

Input files

Expression: (sample)
        An m x n expression matrix with n arrays for m genes/probesets/probes.


Demographic: (sample)
        An m x 2 matrix with m individuals/arrays divided into groups.
        Column1 = Unique Identifier (ID).
        Column2 = 1 (Group1 arrays for e.g. "cases") and 0 (Group2 arrays for e.g. "controls").


Output files (sample)

        A copy of your Expression matrix input.


        A copy of your Demographic matrix input.


        An m x n matrix with m individuals/arrays and n surrogate variables predicted for the data set by SVAw.


        A table of results for regression of gene expression on the primary variable of interest before
        and after correcting for surrogate variables.
        The output includes the following colums (in order) for each gene/probesets/probes:

        1. coefficient_unadjusted = linear regression coefficient from the fitted model.
        2. fold_change_up_or_down_regulated_unadjusted = log2 ratio of mean expression of (Group1
        for e.g. "cases") vs. (Group2 for e.g. "controls") from the fitted model.
        3. p_unadjusted = p-value for the test of signifance for each genes/probesets/probes mean expression
        comparison of (Group1 for e.g. "cases") vs. (Group2 for e.g. "controls") from the fitted model.

        [post-sva] [surrogate variables included as "covariates" in the linear regression model of
        gene expression on the primary variable of interest]
        4. coefficient_adjusted.
        5. fold_change_up_or_down_regulated_adjusted.
        6. p_adjusted.


        A subset of the Probe_Statistics.txt table (above) of results by significance
        (p-value [post-sva] < 0.05). All genes/probesets/probes meeting this criteria are shown in this table.


        Volcano Plot = Plot of significance vs. fold change
        MA Plot


        Volcano Plot
        MA Plot


  • Usage

For small to mid-size dataset simply use the web applicatoin.


For large dataset download the standalone application.
It can be run without any option to view the usage:

  > sh 

|     SVAw v1.0    release: 2/29/2012          |
|  For questions & comments:                   |
|      |

usage: options
 -e  Expression Filename            (required)
 -d  Demographic Filename           (required)
 -v  Surrogate Variable Filename    (optional) default: Surrogate_Variable.txt
 -p  Probe Statistics Filename      (optional) default: Probe_Statistics.txt
 -s  Probe Statistics Significants  (optional) default: Probe_Statistics_Significants.txt
 -u  Unadjusted Graph Filename      (optional) default: Unadjusted_Graph.png
 -a  Adjusted Graph Filename        (optional) default: Adjusted_Graph.png
 -m  SVA Method                     (optional) default: irw
 -o  output folder                  (optional) default: SVAw_output




Running sample data:
  > sh -e svadata.txt -d svadata-Dx.txt 

|     SVAw v1.0    release: 2/29/2012          |
|  For questions & comments:                   |
|      |

Loading required package: corpcor
Loading required package: mgcv
This is mgcv 1.7-2. For overview type 'help("mgcv-package")'.
Loading required package: qvalue
Loading Tcl/Tk interface ... done

Analysis completed! 

Creating output folder: SVAw_output

Generating report ...

open the "SVAw_output/SVAw_report.htm" in your browser to see the SVAw analysis outcome.


download and view the sample output: SVAw_output.tar.gz