<tool id="hgv_gpass" name="GPASS" version="1.0.0">
  <description>significant single-SNP associations in case-control studies</description>

  <command interpreter="perl">
    gpass.pl ${input1.extra_files_path}/${input1.metadata.base_name}.map ${input1.extra_files_path}/${input1.metadata.base_name}.ped $output $fdr
  </command>

  <inputs>
    <param name="input1" type="data" format="lped" label="Dataset"/>
    <param name="fdr" type="float" value="0.05" label="FDR"/>
  </inputs>

  <outputs>
    <data name="output" format="tabular" />
  </outputs>

  <requirements>
    <requirement type="package">gpass</requirement>
  </requirements>

  <!-- we need to be able to set the seed for the random number generator
  <tests>
    <test>
      <param name='input1' value='gpass_and_beam_input' ftype='lped' >
        <metadata name='base_name' value='gpass_and_beam_input' />
        <composite_data value='gpass_and_beam_input.ped' />
        <composite_data value='gpass_and_beam_input.map' />
        <edit_attributes type='name' value='gpass_and_beam_input' />
      </param>
      <param name="fdr" value="0.05" />
      <output name="output" file="gpass_output.txt" />
    </test>
  </tests>
  -->

  <help>
**Dataset formats**

The input dataset must be in lped_ format, and the output is tabular_.
(`Dataset missing?`_)

.. _lped: ./static/formatHelp.html#lped
.. _tabular: ./static/formatHelp.html#tab
.. _Dataset missing?: ./static/formatHelp.html

-----

**What it does**

GPASS (Genome-wide Poisson Approximation for Statistical Significance)
detects significant single-SNP associations in case-control studies at a user-specified FDR.  Unlike previous methods, this tool can accurately approximate the genome-wide significance and FDR of SNP associations, while adjusting for millions of multiple comparisons, within seconds or minutes.

The program has two main functionalities:

1. Detect significant single-SNP associations at a user-specified false
   discovery rate (FDR).

   *Note*: a "typical" definition of FDR could be
            FDR = E(# of false positive SNPs / # of significant SNPs)

   This definition however is very inappropriate for association mapping, since SNPs are
   highly correlated.  Our FDR is
   defined differently to account for SNP correlations, and thus will obtain
   a proper FDR in terms of "proportion of false positive loci".

2. Approximate the significance of a list of candidate SNPs, adjusting for
   multiple comparisons. If you have isolated a few SNPs of interest and want 
   to know their significance in a GWAS, you can supply the GWAS data and let 
   the program specifically test those SNPs.


*Also note*: the number of SNPs in a study cannot be both too small and at the same
time too clustered in a local region. A few hundreds of SNPs, or tens of SNPs
spread in different regions, will be fine. The sample size cannot be too small
either; around 100 or more individuals (case + control combined) will be fine.
Otherwise use permutation.

-----

**Example**

- input map file::

    1  rs0  0  738547
    1  rs1  0  5597094
    1  rs2  0  9424115
    etc.

- input ped file::

    1 1 0 0 1  1  G G  A A  A A  A A  A A  A G  A A  G G  G G  A A  G G  G G  G G  A A  A A  A G  A A  G G  A G  A G  A A  G G  A A  G G  A A  G G  A G  A A  G G  A A  G G  A A  A G  A G  G G  A G  G G  G G  A A  A G  A A  G G  G G  G G  G G  A G  A A  A A  A A  A A
    1 1 0 0 1  1  G G  A G  G G  A A  A A  A G  A A  G G  G G  G G  A A  G G  A G  A G  G G  G G  A G  G G  A G  A A  G G  A G  G G  A A  G G  G G  A G  A G  G G  A G  A A  A A  G G  G G  A G  A G  G G  A G  A A  A A  A G  G G  A G  G G  A G  G G  G G  A A  G G  A G
    etc.

- output dataset, showing significant SNPs and their p-values and FDR::

    #ID   chr   position   Statistics  adj-Pvalue  FDR
    rs35  chr1  136606952  4.890849    0.991562    0.682138
    rs36  chr1  137748344  4.931934    0.991562    0.795827
    rs44  chr2  14423047   7.712832    0.665086    0.218776
    etc.

-----

**Reference**

Zhang Y, Liu JS. (2010)
Fast and accurate significance approximation for genome-wide association studies.
Submitted.

  </help>
</tool>