<tool id="align2database" name="align-to-database">
  <description> features </description>
  <command interpreter="python"> align2database.py $query $database $output_coverage $output_standarderror $output_plot $minfeat $windowsize $anchor $span> $outlog </command>
  <inputs>
    <param name="query" type="data" format="interval" label="Query intervals" help= "keep it small (less than 1,000,000 lines)"/>
    <param name="database" type="select" label="Feature database">
     <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/mm9/feature_database" selected="true">All mm9 features (over 200)</option>
     <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/mm9/annotation">Annotated mm9 features</option>   
     <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/mm9/CLIP">protein bound RNA (CLIP) mm9 </option>
     <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/mm9/conservedmiRNAseedsite">conserved miRNA target sites mm9 </option>
     <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/hg18/all-feature">Human ChIP hmChIP database hg18</option>
      <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/hg18/gene-feature">Human gene features hg18</option>
       <option value="/Users/xuebing/galaxy-dist/tool-data/aligndb/hg19/conservedmiRNAseedsite">conserved miRNA target sites hg19 </option>
    </param>
    <param name="anchor" label="Anchor to query features" help="default anchoring to database featuers" type="boolean" truevalue="query" falsevalue="database" checked="False"/>
        <param name="windowsize" size="10" type="integer" value="5000" label="Window size (-w)"  help="will create new intervals of w bp flanking the original center. set to 0 will not change input interval size)"/>
    <param name="minfeat" size="10" type="integer" value="100" label="Minimum number of query intervals hits" help="database features overlapping with too few query intervals are discarded"/>
        <param name="span" size="10" type="float" value="0.1" label="loess span: smoothing parameter" help="value less then 0.1 disables smoothing"/>
    <param name="outputlabel" size="80" type="text" label="Output label" value="test"/>
   
</inputs>
  <outputs>
      <data format="txt" name="outlog" label="${outputlabel} (log)"/> 
    <data format="tabular" name="output_standarderror" label="${outputlabel} (standard error)"/> 
    <data format="tabular" name="output_coverage" label="${outputlabel} (coverage)"/> 
    <data format="pdf" name="output_plot" label="${outputlabel} (plot)"/> 
  </outputs>
  <help>

**Example output**

.. image:: ./static/operation_icons/align_multiple2.png


**What it does**

This tool aligns a query interval set (such as ChIP peaks) to a database of features (such as other ChIP peaks or TSS/splice sites), calculates and plots the relative distance of database features to the query intervals. Currently two databases are available:  

-- **ChIP peaks** from 191 ChIP experiments (processed from hmChIP database, see individual peak/BED files in **Shared Data**)

-- **Annotated gene features**, such as: TSS, TES, 5'ss, 3'ss, CDS start and end, miRNA seed matches, enhancers, CpG island, microsatellite, small RNA, poly A sites (3P-seq-tags), miRNA genes, and tRNA genes. 

Two output files are generated. One is the coverage/profile for each feature in the database that has a minimum overlap with the query set. The first two columns are feature name and the total number of overlapping intervals from the query. Column 3 to column 102 are coverage at each bin. The other file is an PDF file plotting both the heatmap for all features and the average coverage for each individual database feature.


**How it works**

For each interval/peak in the query file, a window (default 10,000bp) is created around the center of the interval and is divided into 100 bins. For each database feature set (such as Pol II peaks), the tool counts how many intervals in the database feature file overlap with each bin. The count is then averaged over all query intervals that have at least one hit in at least one bin. Overall the plotted 'average coverage' represnts the fraction of query features (only those with hits, number shown in individual plot title) that has database feature interval covering that bin. The extreme is when the database feature is the same as the query, then every query interval is covered at the center, the average coverage of the center bin will be 1.    

The heatmap is scaled for each row before clustering.

  </help>
</tool>