Script gives wrong result

  • Christin


    I want to run a HCL using the Spearman rank correlation and complete linkage.

    Doing that manually in the session produces one result, but if I create a script to do the same thing I get a different result with quite a different sample clustering.

    It seems that when running a HCL in a script the Euclidean distance metric is always used, no matter what I pick. If I use e.g. an ST in my script instead, I actually get the distance metric I asked for.

    Is this a known bug, or am I doing something wrong? Below I have pasted an example of a script that does not produce the expected result.

    <?xml version="1.0"?>
    <!DOCTYPE TM4ML SYSTEM "C:\Users\christin\.mev\mev_script_dtd.dtd">
    <TM4ML version="1.0">
       <mev version="1.0">
          <!- Original Script Creation Date: Apr 21, 2010 2:15:43 PM ->

          <!- Script Name: Test ->

          <!- Script Description: Test ->

          <primary_data id="1"/>
             <alg_set input_data_ref="1" set_id="1">
                <algorithm alg_id="1" alg_name="HCL" alg_type="cluster" input_data_ref="1">
                      <param key="distance-absolute" value="false"/>
                      <param key="calculate-genes" value="true"/>
                      <param key="calculate-experiments" value="true"/>
                      <param key="method-linkage" value="1"/>
                      <param key="distance-factor" value="1.0"/>
                      <param key="distance-function" value="9"/>
                   <output_data output_class="single-output">
                      <data_node data_node_id="2" name="Single Ordered Output"/>