Changeset 4318

Timestamp:

May 27, 2008, 10:08:15 AM (15 years ago)

Author:

Jari Häkkinen

Message:

Addresses #639. Exporting plain matrix for 1 and 2 channel data.

File:

• branches/2.7-stable/src/plugins/core/net/sf/basedb/plugins/BioAssaySetExporter.java (modified) (11 diffs)

branches/2.7-stable/src/plugins/core/net/sf/basedb/plugins/BioAssaySetExporter.java

@@ -4 +4 @@
   Copyright (C) 2006 Johan Enell, Jari Hakkinen, Nicklas Nordborg, Martin Svensson
   Copyright (C) 2007 Johan Enell, Jari Hakkinen, Peter Johansson, Nicklas Nordborg, Martin Svensson
+  Copyright (C) 2008 Jari Hakkinen, Nicklas Nordborg
 
   This file is part of BASE - BioArray Software Environment.
@@ -105 +106 @@
      use in MeV (available from http://www.tm4.org).
    - BASEfile.
-   - Plain Matrix
+   - Plain Matrix, see documentation for exportPlainMatrix method for
+     a description of the format.
 
    @author enell, jari, peter
@@ -602 +604 @@
   
   /**
-    Export bioassayset data in MeV format. The export includes
-    all reporter information and all experimental factor information. If the
-    expriment is a two-channel experiment the log ratio is exported as the
-    expression value otherwise the raw intensity of channel 1 is exported.
+    Export bioassayset data in MeV format. The export includes all
+    reporter information and all experimental factor information. If
+    the expriment is a two-channel experiment the ch1/ch2 ratio
+    (non-log value!) is exported as the expression value otherwise the
+    raw intensity of channel 1 is exported.
     
     @param bioassayset The bioassayset to export
@@ -708 +711 @@
     int position=-1;
     short column=nof_bioassays;
-    double log2=Math.log(2.0);
     int nof_reporters=bioassayset.getNumReporters();
     int progress_report_interval=nof_reporters/10;
@@ -771 +773 @@
 
 
+  /**
+    Export bioassayset data in plain matrix format.
+
+    Plain matrix format is simply a matrix with no header nor reporter
+    information. Each column in the matrix corresponds to an assay,
+    and each row corresponds to a reporter.
+
+    Missing values are reported as entries with no number and 'NaN' is
+    reported for undefined values (such as division with zero). Note,
+    reporters with no measurement within a bioassay set are not
+    reported at all, i.e., a row is not exported with misssing values
+    only. To be clear, we state this once more: Rows (reporters) are
+    only exported if one or more columns (assays) have a value. The
+    value can be 'NaN'.
+
+    The data exported depends on the underlying microarray platform;
+     - 1-channel: raw ch1 intensity
+     - 2-channel: the fraction ch1/ch2, note! non-log value
+    The justification for exporting non-log values for 2-channel data
+    is to follow the MeV export format where non-log data is exported
+    for 2-channel data. 'NaN' is returned for zero ch2 values.
+
+    The plug-in or user that uses this method must keep track of the
+    row and column information themselves. The row is sorted in
+    ascending position order, the columns are sorted in ascending
+    column order.
+
+    @param bioassayset The bioassayset to export
+    @param os The stream to write the exported data to
+    @param progress An optional progress reporter
+    @throws IOException If there is an IO error
+    @throws SQLException If there is an error reading from the database
+  */
   public void exportPlainMatrix(BioAssaySet bioassayset, OutputStream os,
                                 ProgressReporter progress)
@@ -778 +813 @@
     DbControl dc=bioassayset.getDbControl();
     PrintWriter out = new PrintWriter(os);
-    List<ExtendedProperty> reporterProperties =
-      ExtendedProperties.getProperties("ReporterData");
 
     int nof_bioassays=0;
     nof_bioassays=bioassayset.getBioAssays().list(dc).size();
+    int nof_channels=bioassayset.getRawDataType().getChannels();
 
     if (progress != null)
@@ -792 +826 @@
     query.select(Dynamic.select(VirtualColumn.POSITION));
     query.select(Dynamic.select(VirtualColumn.COLUMN));
-    query.select(Dynamic.select(VirtualColumn.channel(1)));
-    
+    for (int i=1; i<=nof_channels; ++i)
+    {
+      query.select(Dynamic.select(VirtualColumn.channel(i)));
+    }
+
     query.order(Orders.asc(Dynamic.column(VirtualColumn.POSITION)));
     query.order(Orders.asc(Dynamic.column(VirtualColumn.COLUMN)));
@@ -801 +838 @@
     int column_position=spotData.getIndex(VirtualColumn.POSITION.getName());
     int column_column=spotData.getIndex(VirtualColumn.COLUMN.getName());
-    int column_channel=spotData.getIndex(VirtualColumn.channel(1).getName());
-    
+    int[] column_channel=new int[nof_channels];
+    for (int i=0; i<nof_channels; ++i)
+    {
+      column_channel[i]=spotData.getIndex(VirtualColumn.channel(i+1).getName());
+    }
+
     int position=-1;
     int column=nof_bioassays;
-    //double log2=Math.log(2.0);
     int nof_spots=bioassayset.getNumSpots();
     int progress_spot_interval=nof_spots/10;
@@ -848 +888 @@
 
       // Calculate expression value
-      String ch1=item.getString(column_channel);
-      out.write(ch1);
-    }
-    out.flush();
+      double exValue=item.getFloat(column_channel[0]);
+      boolean goodvalue=true;
+      if (nof_channels==2)
+      {
+        double ch2=item.getFloat(column_channel[1]);
+        if (ch2!=0)
+        {
+          exValue/=ch2;
+        }
+        else
+        {
+          goodvalue=false;
+        }
+      }
+
+      out.write( "\t" + (goodvalue ? exValue : "NaN"));
+    }
+    out.close();
   }
 
@@ -1095 +1149 @@
       formats.add(FORMAT_MEV, "MultiExperiment Viewer (MeV)");
       formats.add(FORMAT_BASEFILE, "BASEfile");
-//      formats.add(FORMAT_PLAIN_MATRIX, "Plain Matrix");
+      /**
+         Plain matrix export is currently not supported from the web
+         interface. Simply move the code line below outside this
+         comment block to enable plain matrix export in the web
+         interface. If the export is enabled the About information
+         should be updated to reflect the change.
+
+         formats.add(FORMAT_PLAIN_MATRIX, "Plain Matrix");
+      */
       formatParameter = new PluginParameter<String>
         ( "fileformat",
@@ -1292 +1354 @@
       "- Tab Delimited, Multiple Sample Files (TDMS Format) for " +
       "use in MeV (available from http://www.tm4.org).\n"+
-      "- BASEfile",
+      "- BASEfile\n",
       "2.0",
       "2006, Base 2 development team",