Changeset 3887

Timestamp:

Oct 29, 2007, 9:45:30 AM (15 years ago)

Author:

Nicklas Nordborg

Message:

References #797: Enhance performance for LOWESS and Median-ratio plug-ins

Must sort by 'A' or 'ratio' in memory instead of in the database to avoid interference with the
sorting on 'block' when blockgroupsize is > 1

Location:

trunk/src/plugins/core/net/sf/basedb/plugins

Files:

• LowessNormalization.java (modified) (5 diffs)
• MedianRatioNormalization.java (modified) (6 diffs)

trunk/src/plugins/core/net/sf/basedb/plugins/LowessNormalization.java

@@ -333 +333 @@
     Expression ch1 = Dynamic.column(VirtualColumn.channel(1));
     Expression ch2 = Dynamic.column(VirtualColumn.channel(2));
-    // A = log10(sqrt(ch1 * ch2))
-    Expression pseudoA = Expressions.multiply(ch1, ch2);
-    Expression A = Expressions.log10(Expressions.sqrt(pseudoA));
 
     // Create restriction: ch1 > 0 and ch2 > 0
@@ -364 +361 @@
     query.restrict(bioAssayRestriction);
     query.order(Orders.asc(block));
-    query.order(Orders.asc(pseudoA));
       
     // Normalize one bioassay at a time
@@ -424 +420 @@
           if (toIndex > fromIndex)
           {
-            List<Double> smoothCurve = lowess(data.subList(fromIndex, toIndex), fitFraction, iterations, delta);
+            List <SpotData> toNormalize = data.subList(fromIndex, toIndex);
+            Collections.sort(toNormalize);
+            List<Double> smoothCurve = lowess(toNormalize, fitFraction, iterations, delta);
             for (int j = 0; j < smoothCurve.size(); ++j)
             {
-              SpotData spot = data.get(fromIndex + j);
+              SpotData spot = toNormalize.get(j);
               double factor = Math.exp(smoothCurve.get(j) * 0.5);
               batcher.insert(bioassayColumn, spot.position, (float)(spot.ch1/factor), (float)(spot.ch2*factor));
@@ -637 +635 @@
 
   private static class SpotData
+    implements Comparable<SpotData>
   {
     final int position;
@@ -655 +654 @@
       this.A = Math.log10(Math.sqrt(ch1 * ch2));
     }
+
+    /**
+      Sort by A in ascending order.
+    */
+    public int compareTo(SpotData o)
+    {
+      return this.A < o.A ? -1 : 1;
+    }
+    
   }
 

trunk/src/plugins/core/net/sf/basedb/plugins/MedianRatioNormalization.java

@@ -30 +30 @@
 import java.util.Arrays;
 import java.util.Collection;
+import java.util.Collections;
 import java.util.EnumSet;
 import java.util.HashSet;
@@ -329 +330 @@
     Expression ch1 = Dynamic.column(VirtualColumn.channel(1));
     Expression ch2 = Dynamic.column(VirtualColumn.channel(2));
-    // ratio = ch1 / ch2
-    Expression ratio = Expressions.divide(ch1, ch2);
     
     //   Create restriction: ch1 > minIntensity and ch2 > minIntensity
@@ -360 +359 @@
     query.restrict(bioAssayRestriction);
     query.order(Orders.asc(block));
-    query.order(Orders.asc(ratio));
 
     // Normalize one bioassay at a time
@@ -425 +423 @@
             if (highIndex > lowIndex)
             {
+              List <SpotData> toNormalize = data.subList(fromIndex, toIndex);
+              Collections.sort(toNormalize);
               // Find median ratio in the current block range
-              double median = median(data.subList(lowIndex, highIndex));
+              double median = median(toNormalize);
               
               // Correct all spot intensities
@@ -509 +509 @@
   
   private static class SpotData
+    implements Comparable<SpotData>
   {
     final int position;
@@ -524 +525 @@
       this.ratio = ch1 / ch2;
     }
+    /**
+      Sort by ratio in ascending order.
+    */
+    public int compareTo(SpotData o)
+    {
+      return this.ratio < o.ratio ? -1 : 1;
+    }
   }
 }