From 80d547ae71627e2f292a1a9c3d2b70f8e7efd76a Mon Sep 17 00:00:00 2001
From: droazen <droazen@348d0f76-0448-11de-a6fe-93d51630548a>
Date: Fri, 8 Apr 2011 18:33:10 +0000
Subject: [PATCH] Fix for bug GSA-445: Sequence dictionary validation can be
 very slow with large numbers of contigs.
 SequenceDictionaryUtils.getCommonContigsByName() was running in O(n^2) time
 due to poor choice of data structure -- modified it to run in O(n) time. Also
 removed an unnecessary O(n log n) step at another stage in the sequence
 dictionary validation process. In tests with a 181,813-entry sequence
 dictionary, runtime improved from an average of 21.4 minutes to 45.1 seconds.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5604 348d0f76-0448-11de-a6fe-93d51630548a
---
 .../broadinstitute/sting/gatk/GenomeAnalysisEngine.java    | 7 +------
 .../sting/utils/SequenceDictionaryUtils.java               | 6 +++---
 2 files changed, 4 insertions(+), 9 deletions(-)

diff --git a/java/src/org/broadinstitute/sting/gatk/GenomeAnalysisEngine.java b/java/src/org/broadinstitute/sting/gatk/GenomeAnalysisEngine.java
index d8f30c5da..7cfbc0762 100755
--- a/java/src/org/broadinstitute/sting/gatk/GenomeAnalysisEngine.java
+++ b/java/src/org/broadinstitute/sting/gatk/GenomeAnalysisEngine.java
@@ -869,12 +869,7 @@ public class GenomeAnalysisEngine {
             // Compile a set of sequence names that exist in the BAM files.
             SAMSequenceDictionary readsDictionary = reads.getHeader().getSequenceDictionary();
 
-            Set<String> readsSequenceNames = new TreeSet<String>();
-            for (SAMSequenceRecord dictionaryEntry : readsDictionary.getSequences())
-                readsSequenceNames.add(dictionaryEntry.getSequenceName());
-
-
-            if (readsSequenceNames.size() == 0) {
+            if (readsDictionary.size() == 0) {
                 logger.info("Reads file is unmapped.  Skipping validation against reference.");
                 return;
             }
diff --git a/java/src/org/broadinstitute/sting/utils/SequenceDictionaryUtils.java b/java/src/org/broadinstitute/sting/utils/SequenceDictionaryUtils.java
index fc28d6c26..ee47a0f2b 100755
--- a/java/src/org/broadinstitute/sting/utils/SequenceDictionaryUtils.java
+++ b/java/src/org/broadinstitute/sting/utils/SequenceDictionaryUtils.java
@@ -353,13 +353,13 @@ public class SequenceDictionaryUtils {
      * @return
      */
     public static Set<String> getCommonContigsByName(SAMSequenceDictionary dict1, SAMSequenceDictionary dict2) {
-        Set<String> intersectingSequenceNames = new HashSet<String>(getContigNames(dict1));
+        Set<String> intersectingSequenceNames = getContigNames(dict1);
         intersectingSequenceNames.retainAll(getContigNames(dict2));
         return intersectingSequenceNames;
     }
 
-    public static List<String> getContigNames(SAMSequenceDictionary dict) {
-        List<String> contigNames = new ArrayList<String>();
+    public static Set<String> getContigNames(SAMSequenceDictionary dict) {
+        Set<String> contigNames = new HashSet<String>((int)(dict.size() / 0.75f) + 1, 0.75f);
         for (SAMSequenceRecord dictionaryEntry : dict.getSequences())
             contigNames.add(dictionaryEntry.getSequenceName());
         return contigNames;