From f1d708f4d4fde3997a1cadb3a2825a83006885ee Mon Sep 17 00:00:00 2001
From: delangel <delangel@348d0f76-0448-11de-a6fe-93d51630548a>
Date: Thu, 17 Feb 2011 18:57:21 +0000
Subject: [PATCH] Fixes for HRun annotation in case of indels: a) In case of a
 deletion value was completely broken, we'd report 0 or -1. b) For indels, we
 report maximum of forward and backward values - I've seen empirically many
 sites which are not strand biased but which seem to be artifacts and the
 homopolymer run is always to the right only (because we left align by
 convention).

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5260 348d0f76-0448-11de-a6fe-93d51630548a
---
 .../gatk/walkers/annotator/HomopolymerRun.java   | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/java/src/org/broadinstitute/sting/gatk/walkers/annotator/HomopolymerRun.java b/java/src/org/broadinstitute/sting/gatk/walkers/annotator/HomopolymerRun.java
index ab0115160..bf02e816a 100755
--- a/java/src/org/broadinstitute/sting/gatk/walkers/annotator/HomopolymerRun.java
+++ b/java/src/org/broadinstitute/sting/gatk/walkers/annotator/HomopolymerRun.java
@@ -27,7 +27,7 @@ public class HomopolymerRun implements InfoFieldAnnotation, StandardAnnotation {
 
         int run;
         if ( vc.isSNP() ) {
-            run = computeHomopolymerRun(vc.getAlternateAllele(0).getBases()[0], ref);
+            run = computeHomopolymerRun(vc.getAlternateAllele(0).getBases()[0], ref, true);
         } else if ( vc.isIndel() && ANNOTATE_INDELS ) {
             run = computeIndelHomopolymerRun(vc,ref);
         } else {
@@ -45,7 +45,7 @@ public class HomopolymerRun implements InfoFieldAnnotation, StandardAnnotation {
 
     public boolean useZeroQualityReads() { return false; }
 
-    private static int computeHomopolymerRun(byte altAllele, ReferenceContext ref) {
+    private static int computeHomopolymerRun(byte altAllele, ReferenceContext ref, boolean domin) {
 
         // TODO -- this needs to be computed in a more accurate manner
         // We currently look only at direct runs of the alternate allele adjacent to this position
@@ -70,7 +70,11 @@ public class HomopolymerRun implements InfoFieldAnnotation, StandardAnnotation {
             rightRun++;
         }
 
-        return Math.max(leftRun, rightRun);
+        if (domin)
+            return Math.min(leftRun, rightRun);
+        else
+            return Math.max(leftRun, rightRun);
+
      }
 
     private static int computeIndelHomopolymerRun(VariantContext vc, ReferenceContext ref) {
@@ -81,13 +85,13 @@ public class HomopolymerRun implements InfoFieldAnnotation, StandardAnnotation {
         if ( vc.isDeletion() ) {
             // check that deleted bases are the same
             byte dBase = bases[refBasePos];
-            for ( int i = 0; i < vc.getAlternateAllele(0).length(); i ++ ) {
+            for ( int i = 0; i < vc.getReference().length(); i ++ ) {
                 if ( bases[refBasePos+i] != dBase ) {
                     return 0;
                 }
             }
 
-            return computeHomopolymerRun(dBase,ref)-1; // remove the extra match from the base itself
+            return computeHomopolymerRun(dBase, ref, false); // do max in both directions
         } else {
             // check that inserted bases are the same
             byte insBase = vc.getAlternateAllele(0).getBases()[0];
@@ -97,7 +101,7 @@ public class HomopolymerRun implements InfoFieldAnnotation, StandardAnnotation {
                 }
             }
 
-            return computeHomopolymerRun(insBase,ref);
+            return computeHomopolymerRun(insBase,ref, false);
         }
     }
 }
\ No newline at end of file