Moving VQSR v2 to core.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5740 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/ApplyRecalibration.java

@@ -23,7 +23,7 @@
  * THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broad.tribble.util.variantcontext.VariantContext;
 import org.broad.tribble.vcf.*;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/GaussianMixtureModel.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import Jama.Matrix;
 import cern.jet.random.Normal;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/MultivariateGaussian.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import Jama.Matrix;
 import org.apache.commons.math.special.Gamma;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/TrainingSet.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.apache.log4j.Logger;
 import org.broadinstitute.sting.commandline.Tags;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/Tranche.java

@@ -23,7 +23,7 @@
  * THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broadinstitute.sting.utils.exceptions.UserException;
 import org.broadinstitute.sting.utils.exceptions.ReviewedStingException;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/TrancheManager.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.apache.log4j.Logger;
 import org.broadinstitute.sting.utils.exceptions.UserException;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VQSRCalibrationCurve.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.apache.log4j.Logger;
 import org.broad.tribble.util.variantcontext.VariantContext;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantDataManager.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import cern.jet.random.Normal;
 import org.apache.log4j.Logger;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantDatum.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broadinstitute.sting.utils.GenomeLoc;
 

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantRecalibrator.java

@@ -23,7 +23,7 @@
  * THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broad.tribble.util.variantcontext.VariantContext;
 import org.broadinstitute.sting.commandline.Argument;

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantRecalibratorArgumentCollection.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broadinstitute.sting.commandline.Argument;
 

java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantRecalibratorEngine.java

@@ -1,4 +1,4 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
+package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.apache.log4j.Logger;
 

java/test/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantRecalibrationWalkersIntegrationTest.java

@@ -1,12 +1,10 @@
 package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broadinstitute.sting.WalkerTest;
-import org.broadinstitute.sting.utils.exceptions.UserException;
 import org.testng.annotations.Test;
 import org.testng.annotations.DataProvider;
 
 import java.util.*;
-import java.io.File;
 
 public class VariantRecalibrationWalkersIntegrationTest extends WalkerTest {
     static HashMap<String, String> clusterFiles = new HashMap<String, String>();
@@ -15,111 +13,61 @@ public class VariantRecalibrationWalkersIntegrationTest extends WalkerTest {
 
     private static class VRTest {
         String inVCF;
-        String clusterMD5;
         String tranchesMD5;
-        String recalVCFMD5;
+        String recalMD5;
         String cutVCFMD5;
-        public VRTest(String inVCF, String clusterMD5, String tranchesMD5, String recalVCFMD5, String cutVCFMD5) {
+        public VRTest(String inVCF, String tranchesMD5, String recalMD5, String cutVCFMD5) {
             this.inVCF = validationDataLocation + inVCF;
-            this.clusterMD5 = clusterMD5;
             this.tranchesMD5 = tranchesMD5;
-            this.recalVCFMD5 = recalVCFMD5;
+            this.recalMD5 = recalMD5;
             this.cutVCFMD5 = cutVCFMD5;
         }
     }
 
-    VRTest yriTrio = new VRTest("yri.trio.gatk_glftrio.intersection.annotated.filtered.chr1.vcf",
+    VRTest lowPass = new VRTest("phase1.projectConsensus.chr20.raw.snps.vcf",
-            "f65c27ee40053adc72dd0bfbb628e4d7",  // cluster file
+            "920b12d7765eb4f6f4a1bab045679b31",  // tranches
-            "dce581b880ffb6ea39cbada1ecc95915",  // tranches
+            "41bbc5f07c8a9573d5bb638f01808bba",  // recal file
-            "c3e8a2f43656eab7d847dbf850f844a6",  // recalVCF
+            "d9aa7a0f8fb886df4394f1636605adca"); // cut VCF
-            "50f752a72643db9ad0aa94b3fc4e23d6"); // cut VCF
-
-    VRTest lowPass = new VRTest("lowpass.N3.chr1.raw.vcf",
-            "bda8f17cfc19d23e7e51f99e547f4b3d",  // cluster file
-            "66edae83c50f4e8601fef7fafba774af",  // tranches
-            "0123537e373657386068a534c0f5c91b",  // recalVCF
-            "2172368e8585841e5ad96c95d0827c4b"); // cut VCF
 
     @DataProvider(name = "VRTest")
     public Object[][] createData1() {
-        return new Object[][]{ {yriTrio}, {lowPass} };
+        return new Object[][]{ {lowPass} };
+        //return new Object[][]{ {yriTrio}, {lowPass} }; // Add hg19 chr20 trio calls here
     }
 
-    @Test(dataProvider = "VRTest", enabled = true)
+    @Test(dataProvider = "VRTest")
-    public void testGenerateVariantClusters(VRTest params) {
-        WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b36KGReference +
-                        " -NO_HEADER" +
-                        " --DBSNP " + GATKDataLocation + "dbsnp_129_b36.rod" +
-                        " -B:hapmap,VCF " + comparisonDataLocation + "Validated/HapMap/3.2/sites_r27_nr.b36_fwd.vcf" +
-                        " -weightDBSNP 1.0 -weightHapMap 1.0" +
-                        " -T GenerateVariantClusters" +
-                        " -B:input,VCF " + params.inVCF +
-                        " -L 1:50,000,000-200,000,000" +
-                        " -qual 50.0" +
-                        " --ignore_filter GATK_STANDARD" +
-                        " -an QD -an MQ -an SB" +
-                        " -clusterFile %s",
-                Arrays.asList(params.clusterMD5));
-        executeTest("testGenerateVariantClusters-"+params.inVCF, spec).getFirst();
-    }
-
-    @Test(dataProvider = "VRTest",dependsOnMethods="testGenerateVariantClusters")
     public void testVariantRecalibrator(VRTest params) {
         //System.out.printf("PARAMS FOR %s is %s%n", vcf, clusterFile);
         WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b36KGReference +
+                "-R " + b37KGReference +
-                        " -NO_HEADER" +
+                        " -B:dbsnp,VCF,known=true,training=false,truth=false,prior=10.0 " + GATKDataLocation + "dbsnp_132_b37.leftAligned.vcf" +
-                        " --DBSNP " + GATKDataLocation + "dbsnp_129_b36.rod" +
+                        " -B:hapmap,VCF,known=false,training=true,truth=true,prior=15.0 " + comparisonDataLocation + "Validated/HapMap/3.3/sites_r27_nr.b37_fwd.vcf" +
-                        " -B:hapmap,VCF " + comparisonDataLocation + "Validated/HapMap/3.2/sites_r27_nr.b36_fwd.vcf" +
+                        " -B:omni,VCF,known=false,training=true,truth=true,prior=12.0 " + comparisonDataLocation + "Validated/Omni2.5_chip/Omni25_sites_1525_samples.b37.vcf" +
-                        " -B:truthHapmap,VCF " + comparisonDataLocation + "Validated/HapMap/3.2/sites_r27_nr.b36_fwd.vcf" +
                         " -T VariantRecalibrator" +
                         " -B:input,VCF " + params.inVCF +
-                        " -L 1:20,000,000-100,000,000" +
+                        " -L 20:1,000,000-40,000,000" +
-                        " --ignore_filter GATK_STANDARD" +
+                        " -an QD -an HaplotypeScore -an HRun" +
-                        " --ignore_filter HARD_TO_VALIDATE" +
+                        " -percentBad 0.07" +
-                        " -clusterFile " + getFileForMD5(params.clusterMD5) +
+                        " --trustAllPolymorphic" + // for speed
-                        " -sm TRUTH_SENSITIVITY" +
+                        " -recalFile %s" +
-                        " -o %s" +
                         " -tranchesFile %s",
-                Arrays.asList(params.recalVCFMD5, params.tranchesMD5));
+                Arrays.asList(params.recalMD5, params.tranchesMD5));
         executeTest("testVariantRecalibrator-"+params.inVCF, spec).getFirst();
     }
 
     @Test(dataProvider = "VRTest",dependsOnMethods="testVariantRecalibrator")
-    public void testApplyVariantCuts(VRTest params) {
+    public void testApplyRecalibration(VRTest params) {
         WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b36KGReference +
+                "-R " + b37KGReference +
+                        " -T ApplyRecalibration" +
+                        " -L 20:12,000,000-30,000,000" +
                         " -NO_HEADER" +
-                        " --DBSNP " + GATKDataLocation + "dbsnp_129_b36.rod" +
+                        " -B:input,VCF " + params.inVCF +
-                        " -T ApplyVariantCuts" +
-                        " -L 1:20,000,000-100,000,000" +
-                        " -B:input,VCF " + getFileForMD5(params.recalVCFMD5) +
                         " -o %s" +
-                        " -tranchesFile " + getFileForMD5(params.tranchesMD5),
+                        " -tranchesFile " + getFileForMD5(params.tranchesMD5) +
+                        " -recalFile " + getFileForMD5(params.recalMD5),
                 Arrays.asList(params.cutVCFMD5));
-        executeTest("testApplyVariantCuts-"+params.inVCF, spec);
+        executeTest("testApplyRecalibration-"+params.inVCF, spec);
-    }
-
-
-    @Test()
-    public void testFailWithBadAnnotation() {
-        WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b36KGReference +
-                        " -NO_HEADER" +
-                        " --DBSNP " + GATKDataLocation + "dbsnp_129_b36.rod" +
-                        " -B:hapmap,VCF " + comparisonDataLocation + "Validated/HapMap/3.2/sites_r27_nr.b36_fwd.vcf" +
-                        " -weightDBSNP 0.2 -weightHapMap 1.0" +
-                        " -T GenerateVariantClusters" +
-                        " -B:input,VCF " + lowPass.inVCF +
-                        " -L 1:50,000,000-200,000,000" +
-                        " -qual 50.0" +
-                        " --ignore_filter GATK_STANDARD" +
-                        " -an QD -an HRun -an ThisAnnotationIsBAD" + // There is a bad annotation here
-                        " -clusterFile %s",
-                1, // just one output file
-                UserException.MalformedFile.class);
-        executeTest("testFailWithBadAnnotation", spec);
     }
 }
 

java/test/org/broadinstitute/sting/playground/gatk/walkers/variantrecalibration/VariantRecalibrationWalkersV2IntegrationTest.java

@@ -1,73 +0,0 @@
-package org.broadinstitute.sting.playground.gatk.walkers.variantrecalibration;
-
-import org.broadinstitute.sting.WalkerTest;
-import org.testng.annotations.Test;
-import org.testng.annotations.DataProvider;
-
-import java.util.*;
-
-public class VariantRecalibrationWalkersV2IntegrationTest extends WalkerTest {
-    static HashMap<String, String> clusterFiles = new HashMap<String, String>();
-    static HashMap<String, String> tranchesFiles = new HashMap<String, String>();
-    static HashMap<String, String> inputVCFFiles = new HashMap<String, String>();
-
-    private static class VRTest {
-        String inVCF;
-        String tranchesMD5;
-        String recalMD5;
-        String cutVCFMD5;
-        public VRTest(String inVCF, String tranchesMD5, String recalMD5, String cutVCFMD5) {
-            this.inVCF = validationDataLocation + inVCF;
-            this.tranchesMD5 = tranchesMD5;
-            this.recalMD5 = recalMD5;
-            this.cutVCFMD5 = cutVCFMD5;
-        }
-    }
-
-    VRTest lowPass = new VRTest("phase1.projectConsensus.chr20.raw.snps.vcf",
-            "920b12d7765eb4f6f4a1bab045679b31",  // tranches
-            "41bbc5f07c8a9573d5bb638f01808bba",  // recal file
-            "d9aa7a0f8fb886df4394f1636605adca"); // cut VCF
-
-    @DataProvider(name = "VRTest")
-    public Object[][] createData1() {
-        return new Object[][]{ {lowPass} };
-        //return new Object[][]{ {yriTrio}, {lowPass} }; // Add hg19 chr20 trio calls here
-    }
-
-    @Test(dataProvider = "VRTest")
-    public void testVariantRecalibrator(VRTest params) {
-        //System.out.printf("PARAMS FOR %s is %s%n", vcf, clusterFile);
-        WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b37KGReference +
-                        " -B:dbsnp,VCF,known=true,training=false,truth=false,prior=10.0 " + GATKDataLocation + "dbsnp_132_b37.leftAligned.vcf" +
-                        " -B:hapmap,VCF,known=false,training=true,truth=true,prior=15.0 " + comparisonDataLocation + "Validated/HapMap/3.3/sites_r27_nr.b37_fwd.vcf" +
-                        " -B:omni,VCF,known=false,training=true,truth=true,prior=12.0 " + comparisonDataLocation + "Validated/Omni2.5_chip/Omni25_sites_1525_samples.b37.vcf" +
-                        " -T VariantRecalibrator" +
-                        " -B:input,VCF " + params.inVCF +
-                        " -L 20:1,000,000-40,000,000" +
-                        " -an QD -an HaplotypeScore -an HRun" +
-                        " -percentBad 0.07" +
-                        " --trustAllPolymorphic" + // for speed
-                        " -recalFile %s" +
-                        " -tranchesFile %s",
-                Arrays.asList(params.recalMD5, params.tranchesMD5));
-        executeTest("testVariantRecalibrator-"+params.inVCF, spec).getFirst();
-    }
-
-    @Test(dataProvider = "VRTest",dependsOnMethods="testVariantRecalibrator")
-    public void testApplyRecalibration(VRTest params) {
-        WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
-                "-R " + b37KGReference +
-                        " -T ApplyRecalibration" +
-                        " -L 20:12,000,000-30,000,000" +
-                        " -NO_HEADER" +
-                        " -B:input,VCF " + params.inVCF +
-                        " -o %s" +
-                        " -tranchesFile " + getFileForMD5(params.tranchesMD5) +
-                        " -recalFile " + getFileForMD5(params.recalMD5),
-                Arrays.asList(params.cutVCFMD5));
-        executeTest("testApplyRecalibration-"+params.inVCF, spec);
-    }
-}
-