zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Split long-running integration test classes into multiple classes 

This is to facilitate the current experiment with class-level test
suite parallelism. It's our hope that with these changes, we can get
the runtime of the integration test suite down to 20 minutes or so.

-UnifiedGenotyper tests: these divided nicely into logical categories
 that also happened to distribute the runtime fairly evenly

-UnifiedGenotyperPloidy: these had to be divided arbitrarily into two
 classes in order to halve the runtime

-HaplotypeCaller: turns out that the tests for complex and symbolic
 variants make up half the runtime here, so merely moving these into
 a separate class was sufficient

-BiasedDownsampling: most of these tests use excessively large intervals
 that likely can't be reduced without defeating the goals of the tests. I'm
 disabling these tests for now until they can either be redesigned to use smaller
 intervals around the variants of interest, or refactored into unit tests
 (creating a JIRA for Yossi for this task)
This commit is contained in:
David Roazen 2013-03-01 13:06:58 -05:00
parent 6204e6ccc9
commit c5c99c8339
10 changed files with 761 additions and 425 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/BiasedDownsamplingIntegrationTest.java
View File

@ -67,7 +67,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
//
// --------------------------------------------------------------------------------------------------------------
@Test
@Test(enabled = false)
public void testContaminationDownsamplingFlat() {
WalkerTestSpec spec = new WalkerTestSpec(
baseCommand1 + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,010,000 -contamination 0.20", 1,
@ -75,7 +75,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
executeTest("test contamination_percentage_to_filter 0.20", spec);
}
@Test
@Test(enabled = false)
public void testContaminationDownsamplingFlatAndPerSample() {
WalkerTestSpec spec = new WalkerTestSpec(
baseCommand1 + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,010,000 --contamination_fraction_per_sample_file " + ArtificalBAMLocation + "NA12878.NA19240.contam.txt --contamination_fraction_to_filter 0.10", 1,
@ -83,7 +83,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
executeTest("test contamination_percentage_to_filter per-sample and .20 overall", spec);
}
@Test
@Test(enabled = false)
public void testContaminationDownsamplingPerSampleOnly() {
WalkerTestSpec spec = new WalkerTestSpec(
baseCommand1 + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,010,000 -contaminationFile " + ArtificalBAMLocation + "NA19240.contam.txt", 1,
@ -98,7 +98,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
//
// --------------------------------------------------------------------------------------------------------------
@Test
@Test(enabled = false)
private void testDefaultContamination() {
final String bam1 = "NA11918.with.1.NA12842.reduced.bam";
final String bam2 = "NA12842.with.1.NA11918.reduced.bam";
@ -116,47 +116,47 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
executeTest("test contamination on Artificial Contamination (flat) on " + bam1 + " and " + bam2 + " downsampling " + downsampling.toString(), spec);
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase1() {
testFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.05, "e2e5a8dd313f8d7e382e7d49dfac59a2");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase2() {
testFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.1, "549737002f98775fea8f46e7ea174dde");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase3() {
testFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.2, "529d82c2a33fcc303a5dc55de2d56979");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase4() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.2.NA11918.reduced.bam", 0.1, "b5689972fbb7d230a372ee5f0da1c6d7");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase5() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.2.NA11918.reduced.bam", 0.2, "9dceee2e921b53fbc1ce137a7e0b7b74");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase6() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.2.NA11918.reduced.bam", 0.3, "d6a74061033503af80dcaea065bfa075");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase7() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.1, "7d1b5efab58a1b8f9d99fcf5af82f15a");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase8() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.2, "a7f8d5c79626aff59d7f426f79d8816e");
}
@Test
@Test(enabled = false)
public void testFlatContaminationCase9() {
testFlatContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.3, "fcf482398b7c908e3e2d1e4d5da6377b");
}
@ -168,42 +168,42 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
executeTest("test contamination on Artificial Contamination (per-sample) on " + bam1 + " and " + bam2 + " with " + persampleFile, spec);
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase1() {
testPerSampleContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.1.txt", "e00278527a294833259e9e411728e395");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase2() {
testPerSampleContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.2.txt", "a443e793f0b0e2ffce1b751634d706e2");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase3() {
testPerSampleContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.3.txt", "e11d83a7815ce757afbcf7689568cb25");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase4() {
testPerSampleContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.4.txt", "615042eeeffe042bd1c86279d34f80b6");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase5() {
testPerSampleContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.1.txt", "9bc99fc79ca34744bf26cb19ee4ef44d");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase6() {
testPerSampleContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.2.txt", "143626fe5fce765d6c997a64f058a813");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase7() {
testPerSampleContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.3.txt", "f2593674cef894eda4e0be9cf3158f57");
}
@Test
@Test(enabled = false)
public void testPerSampleContaminationCase8() {
testPerSampleContamination("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.4.txt", "fb7ce0740767ae3896b3e552026da1e4");
}
@ -227,17 +227,17 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
// verify that inputing a file with an effectively flat contamination level is equivalent to handing in a flat contamination level
@Test
@Test(enabled = false)
public void testPerSampleEqualsFlatContaminationCase1() {
testPerSampleEqualsFlat("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.6.txt", 0.0, "");
}
@Test
@Test(enabled = false)
public void testPerSampleEqualsFlatContaminationCase2() {
testPerSampleEqualsFlat("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.7.txt", 0.15, "");
}
@Test
@Test(enabled = false)
public void testPerSampleEqualsFlatContaminationCase3() {
testPerSampleEqualsFlat("NA11918.with.2.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", ArtificalBAMLocation + "contamination.case.8.txt", 0.3, "");
}
@ -250,7 +250,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
// --------------------------------------------------------------------------------------------------------------
@Test
@Test(enabled = false)
public void testHCContaminationDownsamplingFlat() {
final String baseCommand = "-T HaplotypeCaller -R " + b36KGReference + " --no_cmdline_in_header --dbsnp " + b36dbSNP129;
WalkerTestSpec spec = new WalkerTestSpec(
@ -260,7 +260,7 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
}
// HaplotypeCaller can only (currently) use flat contamination reduction, not per-sample. Until that is implemented, this test
@Test
@Test(enabled = false)
public void testHCCannotProcessPerSampleContamination() {
final String baseCommand = "-T HaplotypeCaller -R " + hg19Reference + " --no_cmdline_in_header -L 20:3,000,000-5,000,000";
final String bam1 = "NA11918.with.1.NA12842.reduced.bam";
@ -281,17 +281,17 @@ public class BiasedDownsamplingIntegrationTest extends WalkerTest {
executeTest("HC test contamination on Artificial Contamination (flat) on " + bam1 + " and " + bam2 + " downsampling " + downsampling.toString(), spec);
}
@Test
@Test(enabled = false)
public void testHCFlatContaminationCase1() {
testHCFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.05, "c3e695381d8627e3922d8c642b66c3ce");
}
@Test
@Test(enabled = false)
public void testHCFlatContaminationCase2() {
testHCFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.1, "002d2b45336d88d7c04e19f9f26e29d9");
}
@Test
@Test(enabled = false)
public void testHCFlatContaminationCase3() {
testHCFlatContamination("NA11918.with.1.NA12842.reduced.bam", "NA12842.with.1.NA11918.reduced.bam", 0.2, "1809a33ac112d1a3bd7a071c566794dd");
}

84
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite1IntegrationTest.java 100644
View File

@ -0,0 +1,84 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import static org.broadinstitute.sting.gatk.walkers.genotyper.UnifiedGenotyperGeneralPloidyTestExecutor.*;
/**
* Created by IntelliJ IDEA.
* User: delangel
* Date: 4/5/12
* Time: 11:28 AM
* To change this template use File | Settings | File Templates.
*/
public class UnifiedGenotyperGeneralPloidySuite1IntegrationTest extends WalkerTest {
private final UnifiedGenotyperGeneralPloidyTestExecutor executor = new UnifiedGenotyperGeneralPloidyTestExecutor();
@Test(enabled = true)
public void testSNP_ACS_Pools() {
executor.PC_LSV_Test_short(" -maxAltAlleles 1 -ploidy 6 -out_mode EMIT_ALL_CONFIDENT_SITES", "LSV_SNP_ACS", "SNP", "df0e67c975ef74d593f1c704daab1705");
}
@Test(enabled = true)
public void testBOTH_GGA_Pools() {
executor.PC_LSV_Test(String.format(" -maxAltAlleles 2 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s", LSV_ALLELES), "LSV_BOTH_GGA", "BOTH", "71f16e19b7d52e8edee46f4121e59f54");
}
@Test(enabled = true)
public void testINDEL_GGA_Pools() {
executor.PC_LSV_Test(String.format(" -maxAltAlleles 1 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s", LSV_ALLELES), "LSV_INDEL_GGA", "INDEL", "3f7d763c654f1d708323f369ea4a099b");
}
@Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() {
executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1", "LSV_INDEL_DISC_NOREF_p1", "INDEL", "5812da66811887d834d0379a33e655c0");
}
}

72
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite2IntegrationTest.java 100644
View File

@ -0,0 +1,72 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import static org.broadinstitute.sting.gatk.walkers.genotyper.UnifiedGenotyperGeneralPloidyTestExecutor.*;
public class UnifiedGenotyperGeneralPloidySuite2IntegrationTest extends WalkerTest {
private final UnifiedGenotyperGeneralPloidyTestExecutor executor = new UnifiedGenotyperGeneralPloidyTestExecutor();
@Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() {
executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","3a321896c4b8b6457973c76c486da4d4");
}
@Test(enabled = true)
public void testMT_SNP_DISCOVERY_sp4() {
executor.PC_MT_Test(CEUTRIO_BAM, " -maxAltAlleles 1 -ploidy 8", "MT_SNP_DISCOVERY_sp4","3fc6f4d458313616727c60e49c0e852b");
}
@Test(enabled = true)
public void testMT_SNP_GGA_sp10() {
executor.PC_MT_Test(CEUTRIO_BAM, String.format(" -maxAltAlleles 1 -ploidy 20 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",NA12891_CALLS), "MT_SNP_GGA_sp10", "1bebbc0f28bff6fd64736ccca8839df8");
}
}

73
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidyIntegrationTest.java → protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidyTestExecutor.java
View File

@ -47,90 +47,47 @@
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import java.util.Arrays;
/**
* Created by IntelliJ IDEA.
* User: delangel
* Date: 4/5/12
* Time: 11:28 AM
* To change this template use File | Settings | File Templates.
*/
public class UnifiedGenotyperGeneralPloidyIntegrationTest extends WalkerTest {
public class UnifiedGenotyperGeneralPloidyTestExecutor extends WalkerTest {
final static String REF = b37KGReference;
final String CEUTRIO_BAM = "/humgen/gsa-hpprojects/NA12878Collection/bams/CEUTrio.HiSeq.WGS.b37.list";
final String LSV_BAM = validationDataLocation +"93pools_NA12878_ref_chr20_40m_41m.bam";
final String REFSAMPLE_MT_CALLS = comparisonDataLocation + "Unvalidated/mtDNA/NA12878.snp.vcf";
final String REFSAMPLE_NAME = "NA12878";
final String MTINTERVALS = "MT:1-1000";
final String LSVINTERVALS = "20:40,500,000-41,000,000";
final String LSVINTERVALS_SHORT = "20:40,500,000-40,501,000";
final String NA12891_CALLS = comparisonDataLocation + "Unvalidated/mtDNA/NA12891.snp.vcf";
final String NA12878_WG_CALLS = comparisonDataLocation + "Unvalidated/NA12878/CEUTrio.HiSeq.WGS.b37_decoy.recal.ts_95.snp_indel_combined.vcf";
final String LSV_ALLELES = validationDataLocation + "ALL.chr20_40m_41m.largeScaleValidationSites.vcf";
final static String CEUTRIO_BAM = "/humgen/gsa-hpprojects/NA12878Collection/bams/CEUTrio.HiSeq.WGS.b37.list";
final static String LSV_BAM = validationDataLocation +"93pools_NA12878_ref_chr20_40m_41m.bam";
final static String REFSAMPLE_MT_CALLS = comparisonDataLocation + "Unvalidated/mtDNA/NA12878.snp.vcf";
final static String REFSAMPLE_NAME = "NA12878";
final static String MTINTERVALS = "MT:1-1000";
final static String LSVINTERVALS = "20:40,500,000-41,000,000";
final static String LSVINTERVALS_SHORT = "20:40,500,000-40,501,000";
final static String NA12891_CALLS = comparisonDataLocation + "Unvalidated/mtDNA/NA12891.snp.vcf";
final static String NA12878_WG_CALLS = comparisonDataLocation + "Unvalidated/NA12878/CEUTrio.HiSeq.WGS.b37_decoy.recal.ts_95.snp_indel_combined.vcf";
final static String LSV_ALLELES = validationDataLocation + "ALL.chr20_40m_41m.largeScaleValidationSites.vcf";
private void PC_MT_Test(String bam, String args, String name, String md5) {
public void PC_MT_Test(String bam, String args, String name, String md5) {
final String base = String.format("-T UnifiedGenotyper -dcov 10000 -R %s -I %s -L %s --reference_sample_calls %s -refsample %s -ignoreLane ",
REF, bam, MTINTERVALS, REFSAMPLE_MT_CALLS, REFSAMPLE_NAME) + " --no_cmdline_in_header -o %s";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testPoolCaller:"+name+" args=" + args, spec);
}
private void PC_LSV_Test(String args, String name, String model, String md5) {
public void PC_LSV_Test(String args, String name, String model, String md5) {
final String base = String.format("-T UnifiedGenotyper -dcov 10000 -R %s -I %s -L %s --reference_sample_calls %s -refsample %s -glm %s -ignoreLane ",
REF, LSV_BAM, LSVINTERVALS, NA12878_WG_CALLS, REFSAMPLE_NAME, model) + " --no_cmdline_in_header -o %s";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testPoolCaller:"+name+" args=" + args, spec);
}
private void PC_LSV_Test_short(String args, String name, String model, String md5) {
public void PC_LSV_Test_short(String args, String name, String model, String md5) {
final String base = String.format("-T UnifiedGenotyper -dcov 10000 -R %s -I %s -L %s --reference_sample_calls %s -refsample %s -glm %s -ignoreLane ",
REF, LSV_BAM, LSVINTERVALS_SHORT, NA12878_WG_CALLS, REFSAMPLE_NAME, model) + " --no_cmdline_in_header -o %s";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testPoolCaller:"+name+" args=" + args, spec);
}
private void PC_LSV_Test_NoRef(String args, String name, String model, String md5) {
public void PC_LSV_Test_NoRef(String args, String name, String model, String md5) {
final String base = String.format("-T UnifiedGenotyper -dcov 10000 -R %s -I %s -L %s -glm %s -ignoreLane",
REF, LSV_BAM, LSVINTERVALS, model) + " --no_cmdline_in_header -o %s";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testPoolCaller:"+name+" args=" + args, spec);
}
@Test(enabled = true)
public void testSNP_ACS_Pools() {
PC_LSV_Test_short(" -maxAltAlleles 1 -ploidy 6 -out_mode EMIT_ALL_CONFIDENT_SITES","LSV_SNP_ACS","SNP","df0e67c975ef74d593f1c704daab1705");
}
@Test(enabled = true)
public void testBOTH_GGA_Pools() {
PC_LSV_Test(String.format(" -maxAltAlleles 2 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_BOTH_GGA","BOTH","71f16e19b7d52e8edee46f4121e59f54");
}
@Test(enabled = true)
public void testINDEL_GGA_Pools() {
PC_LSV_Test(String.format(" -maxAltAlleles 1 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_INDEL_GGA","INDEL","3f7d763c654f1d708323f369ea4a099b");
}
@Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() {
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","3a321896c4b8b6457973c76c486da4d4");
}
@Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() {
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1","LSV_INDEL_DISC_NOREF_p1","INDEL","5812da66811887d834d0379a33e655c0");
}
@Test(enabled = true)
public void testMT_SNP_DISCOVERY_sp4() {
PC_MT_Test(CEUTRIO_BAM, " -maxAltAlleles 1 -ploidy 8", "MT_SNP_DISCOVERY_sp4","3fc6f4d458313616727c60e49c0e852b");
}
@Test(enabled = true)
public void testMT_SNP_GGA_sp10() {
PC_MT_Test(CEUTRIO_BAM, String.format(" -maxAltAlleles 1 -ploidy 20 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",NA12891_CALLS), "MT_SNP_GGA_sp10", "1bebbc0f28bff6fd64736ccca8839df8");
}
}

197
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java 100644
View File

@ -0,0 +1,197 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import java.io.File;
import java.util.Arrays;
import java.util.List;
public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
private final static String baseCommandIndels = "-T UnifiedGenotyper -R " + b36KGReference + " --no_cmdline_in_header -glm INDEL -mbq 20 -minIndelFrac 0.0 --dbsnp " + b36dbSNP129;
private final static String baseCommandIndelsb37 = "-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -mbq 20 --dbsnp " + b37dbSNP132;
// --------------------------------------------------------------------------------------------------------------
//
// testing indel caller
//
// --------------------------------------------------------------------------------------------------------------
// Basic indel testing with SLX data
@Test
public void testSimpleIndels() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam" +
" -o %s" +
" -L 1:10,000,000-10,500,000",
1,
Arrays.asList("1cb469b9cc8e6c70430021540bf1af8b"));
executeTest(String.format("test indel caller in SLX"), spec);
}
// Basic indel testing with SLX data
@Test
public void testIndelsWithLowMinAlleleCnt() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam" +
" -o %s" +
" -minIndelCnt 1" +
" -L 1:10,000,000-10,100,000",
1,
Arrays.asList("c7e59f9ab718df4c604626a0f51af606"));
executeTest(String.format("test indel caller in SLX with low min allele count"), spec);
}
@Test
public void testMultiTechnologyIndels() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.allTechs.bam" +
" -o %s" +
" -L 1:10,000,000-10,500,000",
1,
Arrays.asList("4bebbe4ed4a7554285a3b4bb7311101c"));
executeTest(String.format("test indel calling, multiple technologies"), spec);
}
@Test
public void testWithIndelAllelesPassedIn1() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation +
"pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("86880ec78755ae91cb5bb34a0631a32c"));
executeTest("test MultiSample Pilot2 indels with alleles passed in", spec);
}
@Test
public void testWithIndelAllelesPassedIn2() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles "
+ privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation +
"pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("2584d5e3ade1b548f1fe9cdcafbe1b28"));
executeTest("test MultiSample Pilot2 indels with alleles passed in and emitting all sites", spec);
}
@Test(timeOut = 20*1000*60) // this guy can take a long time because it's two steps, so give it 12 minutes
public void testMultiSampleIndels1() {
// since we're going to test the MD5s with GGA only do one here
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10450700-10551000", 1,
Arrays.asList(""));
List<File> result = executeTest("test MultiSample Pilot1 CEU indels", spec1).getFirst();
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + result.get(0).getAbsolutePath() + " -I " + validationDataLocation +
"low_coverage_CEU.chr1.10k-11k.bam -o %s -L " + result.get(0).getAbsolutePath(), 1,
Arrays.asList("08b3a85be00c8f6a4fefd3c671463ecf"));
executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2);
}
@Test
public void testGGAwithNoEvidenceInReads() {
final String vcf = "small.indel.test.vcf";
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndelsb37 + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles " + privateTestDir + vcf + " -I " + validationDataLocation +
"NA12878.HiSeq.WGS.bwa.cleaned.recal.hg19.20.bam -o %s -L " + validationDataLocation + vcf, 1,
Arrays.asList("d76eacc4021b78ccc0a9026162e814a7"));
executeTest("test GENOTYPE_GIVEN_ALLELES with no evidence in reads", spec);
}
@Test
public void testBaseIndelQualityScores() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndelsb37 +
" -I " + privateTestDir + "NA12878.100kb.BQSRv2.example.bam" +
" -o %s" +
" -L 20:10,000,000-10,100,000",
1,
Arrays.asList("8a7966e4b67334bca6083670c5a16b67"));
executeTest(String.format("test UG with base indel quality scores"), spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing MinIndelFraction
//
// --------------------------------------------------------------------------------------------------------------
final static String assessMinIndelFraction = baseCommandIndelsb37 + " -I " + validationDataLocation
+ "978604.bam -L 1:978,586-978,626 -o %s --sites_only -rf Sample -goodSM 7377 -goodSM 22-0022 -goodSM 134 -goodSM 344029-53 -goodSM 14030";
@Test
public void testMinIndelFraction0() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.0", 1,
Arrays.asList("556c214366e82e4682e753ce93307a4e"));
executeTest("test minIndelFraction 0.0", spec);
}
@Test
public void testMinIndelFraction25() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.25", 1,
Arrays.asList("1df02b805d9dfbd532fa3632875a989d"));
executeTest("test minIndelFraction 0.25", spec);
}
@Test
public void testMinIndelFraction100() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 1", 1,
Arrays.asList("3f07efb768e08650a7ce333edd4f9a52"));
executeTest("test minIndelFraction 1.0", spec);
}
}

340
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java
View File

@ -51,10 +51,8 @@ import org.broadinstitute.sting.gatk.GenomeAnalysisEngine;
import org.broadinstitute.sting.utils.exceptions.UserException;
import org.testng.annotations.Test;
import java.io.File;
import java.util.Arrays;
import java.util.Collections;
import java.util.List;
// ********************************************************************************** //
// Note that this class also serves as an integration test for the VariantAnnotator! //
@ -63,128 +61,8 @@ import java.util.List;
public class UnifiedGenotyperIntegrationTest extends WalkerTest {
private final static String baseCommand = "-T UnifiedGenotyper -R " + b36KGReference + " --no_cmdline_in_header -glm BOTH -minIndelFrac 0.0 --dbsnp " + b36dbSNP129;
private final static String baseCommandIndels = "-T UnifiedGenotyper -R " + b36KGReference + " --no_cmdline_in_header -glm INDEL -mbq 20 -minIndelFrac 0.0 --dbsnp " + b36dbSNP129;
private final static String baseCommandIndelsb37 = "-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -mbq 20 --dbsnp " + b37dbSNP132;
private final static String baseCommandNoCmdLineHeaderStdout = "-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "bamExample.ReducedRead.ADAnnotation.bam";
// --------------------------------------------------------------------------------------------------------------
//
// testing normal calling
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testMultiSamplePilot1() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10,022,000-10,025,000", 1,
Arrays.asList("2f15ef1ead56d875a3f1d53772f52b3a"));
executeTest("test MultiSample Pilot1", spec);
}
@Test
public void testWithAllelesPassedIn1() {
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "allelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,025,000", 1,
Arrays.asList("5b31b811072a4df04524e13604015f9b"));
executeTest("test MultiSample Pilot2 with alleles passed in", spec1);
}
@Test
public void testWithAllelesPassedIn2() {
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommand + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "allelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,025,000", 1,
Arrays.asList("d9992e55381afb43742cc9b30fcd7538"));
executeTest("test MultiSample Pilot2 with alleles passed in and emitting all sites", spec2);
}
@Test
public void testSingleSamplePilot2() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,100,000", 1,
Arrays.asList("33ab66c2f062cfa1f7fcc077165f778c"));
executeTest("test SingleSample Pilot2", spec);
}
@Test
public void testMultipleSNPAlleles() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH --dbsnp " + b37dbSNP129 + " -I " + privateTestDir + "multiallelic.snps.bam -o %s -L " + privateTestDir + "multiallelic.snps.intervals", 1,
Arrays.asList("9fac00485419878749b03706ae6b852f"));
executeTest("test Multiple SNP alleles", spec);
}
@Test
public void testBadRead() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH -I " + privateTestDir + "badRead.test.bam -o %s -L 1:22753424-22753464", 1,
Arrays.asList("d915535c1458733f09f82670092fcab6"));
executeTest("test bad read", spec);
}
@Test
public void testReverseTrim() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + validationDataLocation + "CEUTrio.HiSeq.b37.chr20.10_11mb.bam -o %s -L 20:10289124 -L 20:10090289", 1,
Arrays.asList("eb9604b77a7d6baab60c81ac3db5e47b"));
executeTest("test reverse trim", spec);
}
@Test
public void testMismatchedPLs() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + privateTestDir + "mismatchedPLs.bam -o %s -L 1:24020341", 1,
Arrays.asList("de2c5707c1805d17d70acaecd36b7372"));
executeTest("test mismatched PLs", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing compressed output
//
// --------------------------------------------------------------------------------------------------------------
private final static String COMPRESSED_OUTPUT_MD5 = "d5a7326fdcf6d441b73c381912ad3a2a";
@Test
public void testCompressedOutput() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,100,000", 1,
Arrays.asList("gz"), Arrays.asList(COMPRESSED_OUTPUT_MD5));
executeTest("test compressed output", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing parallelization
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testParallelization() {
// Note that we need to turn off any randomization for this to work, so no downsampling and no annotations
String md5 = "d408b4661b820ed86272415b8ea08780";
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000", 1,
Arrays.asList(md5));
executeTest("test parallelization (single thread)", spec1);
GenomeAnalysisEngine.resetRandomGenerator();
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000 -nt 2", 1,
Arrays.asList(md5));
executeTest("test parallelization (2 threads)", spec2);
GenomeAnalysisEngine.resetRandomGenerator();
WalkerTest.WalkerTestSpec spec3 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000 -nt 4", 1,
Arrays.asList(md5));
executeTest("test parallelization (4 threads)", spec3);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing parameters
@ -283,6 +161,54 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
executeTest(String.format("test heterozyosity[%s]", arg), spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing compressed output
//
// --------------------------------------------------------------------------------------------------------------
private final static String COMPRESSED_OUTPUT_MD5 = "d5a7326fdcf6d441b73c381912ad3a2a";
@Test
public void testCompressedOutput() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,100,000", 1,
Arrays.asList("gz"), Arrays.asList(COMPRESSED_OUTPUT_MD5));
executeTest("test compressed output", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing parallelization
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testParallelization() {
// Note that we need to turn off any randomization for this to work, so no downsampling and no annotations
String md5 = "d408b4661b820ed86272415b8ea08780";
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000", 1,
Arrays.asList(md5));
executeTest("test parallelization (single thread)", spec1);
GenomeAnalysisEngine.resetRandomGenerator();
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000 -nt 2", 1,
Arrays.asList(md5));
executeTest("test parallelization (2 threads)", spec2);
GenomeAnalysisEngine.resetRandomGenerator();
WalkerTest.WalkerTestSpec spec3 = new WalkerTest.WalkerTestSpec(
baseCommand + " -dt NONE -G none --contamination_fraction_to_filter 0.0 -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,075,000 -nt 4", 1,
Arrays.asList(md5));
executeTest("test parallelization (4 threads)", spec3);
}
// --------------------------------------------------------------------------------------------------------------
//
@ -321,110 +247,6 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
executeTest(String.format("test calling with BAQ"), spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing indel caller
//
// --------------------------------------------------------------------------------------------------------------
// Basic indel testing with SLX data
@Test
public void testSimpleIndels() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam" +
" -o %s" +
" -L 1:10,000,000-10,500,000",
1,
Arrays.asList("1cb469b9cc8e6c70430021540bf1af8b"));
executeTest(String.format("test indel caller in SLX"), spec);
}
// Basic indel testing with SLX data
@Test
public void testIndelsWithLowMinAlleleCnt() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam" +
" -o %s" +
" -minIndelCnt 1" +
" -L 1:10,000,000-10,100,000",
1,
Arrays.asList("c7e59f9ab718df4c604626a0f51af606"));
executeTest(String.format("test indel caller in SLX with low min allele count"), spec);
}
@Test
public void testMultiTechnologyIndels() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels +
" -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.allTechs.bam" +
" -o %s" +
" -L 1:10,000,000-10,500,000",
1,
Arrays.asList("4bebbe4ed4a7554285a3b4bb7311101c"));
executeTest(String.format("test indel calling, multiple technologies"), spec);
}
@Test
public void testWithIndelAllelesPassedIn1() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation +
"pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("86880ec78755ae91cb5bb34a0631a32c"));
executeTest("test MultiSample Pilot2 indels with alleles passed in", spec);
}
@Test
public void testWithIndelAllelesPassedIn2() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles "
+ privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation +
"pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("2584d5e3ade1b548f1fe9cdcafbe1b28"));
executeTest("test MultiSample Pilot2 indels with alleles passed in and emitting all sites", spec);
}
@Test(timeOut = 20*1000*60) // this guy can take a long time because it's two steps, so give it 12 minutes
public void testMultiSampleIndels1() {
// since we're going to test the MD5s with GGA only do one here
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10450700-10551000", 1,
Arrays.asList(""));
List<File> result = executeTest("test MultiSample Pilot1 CEU indels", spec1).getFirst();
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + result.get(0).getAbsolutePath() + " -I " + validationDataLocation +
"low_coverage_CEU.chr1.10k-11k.bam -o %s -L " + result.get(0).getAbsolutePath(), 1,
Arrays.asList("08b3a85be00c8f6a4fefd3c671463ecf"));
executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2);
}
@Test
public void testGGAwithNoEvidenceInReads() {
final String vcf = "small.indel.test.vcf";
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndelsb37 + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles " + privateTestDir + vcf + " -I " + validationDataLocation +
"NA12878.HiSeq.WGS.bwa.cleaned.recal.hg19.20.bam -o %s -L " + validationDataLocation + vcf, 1,
Arrays.asList("d76eacc4021b78ccc0a9026162e814a7"));
executeTest("test GENOTYPE_GIVEN_ALLELES with no evidence in reads", spec);
}
@Test
public void testBaseIndelQualityScores() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndelsb37 +
" -I " + privateTestDir + "NA12878.100kb.BQSRv2.example.bam" +
" -o %s" +
" -L 20:10,000,000-10,100,000",
1,
Arrays.asList("8a7966e4b67334bca6083670c5a16b67"));
executeTest(String.format("test UG with base indel quality scores"), spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing SnpEff
@ -441,39 +263,6 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
executeTest("testSnpEffAnnotationRequestedWithoutRodBinding", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing MinIndelFraction
//
// --------------------------------------------------------------------------------------------------------------
final static String assessMinIndelFraction = baseCommandIndelsb37 + " -I " + validationDataLocation
+ "978604.bam -L 1:978,586-978,626 -o %s --sites_only -rf Sample -goodSM 7377 -goodSM 22-0022 -goodSM 134 -goodSM 344029-53 -goodSM 14030";
@Test
public void testMinIndelFraction0() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.0", 1,
Arrays.asList("556c214366e82e4682e753ce93307a4e"));
executeTest("test minIndelFraction 0.0", spec);
}
@Test
public void testMinIndelFraction25() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.25", 1,
Arrays.asList("1df02b805d9dfbd532fa3632875a989d"));
executeTest("test minIndelFraction 0.25", spec);
}
@Test
public void testMinIndelFraction100() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 1", 1,
Arrays.asList("3f07efb768e08650a7ce333edd4f9a52"));
executeTest("test minIndelFraction 1.0", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing Ns in CIGAR
@ -487,37 +276,4 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
Arrays.asList("4d36969d4f8f1094f1fb6e7e085c19f6"));
executeTest("test calling on reads with Ns in CIGAR", spec);
}
// --------------------------------------------------------------------------------------------------------------
//
// testing reduced reads
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testReducedBam() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "bamExample.ReducedRead.ADAnnotation.bam -o %s -L 1:67,225,396-67,288,518", 1,
Arrays.asList("8b9a9fc2e7150acbe2dac91b4620f304"));
executeTest("test calling on a ReducedRead BAM", spec);
}
@Test
public void testReducedBamSNPs() {
testReducedCalling("SNP", "b5991dddbfb59366614ff8819062649f");
}
@Test
public void testReducedBamINDELs() {
testReducedCalling("INDEL", "acde5694a74f867256a54a26cbebbf21");
}
private void testReducedCalling(final String model, final String md5) {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "NA12878.HiSeq.b37.chr20.10_11mb.reduced.bam -o %s -L 20:10,000,000-11,000,000 -glm " + model, 1,
Arrays.asList(md5));
executeTest("test calling on a ReducedRead BAM with " + model, spec);
}
}

126
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java 100644
View File

@ -0,0 +1,126 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import java.util.Arrays;
public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{
private final static String baseCommand = "-T UnifiedGenotyper -R " + b36KGReference + " --no_cmdline_in_header -glm BOTH -minIndelFrac 0.0 --dbsnp " + b36dbSNP129;
// --------------------------------------------------------------------------------------------------------------
//
// testing normal calling
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testMultiSamplePilot1() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10,022,000-10,025,000", 1,
Arrays.asList("2f15ef1ead56d875a3f1d53772f52b3a"));
executeTest("test MultiSample Pilot1", spec);
}
@Test
public void testWithAllelesPassedIn1() {
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "allelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,025,000", 1,
Arrays.asList("5b31b811072a4df04524e13604015f9b"));
executeTest("test MultiSample Pilot2 with alleles passed in", spec1);
}
@Test
public void testWithAllelesPassedIn2() {
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommand + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "allelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,025,000", 1,
Arrays.asList("d9992e55381afb43742cc9b30fcd7538"));
executeTest("test MultiSample Pilot2 with alleles passed in and emitting all sites", spec2);
}
@Test
public void testSingleSamplePilot2() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -o %s -L 1:10,000,000-10,100,000", 1,
Arrays.asList("33ab66c2f062cfa1f7fcc077165f778c"));
executeTest("test SingleSample Pilot2", spec);
}
@Test
public void testMultipleSNPAlleles() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH --dbsnp " + b37dbSNP129 + " -I " + privateTestDir + "multiallelic.snps.bam -o %s -L " + privateTestDir + "multiallelic.snps.intervals", 1,
Arrays.asList("9fac00485419878749b03706ae6b852f"));
executeTest("test Multiple SNP alleles", spec);
}
@Test
public void testBadRead() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH -I " + privateTestDir + "badRead.test.bam -o %s -L 1:22753424-22753464", 1,
Arrays.asList("d915535c1458733f09f82670092fcab6"));
executeTest("test bad read", spec);
}
@Test
public void testReverseTrim() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + validationDataLocation + "CEUTrio.HiSeq.b37.chr20.10_11mb.bam -o %s -L 20:10289124 -L 20:10090289", 1,
Arrays.asList("eb9604b77a7d6baab60c81ac3db5e47b"));
executeTest("test reverse trim", spec);
}
@Test
public void testMismatchedPLs() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + privateTestDir + "mismatchedPLs.bam -o %s -L 1:24020341", 1,
Arrays.asList("de2c5707c1805d17d70acaecd36b7372"));
executeTest("test mismatched PLs", spec);
}
}

87
protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperReducedReadsIntegrationTest.java 100644
View File

@ -0,0 +1,87 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.genotyper;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import java.util.Arrays;
public class UnifiedGenotyperReducedReadsIntegrationTest extends WalkerTest {
// --------------------------------------------------------------------------------------------------------------
//
// testing reduced reads
//
// --------------------------------------------------------------------------------------------------------------
@Test
public void testReducedBam() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "bamExample.ReducedRead.ADAnnotation.bam -o %s -L 1:67,225,396-67,288,518", 1,
Arrays.asList("8b9a9fc2e7150acbe2dac91b4620f304"));
executeTest("test calling on a ReducedRead BAM", spec);
}
@Test
public void testReducedBamSNPs() {
testReducedCalling("SNP", "b5991dddbfb59366614ff8819062649f");
}
@Test
public void testReducedBamINDELs() {
testReducedCalling("INDEL", "acde5694a74f867256a54a26cbebbf21");
}
private void testReducedCalling(final String model, final String md5) {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "NA12878.HiSeq.b37.chr20.10_11mb.reduced.bam -o %s -L 20:10,000,000-11,000,000 -glm " + model, 1,
Arrays.asList(md5));
executeTest("test calling on a ReducedRead BAM with " + model, spec);
}
}

98
protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java 100644
View File

@ -0,0 +1,98 @@
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE - SOFTWARE LICENSE AGREEMENT - FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 7 Cambridge Center, Cambridge, MA 02142 (BROAD) and the LICENSEE and is effective at the date the downloading is completed (EFFECTIVE DATE).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/GATK on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012 Broad Institute, Inc.
* Notice of attribution: The GATK2 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 4. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 5. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 6. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 7. MISCELLANEOUS
* 7.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 7.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 7.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 7.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 7.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 7.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 7.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.sting.gatk.walkers.haplotypecaller;
import org.broadinstitute.sting.WalkerTest;
import org.testng.annotations.Test;
import static org.broadinstitute.sting.gatk.walkers.haplotypecaller.HaplotypeCallerIntegrationTest.*;
import java.util.Arrays;
public class HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest extends WalkerTest {
private void HCTestComplexVariants(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " -L 20:10028767-10028967 -L 20:10431524-10431924 -L 20:10723661-10724061 -L 20:10903555-10903955 --no_cmdline_in_header -o %s -minPruning 4";
final WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerComplexVariants: args=" + args, spec);
}
@Test
public void testHaplotypeCallerMultiSampleComplex() {
HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "a960722c1ae2b6f774d3443a7e5ac27d");
}
private void HCTestSymbolicVariants(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " -L 20:5947969-5948369 -L 20:61091236-61091636 --no_cmdline_in_header -o %s -minPruning 1";
final WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerSymbolicVariants: args=" + args, spec);
}
// TODO -- need a better symbolic allele test
@Test
public void testHaplotypeCallerSingleSampleSymbolic() {
HCTestSymbolicVariants(NA12878_CHR20_BAM, "", "56f2ef9acc6c0d267cf2b7a447d87fb7");
}
private void HCTestComplexGGA(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " --no_cmdline_in_header -o %s -minPruning 3 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles " + validationDataLocation + "combined.phase1.chr20.raw.indels.sites.vcf";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerComplexGGA: args=" + args, spec);
}
@Test
public void testHaplotypeCallerMultiSampleGGAComplex() {
HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:119673-119823 -L 20:121408-121538",
"417174e043dbb8b86cc3871da9b50536");
}
@Test
public void testHaplotypeCallerMultiSampleGGAMultiAllelic() {
HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:133041-133161 -L 20:300207-300337",
"f2df7a8f53ce449e4a8e8f8496e7c745");
}
}

51
protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerIntegrationTest.java
View File

@ -54,11 +54,11 @@ import java.util.Collections;
public class HaplotypeCallerIntegrationTest extends WalkerTest {
final static String REF = b37KGReference;
final String NA12878_BAM = validationDataLocation + "NA12878.HiSeq.b37.chr20.10_11mb.bam";
final String NA12878_CHR20_BAM = validationDataLocation + "NA12878.HiSeq.WGS.bwa.cleaned.recal.hg19.20.bam";
final String CEUTRIO_BAM = validationDataLocation + "CEUTrio.HiSeq.b37.chr20.10_11mb.bam";
final String NA12878_RECALIBRATED_BAM = privateTestDir + "NA12878.100kb.BQSRv2.example.bam";
final String INTERVALS_FILE = validationDataLocation + "NA12878.HiSeq.b37.chr20.10_11mb.test.intervals";
final static String NA12878_BAM = validationDataLocation + "NA12878.HiSeq.b37.chr20.10_11mb.bam";
final static String NA12878_CHR20_BAM = validationDataLocation + "NA12878.HiSeq.WGS.bwa.cleaned.recal.hg19.20.bam";
final static String CEUTRIO_BAM = validationDataLocation + "CEUTrio.HiSeq.b37.chr20.10_11mb.bam";
final static String NA12878_RECALIBRATED_BAM = privateTestDir + "NA12878.100kb.BQSRv2.example.bam";
final static String INTERVALS_FILE = validationDataLocation + "NA12878.HiSeq.b37.chr20.10_11mb.test.intervals";
private void HCTest(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s -L %s", REF, bam, INTERVALS_FILE) + " --no_cmdline_in_header -o %s -minPruning 3";
@ -87,47 +87,6 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest {
"283524b3e3397634d4cf0dc2b8723002");
}
private void HCTestComplexGGA(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " --no_cmdline_in_header -o %s -minPruning 3 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles " + validationDataLocation + "combined.phase1.chr20.raw.indels.sites.vcf";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerComplexGGA: args=" + args, spec);
}
@Test
public void testHaplotypeCallerMultiSampleGGAComplex() {
HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:119673-119823 -L 20:121408-121538",
"417174e043dbb8b86cc3871da9b50536");
}
@Test
public void testHaplotypeCallerMultiSampleGGAMultiAllelic() {
HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:133041-133161 -L 20:300207-300337",
"f2df7a8f53ce449e4a8e8f8496e7c745");
}
private void HCTestComplexVariants(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " -L 20:10028767-10028967 -L 20:10431524-10431924 -L 20:10723661-10724061 -L 20:10903555-10903955 --no_cmdline_in_header -o %s -minPruning 4";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerComplexVariants: args=" + args, spec);
}
@Test
public void testHaplotypeCallerMultiSampleComplex() {
HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "a960722c1ae2b6f774d3443a7e5ac27d");
}
private void HCTestSymbolicVariants(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " -L 20:5947969-5948369 -L 20:61091236-61091636 --no_cmdline_in_header -o %s -minPruning 1";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));
executeTest("testHaplotypeCallerSymbolicVariants: args=" + args, spec);
}
// TODO -- need a better symbolic allele test
@Test
public void testHaplotypeCallerSingleSampleSymbolic() {
HCTestSymbolicVariants(NA12878_CHR20_BAM, "", "56f2ef9acc6c0d267cf2b7a447d87fb7");
}
private void HCTestIndelQualityScores(String bam, String args, String md5) {
final String base = String.format("-T HaplotypeCaller -R %s -I %s", REF, bam) + " -L 20:10,005,000-10,025,000 --no_cmdline_in_header -o %s -minPruning 2";
final WalkerTestSpec spec = new WalkerTestSpec(base + " " + args, Arrays.asList(md5));