From 51d1a267259adf3279892327d7fab95282c2b1a4 Mon Sep 17 00:00:00 2001 From: Eric Banks Date: Tue, 26 Nov 2013 22:36:35 -0500 Subject: [PATCH] Bug fix for something Guillermo added to UG before he left to support calling indels from reduced reads. His code was excessively clipping reads because it was looking at their cigar string instead of just the read length. This meant that it was basically impossible to call large deletions in UG even with perfect evidence in the reads (as reported by Craig D). Integration tests change because (IMO after looking at sites in IGV) reads with indels similar to the one being genotyped used to be given too much likelihood and now give less. Added unit tests for new methods. --- .../indels/PairHMMIndelErrorModel.java | 46 ++++-- ...dGenotyperIndelCallingIntegrationTest.java | 4 +- .../UnifiedGenotyperIntegrationTest.java | 4 +- ...GenotyperNormalCallingIntegrationTest.java | 2 +- .../PairHMMIndelErrorModelUnitTest.java | 133 ++++++++++++++++++ 5 files changed, 172 insertions(+), 17 deletions(-) create mode 100644 protected/java/test/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModelUnitTest.java diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModel.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModel.java index 97431368e..a273cf01d 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModel.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModel.java @@ -68,8 +68,6 @@ import java.util.LinkedHashMap; import java.util.LinkedList; import java.util.Map; -//import org.broadinstitute.sting.utils.pairhmm.LoglessCachingPairHMM; - public class PairHMMIndelErrorModel { public static final int BASE_QUAL_THRESHOLD = 20; @@ -256,6 +254,28 @@ public class PairHMMIndelErrorModel { } + /** + * Should we clip a downstream portion of a read because it spans off the end of a haplotype? + * + * @param read the read in question + * @param refWindowStop the end of the reference window + * @return true if the read needs to be clipped, false otherwise + */ + protected static boolean mustClipDownstream(final GATKSAMRecord read, final int refWindowStop) { + return ( !read.isEmpty() && read.getSoftStart() < refWindowStop && read.getSoftStart() + read.getReadLength() > refWindowStop ); + } + + /** + * Should we clip a upstream portion of a read because it spans off the end of a haplotype? + * + * @param read the read in question + * @param refWindowStart the start of the reference window + * @return true if the read needs to be clipped, false otherwise + */ + protected static boolean mustClipUpstream(final GATKSAMRecord read, final int refWindowStart) { + return ( !read.isEmpty() && read.getSoftStart() < refWindowStart && read.getSoftEnd() > refWindowStart ); + } + @Ensures("result != null && result.length == pileup.getNumberOfElements()") public synchronized double[][] computeGeneralReadHaplotypeLikelihoods(final ReadBackedPileup pileup, final LinkedHashMap haplotypeMap, @@ -285,9 +305,8 @@ public class PairHMMIndelErrorModel { // in them - a value of 1 will in theory do but we use a slightly higher one just for safety sake, mostly // in case bases at edge of reads have lower quality. final int trailingBases = 3; - final int extraOffset = Math.abs(eventLength); - final int refWindowStart = ref.getWindow().getStart()+(trailingBases+extraOffset); - final int refWindowStop = ref.getWindow().getStop()-(trailingBases+extraOffset); + final int refWindowStart = ref.getWindow().getStart() + trailingBases; + final int refWindowStop = ref.getWindow().getStop() - trailingBases; if (DEBUG) { System.out.format("Read Name:%s, aln start:%d aln stop:%d orig cigar:%s\n",p.getRead().getReadName(), p.getRead().getAlignmentStart(), p.getRead().getAlignmentEnd(), p.getRead().getCigarString()); @@ -295,11 +314,13 @@ public class PairHMMIndelErrorModel { GATKSAMRecord read = ReadClipper.hardClipAdaptorSequence(p.getRead()); - if (!read.isEmpty() && (read.getSoftEnd() > refWindowStop && read.getSoftStart() < refWindowStop)) - read = ReadClipper.hardClipByReferenceCoordinatesRightTail(read, refWindowStop); + // if the read extends beyond the downstream (right) end of the reference window, clip it + if ( mustClipDownstream(read, refWindowStop) ) + read = ReadClipper.hardClipByReadCoordinates(read, read.getSoftStart() + read.getReadLength() - refWindowStop + 1, read.getReadLength() - 1); - if (!read.isEmpty() && (read.getSoftStart() < refWindowStart && read.getSoftEnd() > refWindowStart)) - read = ReadClipper.hardClipByReferenceCoordinatesLeftTail (read, refWindowStart); + // if the read extends beyond the upstream (left) end of the reference window, clip it + if ( mustClipUpstream(read, refWindowStart) ) + read = ReadClipper.hardClipByReferenceCoordinatesLeftTail(read, refWindowStart); if (read.isEmpty()) continue; @@ -337,8 +358,9 @@ public class PairHMMIndelErrorModel { * trailingBases is a padding constant(=3) and we additionally add abs(eventLength) to both sides of read to be able to * differentiate context between two haplotypes */ - long startLocationInRefForHaplotypes = Math.max(readStart + numStartSoftClippedBases - trailingBases - ReadUtils.getFirstInsertionOffset(read)-extraOffset, 0); - long stopLocationInRefForHaplotypes = readEnd -numEndSoftClippedBases + trailingBases + ReadUtils.getLastInsertionOffset(read)+extraOffset; + final int absEventLength = Math.abs(eventLength); + long startLocationInRefForHaplotypes = Math.max(readStart + numStartSoftClippedBases - trailingBases - ReadUtils.getFirstInsertionOffset(read) - absEventLength, 0); + long stopLocationInRefForHaplotypes = readEnd - numEndSoftClippedBases + trailingBases + ReadUtils.getLastInsertionOffset(read) + absEventLength; if (DEBUG) System.out.format("orig Start:%d orig stop: %d\n", startLocationInRefForHaplotypes, stopLocationInRefForHaplotypes); @@ -412,7 +434,7 @@ public class PairHMMIndelErrorModel { // Pack the shortened read and its associated gap-continuation-penalty array into a map, as required by PairHMM readGCPArrayMap.put(processedRead,contextLogGapContinuationProbabilities); - // Create a map of alleles to a new set of haplotypes, whose bases have been trimmed to the approprate genomic locations + // Create a map of alleles to a new set of haplotypes, whose bases have been trimmed to the appropriate genomic locations final Map trimmedHaplotypeMap = trimHaplotypes(haplotypeMap, startLocationInRefForHaplotypes, stopLocationInRefForHaplotypes, ref); // Get the likelihoods for our clipped read against each of our trimmed haplotypes. diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java index 7fa716168..9556f9bf1 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java @@ -101,7 +101,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { " -o %s" + " -L 1:10,000,000-10,500,000", 1, - Arrays.asList("c6f0fa039ca5672469838bc9f52c72d3")); + Arrays.asList("3d12bdb816d27bf7c9efb4c13dc2aec7")); executeTest(String.format("test indel calling, multiple technologies"), spec); } @@ -136,7 +136,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { 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("8682738c2c66b502cdbf7db466a5c3e2")); + Arrays.asList("a2c8e83f37cd1e114b42af4b873f57bc")); executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java index 74cd3f16e..2cdddd49f 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java @@ -260,7 +260,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest { " -o %s" + " -L 1:10,000,000-10,100,000", 1, - Arrays.asList("9f4e663e3b156b14fd55df3f5f0336a5")); + Arrays.asList("150b31ba05113ca1996b548be5170d6d")); executeTest(String.format("test multiple technologies"), spec); } @@ -279,7 +279,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest { " -L 1:10,000,000-10,100,000" + " -baq CALCULATE_AS_NECESSARY", 1, - Arrays.asList("260bb73e2900334d5c3ff8123be0d2d8")); + Arrays.asList("7d0ee85cd89f4addd84c5511daaaa5c5")); executeTest(String.format("test calling with BAQ"), spec); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java index afec97b84..18554e157 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java @@ -64,7 +64,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{ 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("7f26ca78e550afa28df11d593c90ec9a")); + Arrays.asList("ec0977e3fd3e2ac29c9821f0ca830455")); executeTest("test MultiSample Pilot1", spec); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModelUnitTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModelUnitTest.java new file mode 100644 index 000000000..38c06c25f --- /dev/null +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/indels/PairHMMIndelErrorModelUnitTest.java @@ -0,0 +1,133 @@ +/* +* 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.indels; + + +import net.sf.picard.reference.IndexedFastaSequenceFile; +import net.sf.samtools.SAMFileHeader; +import org.broadinstitute.sting.BaseTest; +import org.broadinstitute.sting.utils.fasta.CachingIndexedFastaSequenceFile; +import org.broadinstitute.sting.utils.sam.ArtificialSAMUtils; +import org.broadinstitute.sting.utils.sam.GATKSAMRecord; +import org.testng.Assert; +import org.testng.annotations.BeforeClass; +import org.testng.annotations.DataProvider; +import org.testng.annotations.Test; + +import java.io.File; +import java.io.FileNotFoundException; +import java.util.*; + +public class PairHMMIndelErrorModelUnitTest extends BaseTest { + + private SAMFileHeader header; + + @BeforeClass + public void setup() throws FileNotFoundException { + final IndexedFastaSequenceFile seq = new CachingIndexedFastaSequenceFile(new File(b37KGReference)); + header = ArtificialSAMUtils.createArtificialSamHeader(seq.getSequenceDictionary()); + } + + private static final int refWindowStart = 1000; + private static final int refWindowEnd = 1100; + + @DataProvider(name = "ClipUpstreamProvider") + public Object[][] ClipUpstreamTestData() { + List tests = new ArrayList(); + + for ( final int readStart : Arrays.asList(900, 950, 990, 1000) ) { + for ( final int readLength : Arrays.asList(10, 50, 100) ) { + for ( final int delLength : Arrays.asList(0, 5, 10) ) { + tests.add(new Object[]{readStart, readLength, delLength}); + } + } + } + + return tests.toArray(new Object[][]{}); + } + + @Test(dataProvider = "ClipUpstreamProvider", enabled = true) + public void clipUpstreamTest(final int readStart, final int readLength, final int delLength) { + + final GATKSAMRecord read = ArtificialSAMUtils.createArtificialRead(header, "basicRead", 0, readStart, readLength); + if ( delLength == 0 ) + read.setCigarString(readLength + "M"); + else + read.setCigarString((readLength / 2) + "M" + delLength + "D" + (readLength / 2) + "M"); + + final boolean result = PairHMMIndelErrorModel.mustClipUpstream(read, refWindowStart); + Assert.assertEquals(result, read.getSoftStart() < refWindowStart && read.getSoftEnd() > refWindowStart); + } + + @DataProvider(name = "ClipDownstreamProvider") + public Object[][] ClipDownstreamTestData() { + List tests = new ArrayList(); + + for ( final int readStart : Arrays.asList(1000, 1050, 1090, 1100) ) { + for ( final int readLength : Arrays.asList(10, 50, 100) ) { + for ( final int delLength : Arrays.asList(0, 5, 10) ) { + tests.add(new Object[]{readStart, readLength, delLength}); + } + } + } + + return tests.toArray(new Object[][]{}); + } + + @Test(dataProvider = "ClipDownstreamProvider", enabled = true) + public void clipDownstreamTest(final int readStart, final int readLength, final int delLength) { + + final GATKSAMRecord read = ArtificialSAMUtils.createArtificialRead(header, "basicRead", 0, readStart, readLength); + if ( delLength == 0 ) + read.setCigarString(readLength + "M"); + else + read.setCigarString((readLength / 2) + "M" + delLength + "D" + (readLength / 2) + "M"); + + final boolean result = PairHMMIndelErrorModel.mustClipDownstream(read, refWindowEnd); + Assert.assertEquals(result, read.getSoftStart() < refWindowEnd && read.getSoftStart() + readLength > refWindowEnd); + } +} \ No newline at end of file