diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/AlleleLikelihoodMatrixMapper.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/AlleleLikelihoodMatrixMapper.java new file mode 100644 index 000000000..557dfd0ad --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/AlleleLikelihoodMatrixMapper.java @@ -0,0 +1,171 @@ +/* +* 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.gatk.genotyping; + +import htsjdk.variant.variantcontext.Allele; +import org.broadinstitute.gatk.utils.genotyper.ReadLikelihoods; +import org.broadinstitute.gatk.utils.sam.GATKSAMRecord; + +import java.util.List; + +/** + * Creates {@link org.broadinstitute.gatk.utils.genotyper.ReadLikelihoods.Matrix} mappers to be used when working with a subset of the original alleles. + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public abstract class AlleleLikelihoodMatrixMapper { + + public abstract ReadLikelihoods.Matrix map(final ReadLikelihoods.Matrix original); + + /** + * Instantiates a new mapper given an allele-list permutation. + * @param permutation the requested permutation. + * @param the allele type. + * + * @throws IllegalArgumentException if {@code permutation} is {@code null}. + * + * @return never {@code null}. + */ + public static AlleleLikelihoodMatrixMapper newInstance(final AlleleListPermutation permutation) { + if (permutation == null) + throw new IllegalArgumentException("the permutation must not be null"); + if (permutation.isNonPermuted()) + return asIs(); + else + return general(permutation); + } + + /** + * Returns trivial mapper that just maps to the original matrix without changes. + * + * @param the allele type. + * @return never {@code null}. + */ + @SuppressWarnings("unchecked") + private static AlleleLikelihoodMatrixMapper asIs() { + return AS_IS; + } + + @SuppressWarnings("unchecked") + private static final AlleleLikelihoodMatrixMapper AS_IS = new AlleleLikelihoodMatrixMapper() { + @Override + public ReadLikelihoods.Matrix map(final ReadLikelihoods.Matrix original) { + return original; + } + }; + + /** + * Constructs a new mapper instance that work with general permutation without making any assumption. + * @param permutation the permutation to apply to requested matrices wrappers. + * @param allele type. + * @return never {@code null}. + */ + private static AlleleLikelihoodMatrixMapper general(final AlleleListPermutation permutation) { + return new AlleleLikelihoodMatrixMapper() { + @Override + public ReadLikelihoods.Matrix map(final ReadLikelihoods.Matrix original) { + return new ReadLikelihoods.Matrix() { + + @Override + public List reads() { + return original.reads(); + } + + @Override + public List alleles() { + return permutation.toList(); + } + + @Override + public void set(final int alleleIndex, final int readIndex, final double value) { + original.set(permutation.fromIndex(alleleIndex),readIndex,value); + } + + @Override + public double get(final int alleleIndex, final int readIndex) { + return original.get(permutation.fromIndex(alleleIndex),readIndex); + } + + @Override + public int alleleIndex(final A allele) { + return permutation.alleleIndex(allele); + } + + @Override + public int readIndex(GATKSAMRecord read) { + return original.readIndex(read); + } + + @Override + public int alleleCount() { + return permutation.toSize(); + } + + @Override + public int readCount() { + return original.readCount(); + } + + @Override + public A alleleAt(final int alleleIndex) { + return original.alleleAt(permutation.fromIndex(alleleIndex)); + } + + @Override + public GATKSAMRecord readAt(final int readIndex) { + return original.readAt(readIndex); + } + + @Override + public void copyAlleleLikelihoods(final int alleleIndex, final double[] dest, final int offset) { + original.copyAlleleLikelihoods(permutation.fromIndex(alleleIndex),dest,offset); + } + }; + } + }; + } +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingData.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingData.java new file mode 100644 index 000000000..12214d67e --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingData.java @@ -0,0 +1,130 @@ +/* +* 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.gatk.genotyping; + +import htsjdk.variant.variantcontext.Allele; +import org.broadinstitute.gatk.utils.genotyper.ReadLikelihoods; + +/** + * Encapsulates the data use to make the genotype calls. + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public class GenotypingData implements SampleList, AlleleList { + + private final PloidyModel ploidyModel; + + private final ReadLikelihoods likelihoods; + + /** + * Constructs a new genotyping-data collection providing the ploidy model to apply to the input model + * and the read-likelihoods collection. + * + * + * @param ploidyModel the ploidy model. + * @param likelihoods the read-likelihoods collection. + * + * @throws IllegalArgumentException if either {@code ploidyModel} or {@code likelihoods} is {@code null}, + * or they are not compatible in terms of the samples they contain; their lists must match. + */ + public GenotypingData(final PloidyModel ploidyModel, final ReadLikelihoods likelihoods) { + if (ploidyModel == null) + throw new IllegalArgumentException("the ploidy model cannot be null"); + if (likelihoods == null) + throw new IllegalArgumentException("the likelihood object cannot be null"); + this.ploidyModel = ploidyModel; + this.likelihoods = likelihoods; + if (!SampleListUtils.equals(ploidyModel,likelihoods)) + throw new IllegalArgumentException("sample list are different between ploidy-model and read-likelihood collection, perhaps just the order"); + } + + /** + * Returns the ploidy model that corresponds to the data provided. + * @return never {@code null}. + */ + public PloidyModel ploidyModel() { + return ploidyModel; + } + + @Override + public int sampleCount() { + return ploidyModel.sampleCount(); + } + + @Override + public int sampleIndex(final String sample) { + return ploidyModel.sampleIndex(sample); + } + + @Override + public String sampleAt(int sampleIndex) { + return ploidyModel.sampleAt(sampleIndex); + } + + /** + * Returns read-likelihoods to use for genotyping. + * @return never {@code null}. + */ + public ReadLikelihoods readLikelihoods() { + return likelihoods; + } + + @Override + public int alleleCount() { + return likelihoods.alleleCount(); + } + + @Override + public int alleleIndex(final A allele) { + return likelihoods.alleleIndex(allele); + } + + @Override + public A alleleAt(final int index) { + return likelihoods.alleleAt(index); + } +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingLikelihoods.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingLikelihoods.java new file mode 100644 index 000000000..35787107f --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingLikelihoods.java @@ -0,0 +1,159 @@ +/* +* 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.gatk.genotyping; + +import htsjdk.variant.variantcontext.Allele; +import htsjdk.variant.variantcontext.GenotypeLikelihoods; + +import java.util.List; + +/** + * Genotyping Likelihoods collection. + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public class GenotypingLikelihoods implements SampleList, AlleleList { + + private final GenotypeLikelihoods[] likelihoods; + + private final PloidyModel ploidyModel; + + private final AlleleList alleles; + + /** + * Creates a new genotyping-likelihoods collection given the genotype alleles, the sample ploidy model and the + * likelihoods themselves. + *

+ * Notice that this constructor does not check whether the likelihood array lengths corresponds to the sample plodies and + * number of alleles. + *

+ * + * @param alleles the genotyping alleles. + * @param ploidyModel the ploidy model. + * @param likelihoods the actual genotype likelihoods, one element per sample. + * + * @throws IllegalArgumentException if any argument is {@code null}, or the number of samples in {@code ploidyModel} + * does not correspond with the number of likelihoods arrays in {@code likelihoods} + */ + GenotypingLikelihoods(final AlleleList alleles, final PloidyModel ploidyModel, + final List likelihoods) { + if (alleles == null) + throw new IllegalArgumentException("allele list cannot be null"); + if (ploidyModel == null) + throw new IllegalArgumentException("the ploidy model cannot be null"); + if (likelihoods == null) + throw new IllegalArgumentException("the likelihood collection cannot be null"); + if (ploidyModel.sampleCount() != likelihoods.size()) + throw new IllegalArgumentException("there must be exactly one likelihood set for each sample"); + + this.likelihoods = likelihoods.toArray(new GenotypeLikelihoods[likelihoods.size()]); + for (final GenotypeLikelihoods likelihood : this.likelihoods) + if (likelihood == null) + throw new IllegalArgumentException("no genotype likelihood is allowed to be null"); + + this.alleles = alleles; + this.ploidyModel = ploidyModel; + } + + @Override + public int sampleCount() { + return ploidyModel.sampleCount(); + } + + @Override + public int sampleIndex(final String sample) { + return ploidyModel.sampleIndex(sample); + } + + @Override + public String sampleAt(final int sampleIndex) { + return ploidyModel.sampleAt(sampleIndex); + } + + /** + * Returns the ploidy of the sample given its index in the collection. + * + * @param sampleIndex the query sample index. + * + * @throws IllegalArgumentException if {@code sampleIndex} is not a valid index for this collection: + * [0,{@link #sampleCount()). + * + * @return 0 or greater. + */ + public int samplePloidy(final int sampleIndex) { + return ploidyModel.samplePloidy(sampleIndex); + } + + /** + * Returns the genotype-likelihoods of the sample given its index in the collection. + * + * @param sampleIndex the query sample index. + * + * @throws IllegalArgumentException if {@code sampleIndex} is not a valid index for this collection: + * [0,{@link #sampleCount()). + * + * @return never {@code null}. + */ + public GenotypeLikelihoods sampleLikelihoods(final int sampleIndex) { + return likelihoods[sampleIndex]; + } + + @Override + public int alleleCount() { + return alleles.alleleCount(); + } + + @Override + public int alleleIndex(final A allele) { + return alleles.alleleIndex(allele); + } + + @Override + public A alleleAt(final int index) { + return alleles.alleleAt(index); + } +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingModel.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingModel.java new file mode 100644 index 000000000..9c84c8962 --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/GenotypingModel.java @@ -0,0 +1,75 @@ +/* +* 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.gatk.genotyping; + +import htsjdk.variant.variantcontext.Allele; + +/** + * Common interface for genotyping models. + * + * Given a plo + * + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public interface GenotypingModel { + + /** + * Calculate genotype likelihoods given the genotype data and the set of + * alleles to genotype upon. + * + * @param genotypingAlleles the target alleles. + * @param data the data (read-likelihoods and ploidy) to genotype + * @param the allele type. + * + * @throws IllegalArgumentException if {@code genotypingData} or {@code genotypingAlleles} is {@code null}, + * or {@code genotypingData} does not cover the requested alleles. + * + * @return never {@code null}. + */ + public GenotypingLikelihoods calculateLikelihoods(final AlleleList genotypingAlleles, final GenotypingData data); +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/HomogeneousPloidyModel.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/HomogeneousPloidyModel.java new file mode 100644 index 000000000..ed8d59a94 --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/HomogeneousPloidyModel.java @@ -0,0 +1,116 @@ +/* +* 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.gatk.genotyping; + +/** +* {@link PloidyModel} implementation tailored to work with a homogeneous constant ploidy +* across samples and positions. +* +* @author Valentin Ruano-Rubio <valentin@broadinstitute.org> +*/ +public class HomogeneousPloidyModel implements PloidyModel, SampleList { + + private SampleList sampleList; + + private final int ploidy; + + /** + * Constructs a homogeneous ploidy model given the sample list and ploidy. + * + * @param samples the sample list. + * @param ploidy the common ploidy for all samples in {@code samples}. + * + * @throws IllegalArgumentException if {@code samples} is {@code null}, + * or ploidy is 0 or less. + */ + public HomogeneousPloidyModel(final SampleList samples, final int ploidy) { + if (ploidy <= 0) + throw new IllegalArgumentException("does not support negative ploidy"); + this.ploidy = ploidy; + + sampleList = samples; + } + + @Override + public int sampleCount() { + return sampleList.sampleCount(); + } + + @Override + public String sampleAt(final int index) { + return sampleList.sampleAt(index); + } + + @Override + public int sampleIndex(final String sample) { + return sampleList.sampleIndex(sample); + } + + @Override + public int samplePloidy(final int sampleIndex) { + checkSampleIndex(sampleIndex); + return ploidy; + } + + private void checkSampleIndex(final int sampleIndex) { + if (sampleIndex < 0) + throw new IllegalArgumentException("the sample index cannot be negative: " + sampleIndex); + if (sampleIndex >= sampleList.sampleCount()) + throw new IllegalArgumentException("the sample index is equal or larger than the sample count: " + sampleIndex + " >= " + sampleList.sampleCount()); + } + + @Override + public boolean isHomogeneous() { + return true; + } + + @Override + public int totalPloidy() { + return ploidy * sampleList.sampleCount(); + } + +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/InfiniteRandomMatingPopulationModel.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/InfiniteRandomMatingPopulationModel.java new file mode 100644 index 000000000..6b0594ee0 --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/InfiniteRandomMatingPopulationModel.java @@ -0,0 +1,147 @@ +/* +* 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.gatk.genotyping; + +import htsjdk.variant.variantcontext.Allele; +import htsjdk.variant.variantcontext.GenotypeLikelihoods; +import org.broadinstitute.gatk.utils.genotyper.ReadLikelihoods; + +import java.util.ArrayList; +import java.util.Collections; +import java.util.List; + +/** + * The infinite-population-genotyping assumes that samples belong to individuals taken at random + * from a very large population that mate at random. + *

+ * Consequently genotypes calls between samples are totally independent conditional to the frequencies in + * the population they coming from. And genotypes should exhibit the ratios expected under HWE. + *

+ * Therefore each sample genotype likelihoods can be considered to + * be independent from all other samples. + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public class InfiniteRandomMatingPopulationModel implements GenotypingModel { + + @Override + public GenotypingLikelihoods calculateLikelihoods(final AlleleList genotypingAlleles, final GenotypingData data) { + if (genotypingAlleles == null) + throw new IllegalArgumentException("the allele cannot be null"); + if (data == null) + throw new IllegalArgumentException("the genotyping data cannot be null"); + + final AlleleListPermutation permutation = AlleleListUtils.permutation(data,genotypingAlleles); + final AlleleLikelihoodMatrixMapper alleleLikelihoodMatrixMapper = AlleleLikelihoodMatrixMapper.newInstance(permutation); + + final int sampleCount = data.sampleCount(); + + switch (sampleCount) { + case 0: return noSampleLikelihoods(permutation,data); + case 1: return singleSampleLikelihoods(genotypingAlleles,data,alleleLikelihoodMatrixMapper); + default: + final PloidyModel ploidyModel = data.ploidyModel(); + return ploidyModel.isHomogeneous() ? multiSampleHomogeneousPloidyModelLikelihoods(genotypingAlleles, data, alleleLikelihoodMatrixMapper, sampleCount, ploidyModel) + : multiSampleHeterogeneousPloidyModelLikelihoods(genotypingAlleles, data, alleleLikelihoodMatrixMapper, sampleCount, ploidyModel); + } + } + + private GenotypingLikelihoods noSampleLikelihoods(final AlleleList genotypingAlleles, + final GenotypingData data) { + @SuppressWarnings("unchecked") + final List likelihoods = Collections.EMPTY_LIST; + return new GenotypingLikelihoods<>(genotypingAlleles,data.ploidyModel(), likelihoods); + + } + + private GenotypingLikelihoods singleSampleLikelihoods(final AlleleList genotypingAlleles, + final GenotypingData data, + final AlleleLikelihoodMatrixMapper alleleLikelihoodMatrixMapper) { + final PloidyModel ploidyModel = data.ploidyModel(); + final int samplePloidy = ploidyModel.samplePloidy(0); + final int alleleCount = genotypingAlleles.alleleCount(); + final GenotypeLikelihoodCalculator likelihoodsCalculator = + GenotypeLikelihoodCalculator.getInstance(samplePloidy,alleleCount); + final ReadLikelihoods.Matrix sampleLikelihoods = alleleLikelihoodMatrixMapper.map(data.readLikelihoods().sampleMatrix(0)); + final List genotypeLikelihoods = Collections.singletonList(likelihoodsCalculator.genotypeLikelihoods(sampleLikelihoods)); + return new GenotypingLikelihoods<>(genotypingAlleles,ploidyModel,genotypeLikelihoods); + } + + private GenotypingLikelihoods multiSampleHeterogeneousPloidyModelLikelihoods(final AlleleList genotypingAlleles, + final GenotypingData data, + final AlleleLikelihoodMatrixMapper alleleLikelihoodMatrixMapper, + final int sampleCount, + final PloidyModel ploidyModel) { + final List genotypeLikelihoods = new ArrayList<>(sampleCount); + final int alleleCount = genotypingAlleles.alleleCount(); + for (int i = 0; i < sampleCount; i++) { + final int samplePloidy = ploidyModel.samplePloidy(i); + final GenotypeLikelihoodCalculator likelihoodsCalculator = + GenotypeLikelihoodCalculator.getInstance(samplePloidy,alleleCount); + final ReadLikelihoods.Matrix sampleLikelihoods = alleleLikelihoodMatrixMapper.map(data.readLikelihoods().sampleMatrix(i)); + genotypeLikelihoods.add(likelihoodsCalculator.genotypeLikelihoods(sampleLikelihoods)); + } + return new GenotypingLikelihoods<>(genotypingAlleles,ploidyModel,genotypeLikelihoods); + } + + private GenotypingLikelihoods multiSampleHomogeneousPloidyModelLikelihoods(final AlleleList genotypingAlleles, + final GenotypingData data, + final AlleleLikelihoodMatrixMapper alleleLikelihoodMatrixMapper, + final int sampleCount, + final PloidyModel ploidyModel) { + final int samplePloidy = ploidyModel.samplePloidy(0); + final List genotypeLikelihoods = new ArrayList<>(sampleCount); + final int alleleCount = genotypingAlleles.alleleCount(); + final GenotypeLikelihoodCalculator likelihoodsCalculator = + GenotypeLikelihoodCalculator.getInstance(samplePloidy,alleleCount); + for (int i = 0; i < sampleCount; i++) { + final ReadLikelihoods.Matrix sampleLikelihoods = alleleLikelihoodMatrixMapper.map(data.readLikelihoods().sampleMatrix(i)); + genotypeLikelihoods.add(likelihoodsCalculator.genotypeLikelihoods(sampleLikelihoods)); + } + return new GenotypingLikelihoods<>(genotypingAlleles,ploidyModel,genotypeLikelihoods); + } +} diff --git a/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/PloidyModel.java b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/PloidyModel.java new file mode 100644 index 000000000..8ca00611e --- /dev/null +++ b/protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/genotyping/PloidyModel.java @@ -0,0 +1,80 @@ +/* +* 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.gatk.genotyping; + +/** + * Information about the number of chromosome per sample at a given location. + * + * @author Valentin Ruano-Rubio <valentin@broadinstitute.org> + */ +public interface PloidyModel extends SampleList { + + /** + * Return the assumed ploidy for a sample given its index. + * + * @param sampleIndex target sample index. + * @return 0 or greater. + */ + public int samplePloidy(final int sampleIndex); + + /** + * Checks whether the ploidy is homogeneous across all samples. + * + * @return {@code true} if all samples has the same ploidy. + */ + public boolean isHomogeneous(); + + /** + * Sum of all ploidy across all samples. + *

+ * It must match the sum of all ploidies across samples. + *

+ * + * @return 0 or greater. + */ + public int totalPloidy(); +}