Merge pull request #292 from broadinstitute/vrr_analyzeCovariates
Added the AnalyzeCovariates tool to generate BQSR quality assessment plots.
This commit is contained in:
Eric Banks
commit
74415a6a2a
|
|
@ -0,0 +1,583 @@
|
|||
/*
|
||||
* 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.bqsr;
|
||||
|
||||
import com.google.java.contract.Requires;
|
||||
import org.broadinstitute.sting.commandline.Argument;
|
||||
import org.broadinstitute.sting.commandline.Input;
|
||||
import org.broadinstitute.sting.commandline.Output;
|
||||
import org.broadinstitute.sting.gatk.contexts.AlignmentContext;
|
||||
import org.broadinstitute.sting.gatk.contexts.ReferenceContext;
|
||||
import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker;
|
||||
import org.broadinstitute.sting.gatk.walkers.RodWalker;
|
||||
import org.broadinstitute.sting.utils.Utils;
|
||||
import org.broadinstitute.sting.utils.exceptions.UserException;
|
||||
import org.broadinstitute.sting.utils.recalibration.RecalUtils;
|
||||
import org.broadinstitute.sting.utils.recalibration.RecalibrationReport;
|
||||
import org.broadinstitute.sting.utils.recalibration.BaseRecalibration;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.IOException;
|
||||
import java.util.LinkedHashMap;
|
||||
import java.util.Map;
|
||||
|
||||
|
||||
/**
|
||||
* Tool to analyze and evaluate base recalibration ables.
|
||||
* <p/>
|
||||
* For now it generates a plot report to assess the quality of a recalibration.
|
||||
*
|
||||
* <h3>Input</h3>
|
||||
*
|
||||
* The tool can take up to three different sets of recalibration tables.
|
||||
* The resulting plots will be overlaid on top of each other to make
|
||||
* comparisons easy.
|
||||
*
|
||||
* <table style="text-align: left">
|
||||
* <thead>
|
||||
* <tr><th>Set</th><th>Argument</th><th>Label</th><th>Color</th><th>Description</th></tr>
|
||||
* </thead>
|
||||
* <tbody>
|
||||
* <tr><td>Original</td><td>-before</td><td>BEFORE</td><td style="color: #ff34b3">Maroon1</td>
|
||||
* <td>First pass recalibration
|
||||
* tables obtained from applying {@link BaseRecalibration}
|
||||
* on the original alignment.</td></tr>
|
||||
* <tr><td>Recalibrated</td><td>-after</td><td>AFTER</td><td style="color: #0000ff">Blue</td>
|
||||
* <td>Second pass recalibration tables
|
||||
* results from the application of {@link BaseRecalibration}
|
||||
* on the alignment recalibrated using the first pass tables</td></tr>
|
||||
* <tr><td>Input</td><td>-BQSR</td><td>BQSR</td><td style="color: #000000">Black</td>
|
||||
* <td>Any recalibration table without a specific role</td></tr>
|
||||
* </tbody>
|
||||
* </table>
|
||||
* <br/>
|
||||
*
|
||||
* You need to specify one set at least. Multiple sets need to have the same values for the following parameters:
|
||||
* <br/></br>
|
||||
* <i>covariate (order is not important), no_standard_covs, run_without_dbsnp, solid_recal_mode,
|
||||
* solid_nocall_strategy, mismatches_context_size, mismatches_default_quality, deletions_default_quality,
|
||||
* insertions_default_quality, maximum_cycle_value, low_quality_tail, default_platform, force_platform,
|
||||
* quantizing_levels</i> and <i>binary_tag_name</i>
|
||||
* <h3>Output</h3>
|
||||
*
|
||||
* Currently this tool generates two outputs:
|
||||
*
|
||||
* <dl>
|
||||
* <dt style="font-weight: normal">-plots <i>my-report.pdf</i></dt>
|
||||
* <dd>A pdf document that encloses plots to assess the quality of the recalibration.</dd>
|
||||
* <dt style="font-weight: normal">-csv <i>my-report.csv</i></dt>
|
||||
* <dd>A csv file that contains a table with all the data required to generate those plots.</dd>
|
||||
* </dl>
|
||||
*
|
||||
* You need to specify at least one of them.
|
||||
*
|
||||
* <h3>Other Arguments</h3>
|
||||
*
|
||||
* <h4>-ignoreLMT, --ignoreLastModificationTimes</h4>
|
||||
*
|
||||
* when set, no warning message will be displayed in the -before recalibration table file is older than the -after one.
|
||||
*
|
||||
* <h3>Examples</h3>
|
||||
*
|
||||
*
|
||||
* <h4>Plot a single recalibration table</h4>
|
||||
* <pre>
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T AnalyzeCovariates \
|
||||
* -R myrefernce.fasta \
|
||||
* -BQSR myrecal.table \
|
||||
* -plots BQSR.pdf
|
||||
* </pre>
|
||||
*
|
||||
* <h4>Plot before (first pass) and after (second pass) recalibration table to compare them</h4>
|
||||
*
|
||||
* <pre>
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T AnalyzeCovariates \
|
||||
* -R myrefernce.fasta \
|
||||
* -before recal2.table \
|
||||
* -after recal3.table \
|
||||
* -plots recalQC.pdf
|
||||
* </pre>
|
||||
*
|
||||
* <h4>Plot up to three recalibration tables for comparison</h4>
|
||||
*
|
||||
* <pre>
|
||||
*
|
||||
* # You can ignore the before/after semantics completely if you like (if you do add -ignoreLMT
|
||||
* # to avoid a possible warning), but all tables should have been generated using the same parameters.
|
||||
*
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T AnalyzeCovariates \
|
||||
* -R myrefernce.fasta \
|
||||
* -ignoreLMT \
|
||||
* -BQSR recal1.table \ # you can discard any two
|
||||
* -before recal2.table \
|
||||
* -after recal3.table \
|
||||
* -plots myrecals.pdf
|
||||
* </pre>
|
||||
*
|
||||
* <h4>Full BQSR quality assessment pipeline</h4>
|
||||
*
|
||||
* <pre>
|
||||
* # Generate the first pass recalibration table file.
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T BaseRecalibrator \
|
||||
* -R myreference.fasta \
|
||||
* -I myinput.bam \
|
||||
* -knownSites bundle/my-trusted-snps.vcf \ # optional but recommendable
|
||||
* -knownSites bundle/my-trusted-indels.vcf \ # optional but recommendable
|
||||
* ... other options
|
||||
* -o firstpass.table
|
||||
*
|
||||
* # Generate the second pass recalibration table file.
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T BaseRecalibrator \
|
||||
* -BQSR firstpass.table \
|
||||
* -R myreference.fasta \
|
||||
* -I myinput.bam \
|
||||
* -knownSites bundle/my-trusted-snps.vcf \
|
||||
* -knownSites bundle/my-trusted-indels.vcf \
|
||||
* ... other options \
|
||||
* -o secondpass.table
|
||||
*
|
||||
* # Finally generate the plots report and also keep a copy of the csv (optional).
|
||||
* java -jar GenomeAnalysisTK.jar \
|
||||
* -T AnalyzeCovariates \
|
||||
* -R myrefernce.fasta \
|
||||
* -before firstpass.table \
|
||||
* -after secondpass.table \
|
||||
* -csv BQSR.csv \ # optional
|
||||
* -plots BQSR.pdf
|
||||
* </pre>
|
||||
*
|
||||
* @author Valentin Ruano-Rubio <valentin@broadinstitute.org>
|
||||
* @version 6/16/2013
|
||||
* @since 2.6
|
||||
*/
|
||||
public final class AnalyzeCovariates extends RodWalker<AnalyzeCovariates.None,AnalyzeCovariates.None> {
|
||||
|
||||
|
||||
// Constants on option short names that are used in some error/warning messages:
|
||||
|
||||
static final String CSV_ARG_SHORT_NAME = "csv";
|
||||
static final String PDF_ARG_SHORT_NAME = "plots";
|
||||
static final String BEFORE_ARG_SHORT_NAME = "before";
|
||||
static final String AFTER_ARG_SHORT_NAME = "after";
|
||||
|
||||
/**
|
||||
* File containing the recalibration tables from the first pass.
|
||||
*/
|
||||
@Input(shortName=BEFORE_ARG_SHORT_NAME,fullName="beforeReportFile", doc = "file containing the BQSR first-pass report file",required = false)
|
||||
protected File beforeFile = null;
|
||||
|
||||
/**
|
||||
* File containing the recalibration tables from the second pass.
|
||||
*/
|
||||
@Input(shortName=AFTER_ARG_SHORT_NAME, fullName="afterReportFile", doc = "file containing the BQSR second-pass report file",required = false)
|
||||
protected File afterFile = null;
|
||||
|
||||
/**
|
||||
* If true, it won't show a warning if the last-modification time of the before and after input files suggest that they have been reversed.
|
||||
*/
|
||||
@Argument(shortName="ignoreLMT", fullName="ignoreLastModificationTimes", doc= "do not emit warning messages related to suspicious last modification time order of inputs", required = false)
|
||||
protected boolean ignoreLastModificationTime = false;
|
||||
|
||||
/**
|
||||
* Output report file name.
|
||||
*/
|
||||
@Output(shortName=PDF_ARG_SHORT_NAME, fullName="plotsReportFile" ,doc = "location of the output report", required = false)
|
||||
protected File pdfFile = null;
|
||||
|
||||
/**
|
||||
* Output csv file name.
|
||||
*/
|
||||
@Output(shortName=CSV_ARG_SHORT_NAME,fullName="intermediateCsvFile" ,doc = "location of the csv intermediate file", required = false)
|
||||
protected File csvFile = null;
|
||||
|
||||
/**
|
||||
* Convenience reference to the RECAL_BQSR_FILE argument value.
|
||||
* <p/>
|
||||
* This field value is resolved by {@link #initialize()}.
|
||||
*/
|
||||
protected File bqsrFile = null;
|
||||
|
||||
/**
|
||||
* Checks inputs and argument values.
|
||||
* <p/>
|
||||
* Notice that this routine will not validate the content of files. It may have some minor side effects as
|
||||
* the output of warning messages back to the user.
|
||||
*
|
||||
* @throw IllegalStateException there is some required argument value that has not been loaded yet.
|
||||
* @throw UserException if there is some error caused by or under the end user's control.
|
||||
*/
|
||||
private void checkArgumentsValues() {
|
||||
checkInputReportFile("BQSR",bqsrFile);
|
||||
checkInputReportFile("before",beforeFile);
|
||||
checkInputReportFile("after",afterFile);
|
||||
if (bqsrFile == null && beforeFile == null && afterFile == null) {
|
||||
throw new UserException("you must provide at least one recalibration report file "
|
||||
+ "(arguments -BQSR, -" + BEFORE_ARG_SHORT_NAME + " or -" + AFTER_ARG_SHORT_NAME);
|
||||
}
|
||||
|
||||
checkOutputFile(PDF_ARG_SHORT_NAME,pdfFile);
|
||||
checkOutputFile(CSV_ARG_SHORT_NAME, csvFile);
|
||||
checkInputReportFileLMT(beforeFile,afterFile);
|
||||
checkOutputRequested();
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks whether the last-modification-time of the inputs is consistent with their relative roles.
|
||||
*
|
||||
* This routine does not thrown an exception but may output a warning message if inconsistencies are spotted.
|
||||
*
|
||||
* @param beforeFile the before report file.
|
||||
* @param afterFile the after report file.
|
||||
*/
|
||||
private void checkInputReportFileLMT(final File beforeFile, final File afterFile) {
|
||||
|
||||
if (ignoreLastModificationTime || beforeFile == null || afterFile == null) {
|
||||
return; // nothing to do here
|
||||
} else if (beforeFile.lastModified() > afterFile.lastModified()) {
|
||||
Utils.warnUser("Last modification timestamp for 'Before' and 'After'"
|
||||
+ "recalibration reports are in the wrong order. Perhaps, have they been swapped?");
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks that at least one output was requested.
|
||||
*
|
||||
* @throw UserException if no output was requested.
|
||||
*/
|
||||
private void checkOutputRequested() {
|
||||
if (pdfFile == null && csvFile == null) {
|
||||
throw new UserException("you need to request at least one output:"
|
||||
+ " the intermediate csv file (-" + CSV_ARG_SHORT_NAME + " FILE)"
|
||||
+ " or the final plot file (-" + PDF_ARG_SHORT_NAME + " FILE).");
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks the value provided to input file arguments.
|
||||
*
|
||||
* @throw UserException if there is any problem cause by or under the end user's control
|
||||
*
|
||||
* @param name command line argument short name.
|
||||
* @param value the argument value.
|
||||
*/
|
||||
private void checkInputReportFile(final String name,final File value) {
|
||||
if (value == null) {
|
||||
return;
|
||||
} else if (!value.exists()) {
|
||||
throw new UserException.BadArgumentValue(name, "input report '" +
|
||||
value + "' does not exist or is unreachable");
|
||||
} else if (!value.isFile()) {
|
||||
throw new UserException.BadArgumentValue(name, "input report '" +
|
||||
value + "' is not a regular file");
|
||||
} else if (!value.canRead()) {
|
||||
throw new UserException.BadArgumentValue(name, "input report '" +
|
||||
value + "' cannot be read");
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks the value provided for output arguments.
|
||||
*
|
||||
* @throw UserException if there is any problem cause by or under the end user's control
|
||||
*
|
||||
* @param name command line argument short name.
|
||||
* @param value the argument value.
|
||||
*/
|
||||
private void checkOutputFile(final String name, final File value) {
|
||||
if (value == null) {
|
||||
return;
|
||||
}
|
||||
if (value.exists() && !value.isFile()) {
|
||||
throw new UserException.BadArgumentValue(name, "the output file location '"
|
||||
+ value + "' exists as not a file");
|
||||
}
|
||||
final File parent = value.getParentFile();
|
||||
if (parent == null) {
|
||||
return;
|
||||
}
|
||||
if (!parent.exists()) {
|
||||
throw new UserException.BadArgumentValue(name, "the output file parent directory '"
|
||||
+ parent + "' does not exists or is unreachable");
|
||||
} else if (!parent.isDirectory()) {
|
||||
throw new UserException.BadArgumentValue(name, "the output file parent directory '"
|
||||
+ parent + "' is not a directory");
|
||||
} else if (!parent.canWrite()) {
|
||||
throw new UserException.BadArgumentValue(name, "the output file parent directory '"
|
||||
+ parent + "' cannot be written");
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Generates the plots using the external R script.
|
||||
*
|
||||
* <p/>
|
||||
* If <code>plotsFile</code> is <code>null</code>, it does not perform any plotting.
|
||||
*
|
||||
* @param csvFile the intermediary csv file.
|
||||
* @param plotsFile the output plot location.
|
||||
*/
|
||||
private void generatePlots(final File csvFile, final Map<String,File> reportFiles, final File plotsFile) {
|
||||
|
||||
if (plotsFile == null) {
|
||||
return;
|
||||
}
|
||||
logger.info("Generating plots file '" + plotsFile + "'");
|
||||
final File exampleReportFile = reportFiles.values().iterator().next();
|
||||
RecalUtils.generatePlots(csvFile,exampleReportFile,plotsFile);
|
||||
}
|
||||
|
||||
@Override
|
||||
public void initialize() {
|
||||
super.initialize();
|
||||
bqsrFile = getToolkit().getArguments().BQSR_RECAL_FILE;
|
||||
checkArgumentsValues();
|
||||
final Map<String, File> reportFiles = buildReportFileMap();
|
||||
final Map<String, RecalibrationReport> reports = buildReportMap(reportFiles);
|
||||
checkReportConsistency(reports);
|
||||
final File csvFile = resolveCsvFile();
|
||||
generateCsvFile(csvFile,reports);
|
||||
final File plotFile = resolvePlotFile();
|
||||
generatePlots(csvFile, reportFiles, plotFile);
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the plot output file
|
||||
* @return might be <code>null</code> if the user has not indicated and output file.
|
||||
*/
|
||||
private File resolvePlotFile() {
|
||||
return pdfFile;
|
||||
}
|
||||
|
||||
/**
|
||||
* Generates the intermediary Csv file.
|
||||
*
|
||||
* @param csvFile where to write the file.
|
||||
* @param reports the reports to be included.
|
||||
*/
|
||||
private void generateCsvFile(final File csvFile, final Map<String, RecalibrationReport> reports) {
|
||||
try {
|
||||
logger.info("Generating csv file '" + csvFile + "'");
|
||||
RecalUtils.generateCsv(csvFile, reports);
|
||||
} catch (FileNotFoundException e) {
|
||||
throw new UserException(
|
||||
String.format("There is a problem creating the intermediary Csv file '%s': %s",
|
||||
csvFile,e.getMessage()),e);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks whether multiple input recalibration report files argument values are consistent (equal).
|
||||
*
|
||||
* @param reports map with report to verify.
|
||||
*
|
||||
* @throw UserException if there is any inconsistency.
|
||||
*/
|
||||
private void checkReportConsistency(final Map<String, RecalibrationReport> reports) {
|
||||
final Map.Entry<String,RecalibrationReport>[] reportEntries =
|
||||
reports.entrySet().toArray((Map.Entry<String,RecalibrationReport>[]) new Map.Entry[reports.size()]);
|
||||
|
||||
final Map.Entry<String,RecalibrationReport> exampleEntry = reportEntries[0];
|
||||
|
||||
for (int i = 1; i < reportEntries.length; i++) {
|
||||
final Map<String,? extends CharSequence> diffs = exampleEntry.getValue().getRAC().compareReportArguments(
|
||||
reportEntries[i].getValue().getRAC(),exampleEntry.getKey(),reportEntries[i].getKey());
|
||||
if (diffs.size() != 0) {
|
||||
throw new UserException.IncompatibleRecalibrationTableParameters("There are differences in relevant arguments of"
|
||||
+ " two or more input recalibration reports. Please make sure"
|
||||
+ " they have been created using the same recalibration parameters."
|
||||
+ " " + Utils.join("// ", reportDifferencesStringArray(diffs)));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Creates a map with all input recalibration files indexed by their "role".
|
||||
* <p/>
|
||||
* The key is the role and the value the corresponding report file.
|
||||
* <p/>
|
||||
* Roles: "Before" (recalibration), "After" (recalibration), "BQSR" (the tool standard argument recalibration file)
|
||||
*
|
||||
* @return never <code>null</code>
|
||||
*/
|
||||
private Map<String, File> buildReportFileMap() {
|
||||
final Map<String,File> reports = new LinkedHashMap<>(3);
|
||||
if (bqsrFile != null) {
|
||||
reports.put("BQSR",bqsrFile);
|
||||
}
|
||||
if (beforeFile != null) {
|
||||
reports.put("Before",beforeFile);
|
||||
}
|
||||
if (afterFile != null) {
|
||||
reports.put("After",afterFile);
|
||||
}
|
||||
return reports;
|
||||
}
|
||||
|
||||
/**
|
||||
* Transforms a recalibration file map into a report object map.
|
||||
*
|
||||
* @param reportFileMap the file map to transforms.
|
||||
* @return never <code>null</code>, a new map with the same size as
|
||||
* <code>reportFileMap</code> and the same key set.
|
||||
*/
|
||||
@Requires("reportFileMap != null")
|
||||
private Map<String, RecalibrationReport> buildReportMap(final Map<String, File> reportFileMap) {
|
||||
final Map<String,RecalibrationReport> reports = new LinkedHashMap<>(reportFileMap.size());
|
||||
for (final Map.Entry<String,File> e : reportFileMap.entrySet()) {
|
||||
reports.put(e.getKey(),new RecalibrationReport(e.getValue()));
|
||||
}
|
||||
return reports;
|
||||
}
|
||||
|
||||
/**
|
||||
* Generates a flatter String array representation of recalibration argument differences.
|
||||
* @param diffs the differences to represent.
|
||||
*
|
||||
* @return never <code>null</code>, an array of the same length as the size of the input <code>diffs</code>.
|
||||
*/
|
||||
@Requires("diffs != null")
|
||||
private String[] reportDifferencesStringArray(final Map<String, ? extends CharSequence> diffs) {
|
||||
final String[] result = new String[diffs.size()];
|
||||
int i = 0;
|
||||
for (final Map.Entry<String, ? extends CharSequence> e : diffs.entrySet()) {
|
||||
result[i++] = capitalize(e.getKey()) + ": " + e.getValue();
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the input string capitalizing the first letter.
|
||||
*
|
||||
* @param str the string to capitalize
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
@Requires("str != null")
|
||||
private String capitalize(final String str) {
|
||||
if (str.isEmpty()) {
|
||||
return str;
|
||||
} else {
|
||||
return Character.toUpperCase(str.charAt(0)) + str.substring(1);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the csv file to use.
|
||||
* <p/>
|
||||
* This is the the one specified by the user if any or a temporary file
|
||||
* that will be deleted as soon as the VM exists by default.
|
||||
*
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
private File resolveCsvFile() {
|
||||
if (csvFile != null) {
|
||||
return csvFile;
|
||||
} else {
|
||||
try {
|
||||
final File result = File.createTempFile("AnalyzeCovariates", ".csv");
|
||||
result.deleteOnExit();
|
||||
return result;
|
||||
} catch (IOException e) {
|
||||
throw new UserException("Could not create temporary Csv file",e);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Always return true, forcing the immediate termination of the travesal.
|
||||
* @return
|
||||
*/
|
||||
@Override
|
||||
public boolean isDone() {
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* {@inheritDoc}
|
||||
*/
|
||||
@Override
|
||||
public None reduceInit() {
|
||||
return new None();
|
||||
}
|
||||
|
||||
/**
|
||||
* Is not supposed to ever be called, thus it always results in an exception.
|
||||
*
|
||||
* @throws IllegalStateException always.
|
||||
*/
|
||||
@Override
|
||||
public None reduce(None value, None sum) {
|
||||
throw new IllegalStateException("AnalyzeCovariates reduce method is not supposed to be invoked ever");
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Is not supposed to ever be called, thus it always results in an exception.
|
||||
*
|
||||
* @throws IllegalStateException always.
|
||||
*/
|
||||
@Override
|
||||
public None map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
|
||||
throw new IllegalStateException("AnalyzeCovariates map method is not supposed to be invoked ever");
|
||||
}
|
||||
|
||||
/**
|
||||
* Dummy map and reduce types for the {@link AnalyzeCovariates} tool that in fact does not do any traversal.
|
||||
*/
|
||||
protected static class None {
|
||||
private None() {
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
|
|
@ -92,18 +92,6 @@ public class BQSRGatherer extends Gatherer {
|
|||
|
||||
generalReport.calculateQuantizedQualities();
|
||||
|
||||
RecalibrationArgumentCollection RAC = generalReport.getRAC();
|
||||
if ( RAC.RECAL_PDF_FILE != null ) {
|
||||
RAC.RECAL_TABLE_FILE = output;
|
||||
if ( RAC.existingRecalibrationReport != null ) {
|
||||
final RecalibrationReport originalReport = new RecalibrationReport(RAC.existingRecalibrationReport);
|
||||
RecalUtils.generateRecalibrationPlot(RAC, originalReport.getRecalibrationTables(), generalReport.getRecalibrationTables(), generalReport.getCovariates());
|
||||
}
|
||||
else {
|
||||
RecalUtils.generateRecalibrationPlot(RAC, generalReport.getRecalibrationTables(), generalReport.getCovariates());
|
||||
}
|
||||
}
|
||||
|
||||
generalReport.output(outputFile);
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -61,6 +61,7 @@ import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker;
|
|||
import org.broadinstitute.sting.gatk.walkers.*;
|
||||
import org.broadinstitute.sting.utils.MathUtils;
|
||||
import org.broadinstitute.sting.utils.BaseUtils;
|
||||
import org.broadinstitute.sting.utils.Utils;
|
||||
import org.broadinstitute.sting.utils.baq.BAQ;
|
||||
import org.broadinstitute.sting.utils.clipping.ReadClipper;
|
||||
import org.broadinstitute.sting.utils.collections.Pair;
|
||||
|
|
@ -124,7 +125,7 @@ import java.util.List;
|
|||
* -R resources/Homo_sapiens_assembly18.fasta \
|
||||
* -knownSites bundle/hg18/dbsnp_132.hg18.vcf \
|
||||
* -knownSites another/optional/setOfSitesToMask.vcf \
|
||||
* -o recal_data.grp
|
||||
* -o recal_data.table
|
||||
* </pre>
|
||||
*/
|
||||
|
||||
|
|
@ -516,11 +517,6 @@ public class BaseRecalibrator extends ReadWalker<Long, Long> implements NanoSche
|
|||
generateReport();
|
||||
logger.info("...done!");
|
||||
|
||||
if ( RAC.RECAL_PDF_FILE != null ) {
|
||||
logger.info("Generating recalibration plots...");
|
||||
generatePlots();
|
||||
}
|
||||
|
||||
logger.info("BaseRecalibrator was able to recalibrate " + result + " reads");
|
||||
}
|
||||
|
||||
|
|
@ -528,16 +524,6 @@ public class BaseRecalibrator extends ReadWalker<Long, Long> implements NanoSche
|
|||
return recalibrationEngine.getFinalRecalibrationTables();
|
||||
}
|
||||
|
||||
private void generatePlots() {
|
||||
File recalFile = getToolkit().getArguments().BQSR_RECAL_FILE;
|
||||
if (recalFile != null) {
|
||||
RecalibrationReport report = new RecalibrationReport(recalFile);
|
||||
RecalUtils.generateRecalibrationPlot(RAC, report.getRecalibrationTables(), getRecalibrationTable(), requestedCovariates);
|
||||
}
|
||||
else
|
||||
RecalUtils.generateRecalibrationPlot(RAC, getRecalibrationTable(), requestedCovariates);
|
||||
}
|
||||
|
||||
/**
|
||||
* go through the quality score table and use the # observations and the empirical quality score
|
||||
* to build a quality score histogram for quantization. Then use the QuantizeQual algorithm to
|
||||
|
|
|
|||
|
|
@ -46,15 +46,17 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.bqsr;
|
||||
|
||||
import com.google.java.contract.Requires;
|
||||
import org.broad.tribble.Feature;
|
||||
import org.broadinstitute.sting.commandline.*;
|
||||
import org.broadinstitute.sting.gatk.report.GATKReportTable;
|
||||
import org.broadinstitute.sting.utils.Utils;
|
||||
import org.broadinstitute.sting.utils.exceptions.StingException;
|
||||
import org.broadinstitute.sting.utils.recalibration.RecalUtils;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.PrintStream;
|
||||
import java.util.Collections;
|
||||
import java.util.List;
|
||||
import java.util.*;
|
||||
|
||||
/**
|
||||
* Created by IntelliJ IDEA.
|
||||
|
|
@ -65,7 +67,7 @@ import java.util.List;
|
|||
* This set of arguments will also be passed to the constructor of every Covariate when it is instantiated.
|
||||
*/
|
||||
|
||||
public class RecalibrationArgumentCollection {
|
||||
public class RecalibrationArgumentCollection implements Cloneable {
|
||||
|
||||
/**
|
||||
* This algorithm treats every reference mismatch as an indication of error. However, real genetic variation is expected to mismatch the reference,
|
||||
|
|
@ -87,21 +89,6 @@ public class RecalibrationArgumentCollection {
|
|||
public File RECAL_TABLE_FILE = null;
|
||||
public PrintStream RECAL_TABLE;
|
||||
|
||||
/**
|
||||
* If not provided, then no plots will be generated (useful for queue scatter/gathering).
|
||||
* However, we *highly* recommend that users generate these plots whenever possible for QC checking.
|
||||
*/
|
||||
@Output(fullName = "plot_pdf_file", shortName = "plots", doc = "The output recalibration pdf file to create", required = false, defaultToStdout = false)
|
||||
public File RECAL_PDF_FILE = null;
|
||||
|
||||
/**
|
||||
* If not provided, then a temporary file is created and then deleted upon completion.
|
||||
* For advanced users only.
|
||||
*/
|
||||
@Advanced
|
||||
@Argument(fullName = "intermediate_csv_file", shortName = "intermediate", doc = "The intermediate csv file to create", required = false)
|
||||
public File RECAL_CSV_FILE = null;
|
||||
|
||||
/**
|
||||
* Note that the --list argument requires a fully resolved and correct command-line to work.
|
||||
*/
|
||||
|
|
@ -282,11 +269,147 @@ public class RecalibrationArgumentCollection {
|
|||
argumentsTable.set("quantizing_levels", RecalUtils.ARGUMENT_VALUE_COLUMN_NAME, QUANTIZING_LEVELS);
|
||||
argumentsTable.addRowID("recalibration_report", true);
|
||||
argumentsTable.set("recalibration_report", RecalUtils.ARGUMENT_VALUE_COLUMN_NAME, existingRecalibrationReport == null ? "null" : existingRecalibrationReport.getAbsolutePath());
|
||||
argumentsTable.addRowID("plot_pdf_file", true);
|
||||
argumentsTable.set("plot_pdf_file", RecalUtils.ARGUMENT_VALUE_COLUMN_NAME, RECAL_PDF_FILE == null ? "null" : RECAL_PDF_FILE.getAbsolutePath());
|
||||
argumentsTable.addRowID("binary_tag_name", true);
|
||||
argumentsTable.set("binary_tag_name", RecalUtils.ARGUMENT_VALUE_COLUMN_NAME, BINARY_TAG_NAME == null ? "null" : BINARY_TAG_NAME);
|
||||
return argumentsTable;
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns a map with the arguments that differ between this an
|
||||
* another {@link RecalibrationArgumentCollection} instance.
|
||||
* <p/>
|
||||
* The key is the name of that argument in the report file. The value is a message
|
||||
* that explains the difference to the end user.
|
||||
* <p/>
|
||||
* Thus, a empty map indicates that there is no differences between both argument collection that
|
||||
* is relevant to report comparison.
|
||||
* <p/>
|
||||
* This method should not throw any exception.
|
||||
*
|
||||
* @param other the argument-collection to compare against.
|
||||
* @param thisRole the name used to refer to this RAC report that makes sense to the end user.
|
||||
* @param otherRole the name used to refer to the other RAC report that makes sense to the end user.
|
||||
*
|
||||
* @return never <code>null</code>, but a zero-size collection if there are no differences.
|
||||
*/
|
||||
@Requires("other != null && thisRole != null && otherRole != null && !thisRole.equalsIgnoreCase(otherRole)")
|
||||
Map<String,? extends CharSequence> compareReportArguments(final RecalibrationArgumentCollection other,final String thisRole, final String otherRole) {
|
||||
final Map<String,String> result = new LinkedHashMap<>(15);
|
||||
compareRequestedCovariates(result, other, thisRole, otherRole);
|
||||
compareSimpleReportArgument(result,"no_standard_covs", DO_NOT_USE_STANDARD_COVARIATES, other.DO_NOT_USE_STANDARD_COVARIATES, thisRole, otherRole);
|
||||
compareSimpleReportArgument(result,"run_without_dbsnp",RUN_WITHOUT_DBSNP,other.RUN_WITHOUT_DBSNP,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"solid_recal_mode", SOLID_RECAL_MODE, other.SOLID_RECAL_MODE,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"solid_nocall_strategy", SOLID_NOCALL_STRATEGY, other.SOLID_NOCALL_STRATEGY,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"mismatches_context_size", MISMATCHES_CONTEXT_SIZE,other.MISMATCHES_CONTEXT_SIZE,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"mismatches_default_quality", MISMATCHES_DEFAULT_QUALITY, other.MISMATCHES_DEFAULT_QUALITY,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"deletions_default_quality", DELETIONS_DEFAULT_QUALITY, other.DELETIONS_DEFAULT_QUALITY,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"insertions_default_quality", INSERTIONS_DEFAULT_QUALITY, other.INSERTIONS_DEFAULT_QUALITY,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"maximum_cycle_value", MAXIMUM_CYCLE_VALUE, other.MAXIMUM_CYCLE_VALUE,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"low_quality_tail", LOW_QUAL_TAIL, other.LOW_QUAL_TAIL,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"default_platform", DEFAULT_PLATFORM, other.DEFAULT_PLATFORM,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"force_platform", FORCE_PLATFORM, other.FORCE_PLATFORM,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"quantizing_levels", QUANTIZING_LEVELS, other.QUANTIZING_LEVELS,thisRole,otherRole);
|
||||
compareSimpleReportArgument(result,"binary_tag_name", BINARY_TAG_NAME, other.BINARY_TAG_NAME,thisRole,otherRole);
|
||||
return result;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Compares the covariate report lists.
|
||||
*
|
||||
* @param diffs map where to annotate the difference.
|
||||
* @param other the argument collection to compare against.
|
||||
* @param thisRole the name for this argument collection that makes sense to the user.
|
||||
* @param otherRole the name for the other argument collection that makes sense to the end user.
|
||||
*
|
||||
* @return <code>true</code> if a difference was found.
|
||||
*/
|
||||
@Requires("diffs != null && other != null && thisRole != null && otherRole != null")
|
||||
private boolean compareRequestedCovariates(final Map<String,String> diffs,
|
||||
final RecalibrationArgumentCollection other, final String thisRole, final String otherRole) {
|
||||
|
||||
final Set<String> beforeNames = new HashSet<>(this.COVARIATES.length);
|
||||
final Set<String> afterNames = new HashSet<>(other.COVARIATES.length);
|
||||
Utils.addAll(beforeNames, this.COVARIATES);
|
||||
Utils.addAll(afterNames,other.COVARIATES);
|
||||
final Set<String> intersect = new HashSet<>(Math.min(beforeNames.size(),afterNames.size()));
|
||||
intersect.addAll(beforeNames);
|
||||
intersect.retainAll(afterNames);
|
||||
|
||||
String diffMessage = null;
|
||||
if (intersect.size() == 0) { // In practice this is not possible due to required covariates but...
|
||||
diffMessage = String.format("There are no common covariates between '%s' and '%s'"
|
||||
+ " recalibrator reports. Covariates in '%s': {%s}. Covariates in '%s': {%s}.",thisRole,otherRole,
|
||||
thisRole,Utils.join(", ",this.COVARIATES),
|
||||
otherRole,Utils.join(",",other.COVARIATES));
|
||||
} else if (intersect.size() != beforeNames.size() || intersect.size() != afterNames.size()) {
|
||||
beforeNames.removeAll(intersect);
|
||||
afterNames.removeAll(intersect);
|
||||
diffMessage = String.format("There are differences in the set of covariates requested in the"
|
||||
+ " '%s' and '%s' recalibrator reports. "
|
||||
+ " Exclusive to '%s': {%s}. Exclusive to '%s': {%s}.",thisRole,otherRole,
|
||||
thisRole,Utils.join(", ",beforeNames),
|
||||
otherRole,Utils.join(", ",afterNames));
|
||||
}
|
||||
if (diffMessage != null) {
|
||||
diffs.put("covariate",diffMessage);
|
||||
return true;
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Annotates a map with any difference encountered in a simple value report argument that differs between this an
|
||||
* another {@link RecalibrationArgumentCollection} instance.
|
||||
* <p/>
|
||||
* The key of the new entry would be the name of that argument in the report file. The value is a message
|
||||
* that explains the difference to the end user.
|
||||
* <p/>
|
||||
*
|
||||
* <p/>
|
||||
* This method should not return any exception.
|
||||
*
|
||||
* @param diffs where to annotate the differences.
|
||||
* @param name the name of the report argument to compare.
|
||||
* @param thisValue this argument collection value for that argument.
|
||||
* @param otherValue the other collection value for that argument.
|
||||
* @param thisRole the name used to refer to this RAC report that makes sense to the end user.
|
||||
* @param otherRole the name used to refer to the other RAC report that makes sense to the end user.
|
||||
*
|
||||
* @type T the argument Object value type.
|
||||
*
|
||||
* @return <code>true</code> if a difference has been spotted, thus <code>diff</code> has been modified.
|
||||
*/
|
||||
private <T> boolean compareSimpleReportArgument(final Map<String,String> diffs,
|
||||
final String name, final T thisValue, final T otherValue, final String thisRole, final String otherRole) {
|
||||
if (thisValue == null && otherValue == null) {
|
||||
return false;
|
||||
} else if (thisValue != null && thisValue.equals(otherValue)) {
|
||||
return false;
|
||||
} else {
|
||||
diffs.put(name,
|
||||
String.format("differences between '%s' {%s} and '%s' {%s}.",
|
||||
thisRole,thisValue == null ? "" : thisValue,
|
||||
otherRole,otherValue == null ? "" : otherValue));
|
||||
return true;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Create a shallow copy of this argument collection.
|
||||
*
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
@Override
|
||||
public RecalibrationArgumentCollection clone() {
|
||||
try {
|
||||
return (RecalibrationArgumentCollection) super.clone();
|
||||
} catch (CloneNotSupportedException e) {
|
||||
throw new StingException("Unreachable code clone not supported thrown when the class "
|
||||
+ this.getClass().getName() + " is cloneable ",e);
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
|
|
|
|||
|
|
@ -70,9 +70,7 @@ import org.broadinstitute.sting.utils.sam.GATKSAMReadGroupRecord;
|
|||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
import org.broadinstitute.sting.utils.sam.ReadUtils;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.IOException;
|
||||
import java.io.PrintStream;
|
||||
import java.io.*;
|
||||
import java.util.*;
|
||||
|
||||
/**
|
||||
|
|
@ -223,6 +221,150 @@ public class RecalUtils {
|
|||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Component used to print out csv representation of the reports that can be use to perform analysis in
|
||||
* external tools. E.g. generate plots using R scripts.
|
||||
* <p/>
|
||||
* A header is always printed into the output stream (or file) when the printer is created. Then you only need
|
||||
* to call {@link #print(RecalibrationReport,String) print} for each report you want to include in the csv file.
|
||||
* Once finished, you close the printer calling {@link #close() close}
|
||||
*
|
||||
*/
|
||||
private static class CsvPrinter {
|
||||
|
||||
private final PrintStream ps;
|
||||
private final Covariate[] covariates;
|
||||
|
||||
/**
|
||||
* Constructs a printer redirected to an output file.
|
||||
* @param out the output file.
|
||||
* @param c covariates to print out.
|
||||
* @throws FileNotFoundException if the file could not be created anew.
|
||||
*/
|
||||
protected CsvPrinter(final File out, final Covariate ... c)
|
||||
throws FileNotFoundException {
|
||||
this(new FileOutputStream(out), c);
|
||||
}
|
||||
|
||||
/**
|
||||
* Constructs a printer redirected to an output stream
|
||||
* @param os the output.
|
||||
* @param c covariates to print out.
|
||||
*/
|
||||
protected CsvPrinter(final OutputStream os, final Covariate ... c) {
|
||||
covariates = c == null ? new Covariate[0] : c.clone();
|
||||
ps = new PrintStream(os);
|
||||
printHeader();
|
||||
}
|
||||
|
||||
/**
|
||||
* Prints the header out.
|
||||
* <p/>
|
||||
* Should only be invoked at creation.
|
||||
*/
|
||||
protected void printHeader() {
|
||||
RecalUtils.printHeader(ps);
|
||||
}
|
||||
|
||||
/**
|
||||
* Prints out a report into the csv file.
|
||||
*
|
||||
*
|
||||
* @param report the report to print out.
|
||||
* @param mode the report associated mode. (typically ORIGINAL, RECALIBRATED
|
||||
*/
|
||||
public void print(final RecalibrationReport report, final String mode) {
|
||||
RecalUtils.writeCSV(ps,report.getRecalibrationTables(),mode,covariates,false);
|
||||
}
|
||||
|
||||
/**
|
||||
* Close the csv printer.
|
||||
*
|
||||
* No further output will be allowed or take place after calling this method.
|
||||
*/
|
||||
public void close() {
|
||||
ps.close();
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns a csv output printer.
|
||||
*
|
||||
* @param out the output file. It will be overridden
|
||||
* @param c list of covariates to print out.
|
||||
*
|
||||
* @throws FileNotFoundException if <code>out</code> could not be created anew.
|
||||
*
|
||||
* @return never <code>null</code>
|
||||
*/
|
||||
protected static CsvPrinter csvPrinter(final File out, final Covariate ... c)
|
||||
throws FileNotFoundException
|
||||
{
|
||||
if (c == null) {
|
||||
throw new IllegalArgumentException("the input covariate array cannot be null");
|
||||
}
|
||||
return new CsvPrinter(out,c);
|
||||
}
|
||||
|
||||
/**
|
||||
* Prints out a collection of reports into a file in Csv format in a way
|
||||
* that can be used by R scripts (such as the plot generator script).
|
||||
* <p/>
|
||||
* The set of covariates is take as the minimum common set from all reports.
|
||||
*
|
||||
* @param out the output file. It will be overridden.
|
||||
* @param reports map where keys are the unique 'mode' (ORIGINAL, RECALIBRATED, ...)
|
||||
* of each report and the corresponding value the report itself.
|
||||
* @throws FileNotFoundException if <code>out</code> could not be created anew.
|
||||
*/
|
||||
public static void generateCsv(final File out, final Map<String, RecalibrationReport> reports)
|
||||
throws FileNotFoundException {
|
||||
if (reports.size() == 0) {
|
||||
writeCsv(out, reports, new Covariate[0]);
|
||||
} else {
|
||||
final Iterator<RecalibrationReport> rit = reports.values().iterator();
|
||||
final RecalibrationReport first = rit.next();
|
||||
final Covariate[] firstCovariates = first.getRequestedCovariates();
|
||||
final Set<Covariate> covariates = new LinkedHashSet<>();
|
||||
Utils.addAll(covariates,firstCovariates);
|
||||
while (rit.hasNext() && covariates.size() > 0) {
|
||||
final Covariate[] nextCovariates = rit.next().getRequestedCovariates();
|
||||
final Set<String> nextCovariateNames = new LinkedHashSet<String>(nextCovariates.length);
|
||||
for (final Covariate nc : nextCovariates) {
|
||||
nextCovariateNames.add(nc.getClass().getSimpleName());
|
||||
}
|
||||
final Iterator<Covariate> cit = covariates.iterator();
|
||||
while (cit.hasNext()) {
|
||||
if (!nextCovariateNames.contains(cit.next().getClass().getSimpleName())) {
|
||||
cit.remove();
|
||||
}
|
||||
}
|
||||
}
|
||||
writeCsv(out, reports, covariates.toArray(new Covariate[covariates.size()]));
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Print out a collection of reports into a file in Csv format in a way
|
||||
* that can be used by R scripts (such as the plot generator script).
|
||||
*
|
||||
* @param out
|
||||
* @param reports map where keys are the unique 'mode' (ORIGINAL, RECALIBRATED, ...)
|
||||
* of each report and the corresponding value the report itself.
|
||||
* @param c the covariates to print out.
|
||||
* @throws FileNotFoundException if <code>out</code> could not be created anew.
|
||||
*/
|
||||
private static void writeCsv(final File out,
|
||||
final Map<String, RecalibrationReport> reports, final Covariate[] c)
|
||||
throws FileNotFoundException {
|
||||
final CsvPrinter p = csvPrinter(out,c);
|
||||
for (Map.Entry<String,RecalibrationReport> e : reports.entrySet()) {
|
||||
p.print(e.getValue(),e.getKey());
|
||||
}
|
||||
p.close();
|
||||
}
|
||||
|
||||
public enum SOLID_RECAL_MODE {
|
||||
/**
|
||||
* Treat reference inserted bases as reference matching bases. Very unsafe!
|
||||
|
|
@ -390,36 +532,66 @@ public class RecalUtils {
|
|||
report.print(outputFile);
|
||||
}
|
||||
|
||||
private static void outputRecalibrationPlot(final RecalibrationArgumentCollection RAC) {
|
||||
|
||||
/** s
|
||||
* Write recalibration plots into a file
|
||||
*
|
||||
* @param csvFile location of the intermediary file
|
||||
* @param exampleReportFile where the report arguments are collected from.
|
||||
* @param output result plot file name.
|
||||
*/
|
||||
public static void generatePlots(final File csvFile, final File exampleReportFile, final File output) {
|
||||
final RScriptExecutor executor = new RScriptExecutor();
|
||||
executor.setExceptOnError(true);
|
||||
executor.addScript(new Resource(SCRIPT_FILE, RecalUtils.class));
|
||||
executor.addArgs(RAC.RECAL_CSV_FILE.getAbsolutePath());
|
||||
executor.addArgs(RAC.RECAL_TABLE_FILE.getAbsolutePath());
|
||||
executor.addArgs(RAC.RECAL_PDF_FILE.getAbsolutePath());
|
||||
executor.addArgs(csvFile.getAbsolutePath());
|
||||
executor.addArgs(exampleReportFile.getAbsolutePath());
|
||||
executor.addArgs(output.getAbsolutePath());
|
||||
Logger.getLogger(RecalUtils.class).debug("R command line: " + executor.getApproximateCommandLine());
|
||||
executor.exec();
|
||||
}
|
||||
|
||||
private static void outputRecalibrationPlot(final File csvFile, final RecalibrationArgumentCollection RAC) {
|
||||
|
||||
final RScriptExecutor executor = new RScriptExecutor();
|
||||
executor.addScript(new Resource(SCRIPT_FILE, RecalUtils.class));
|
||||
executor.addArgs(csvFile.getAbsolutePath());
|
||||
executor.addArgs(RAC.RECAL_TABLE_FILE.getAbsolutePath());
|
||||
executor.exec();
|
||||
}
|
||||
|
||||
/**
|
||||
* Please use {@link #generateCsv(java.io.File, java.util.Map)} and {@link #generatePlots(java.io.File, java.io.File, java.io.File)} instead.
|
||||
*
|
||||
* @deprecated
|
||||
*/
|
||||
@Deprecated
|
||||
public static void generateRecalibrationPlot(final RecalibrationArgumentCollection RAC, final RecalibrationTables original, final Covariate[] requestedCovariates) {
|
||||
generateRecalibrationPlot(RAC, original, null, requestedCovariates);
|
||||
}
|
||||
|
||||
/**
|
||||
* Please use {@link #generateCsv(java.io.File, java.util.Map)} and {@link #generatePlots(java.io.File, java.io.File, java.io.File)} instead.
|
||||
*
|
||||
* @deprecated
|
||||
*/
|
||||
@Deprecated
|
||||
public static void generateRecalibrationPlot(final RecalibrationArgumentCollection RAC, final RecalibrationTables original, final RecalibrationTables recalibrated, final Covariate[] requestedCovariates) {
|
||||
final PrintStream csvFile;
|
||||
final PrintStream csvStream;
|
||||
final File csvTempFile = null;
|
||||
try {
|
||||
if ( RAC.RECAL_CSV_FILE == null ) {
|
||||
RAC.RECAL_CSV_FILE = File.createTempFile("BQSR", ".csv");
|
||||
RAC.RECAL_CSV_FILE.deleteOnExit();
|
||||
}
|
||||
csvFile = new PrintStream(RAC.RECAL_CSV_FILE);
|
||||
File csvTmpFile = File.createTempFile("BQSR",".csv");
|
||||
csvTmpFile.deleteOnExit();
|
||||
csvStream = new PrintStream(csvTmpFile);
|
||||
} catch (IOException e) {
|
||||
throw new UserException.CouldNotCreateOutputFile(RAC.RECAL_CSV_FILE, e);
|
||||
throw new UserException("Could not create temporary csv file", e);
|
||||
}
|
||||
|
||||
if ( recalibrated != null )
|
||||
writeCSV(csvFile, recalibrated, "RECALIBRATED", requestedCovariates, true);
|
||||
writeCSV(csvFile, original, "ORIGINAL", requestedCovariates, recalibrated == null);
|
||||
outputRecalibrationPlot(RAC);
|
||||
writeCSV(csvStream, recalibrated, "RECALIBRATED", requestedCovariates, true);
|
||||
writeCSV(csvStream, original, "ORIGINAL", requestedCovariates, recalibrated == null);
|
||||
csvStream.close();
|
||||
outputRecalibrationPlot(csvTempFile, RAC);
|
||||
csvTempFile.delete();
|
||||
}
|
||||
|
||||
private static void writeCSV(final PrintStream deltaTableFile, final RecalibrationTables recalibrationTables, final String recalibrationMode, final Covariate[] requestedCovariates, final boolean printHeader) {
|
||||
|
|
@ -452,18 +624,7 @@ public class RecalUtils {
|
|||
|
||||
// output the csv file
|
||||
if (printHeader) {
|
||||
final List<String> header = new LinkedList<String>();
|
||||
header.add("ReadGroup");
|
||||
header.add("CovariateValue");
|
||||
header.add("CovariateName");
|
||||
header.add("EventType");
|
||||
header.add("Observations");
|
||||
header.add("Errors");
|
||||
header.add("EmpiricalQuality");
|
||||
header.add("AverageReportedQuality");
|
||||
header.add("Accuracy");
|
||||
header.add("Recalibration");
|
||||
deltaTableFile.println(Utils.join(",", header));
|
||||
printHeader(deltaTableFile);
|
||||
}
|
||||
|
||||
final Map<Covariate, String> covariateNameMap = new HashMap<Covariate, String>(requestedCovariates.length);
|
||||
|
|
@ -480,6 +641,21 @@ public class RecalUtils {
|
|||
}
|
||||
}
|
||||
|
||||
private static void printHeader(PrintStream out) {
|
||||
final List<String> header = new LinkedList<String>();
|
||||
header.add("ReadGroup");
|
||||
header.add("CovariateValue");
|
||||
header.add("CovariateName");
|
||||
header.add("EventType");
|
||||
header.add("Observations");
|
||||
header.add("Errors");
|
||||
header.add("EmpiricalQuality");
|
||||
header.add("AverageReportedQuality");
|
||||
header.add("Accuracy");
|
||||
header.add("Recalibration");
|
||||
out.println(Utils.join(",", header));
|
||||
}
|
||||
|
||||
/*
|
||||
* Return an initialized nested integer array with appropriate dimensions for use with the delta tables
|
||||
*
|
||||
|
|
|
|||
|
|
@ -340,9 +340,6 @@ public class RecalibrationReport {
|
|||
else if (argument.equals("recalibration_report"))
|
||||
RAC.existingRecalibrationReport = (value == null) ? null : new File((String) value);
|
||||
|
||||
else if (argument.equals("plot_pdf_file"))
|
||||
RAC.RECAL_PDF_FILE = (value == null) ? null : new File((String) value);
|
||||
|
||||
else if (argument.equals("binary_tag_name"))
|
||||
RAC.BINARY_TAG_NAME = (value == null) ? null : (String) value;
|
||||
|
||||
|
|
@ -369,6 +366,11 @@ public class RecalibrationReport {
|
|||
return RAC;
|
||||
}
|
||||
|
||||
/**
|
||||
*
|
||||
* @deprecated use {@link #getRequestedCovariates()} instead.
|
||||
*/
|
||||
@Deprecated
|
||||
public Covariate[] getCovariates() {
|
||||
return requestedCovariates;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -67,6 +67,8 @@ import java.util.ArrayList;
|
|||
public class ContextCovariate implements StandardCovariate {
|
||||
private final static Logger logger = Logger.getLogger(ContextCovariate.class);
|
||||
|
||||
|
||||
|
||||
private int mismatchesContextSize;
|
||||
private int indelsContextSize;
|
||||
|
||||
|
|
|
|||
|
|
@ -0,0 +1,362 @@
|
|||
/*
|
||||
* 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.bqsr;
|
||||
|
||||
import org.broadinstitute.sting.WalkerTest;
|
||||
import org.broadinstitute.sting.utils.Utils;
|
||||
import org.broadinstitute.sting.utils.exceptions.UserException;
|
||||
import org.testng.annotations.DataProvider;
|
||||
import org.testng.annotations.Test;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.IOException;
|
||||
import java.lang.reflect.Method;
|
||||
import java.util.*;
|
||||
|
||||
import static org.testng.Assert.assertTrue;
|
||||
|
||||
/**
|
||||
* Tests Analyze Covariates.
|
||||
* <p/>
|
||||
* Notice that since PDF report generated by R are different every-time this program
|
||||
* is executed their content won't be tested. It only will verify that file has a healthy size.
|
||||
*
|
||||
*/
|
||||
public class AnalyzeCovariatesIntegrationTest extends WalkerTest {
|
||||
|
||||
private static final String TOOL_NAME = AnalyzeCovariates.class.getSimpleName();
|
||||
|
||||
/**
|
||||
* Directory where the testdata is located.
|
||||
*/
|
||||
private static final File TEST_DATA_DIR = new File(privateTestDir,"AnalyzeCovariates");
|
||||
|
||||
/**
|
||||
* File containing the before report for normal testing.
|
||||
*/
|
||||
private static final File BEFORE_FILE = new File(TEST_DATA_DIR,"before.table");
|
||||
|
||||
/**
|
||||
* File containing the after report for normal testing.
|
||||
*/
|
||||
private static final File AFTER_FILE = new File(TEST_DATA_DIR,"after.table");
|
||||
|
||||
|
||||
/**
|
||||
* File containing the bqsr report for normal testing.
|
||||
*/
|
||||
private static final File BQSR_FILE = new File(TEST_DATA_DIR,"bqsr.table");
|
||||
|
||||
/**
|
||||
* Test the content of the generated csv file.
|
||||
*
|
||||
* @throws IOException should never happen. It would be an indicator of a
|
||||
* problem with the testing environment.
|
||||
*/
|
||||
@Test(enabled = true)
|
||||
public void testCsvGeneration()
|
||||
throws IOException {
|
||||
|
||||
final WalkerTestSpec spec = new WalkerTestSpec(
|
||||
buildCommandLine("%s",null,true,true,true),
|
||||
Collections.singletonList("106709d32e6f0a0a9dd6a6340ec246ab"));
|
||||
executeTest("testCsvGeneration",spec);
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Test the size of the generated pdf.
|
||||
* <p/>
|
||||
* Unfortunately we cannot test the content as it changes slightly
|
||||
* every time the tool is run.
|
||||
*
|
||||
* @throws IOException should never happen. It would be an
|
||||
* indicator of a problem with the testing environment.
|
||||
*/
|
||||
@Test(enabled = true)
|
||||
public void testPdfGeneration()
|
||||
throws IOException {
|
||||
final File pdfFile = File.createTempFile("ACTest",".pdf");
|
||||
pdfFile.delete();
|
||||
pdfFile.deleteOnExit();
|
||||
|
||||
final List<String> md5 = Collections.emptyList();
|
||||
final WalkerTestSpec spec = new WalkerTestSpec(
|
||||
buildCommandLine(null,pdfFile.toString(),true,true,true),md5);
|
||||
executeTest("testPdfGeneration",spec);
|
||||
assertTrue(pdfFile.exists(),"the pdf file was not created");
|
||||
assertTrue(pdfFile.length() > 260000,"the pdf file size does"
|
||||
+ " not reach the minimum of 260Kb");
|
||||
}
|
||||
|
||||
/**
|
||||
* Test the effect of changing some recalibration parameters.
|
||||
* @param afterFileName name of the alternative after recalibration file.
|
||||
* @param description describes what has been changed.
|
||||
* @throws IOException should never happen. It would be an
|
||||
* indicator of a problem with the testing environment.
|
||||
*/
|
||||
@Test(enabled = true, dataProvider="alternativeAfterFileProvider")
|
||||
public void testParameterChangeException(final String afterFileName,
|
||||
final String description)
|
||||
throws IOException {
|
||||
|
||||
final File pdfFile = File.createTempFile("ACTest",".pdf");
|
||||
pdfFile.deleteOnExit();
|
||||
final List<String> md5 = Collections.emptyList();
|
||||
final File afterFile = new File(TEST_DATA_DIR,afterFileName);
|
||||
final WalkerTestSpec spec = new WalkerTestSpec(
|
||||
buildCommandLine(null,"%s",true,true,afterFile),
|
||||
1,UserException.IncompatibleRecalibrationTableParameters.class);
|
||||
executeTest("testParameterChangeException - " + description, spec);
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Test combinations of input and output inclusion exclusion of the command
|
||||
* line that cause an exception to be thrown.
|
||||
*
|
||||
* @param useCsvFile whether to include the output csv file.
|
||||
* @param usePdfFile whether to include the output pdf file.
|
||||
* @param useBQSRFile whether to include the -BQSR input file.
|
||||
* @param useBeforeFile whether to include the -before input file.
|
||||
* @param useAfterFile whether to include the -after input file.
|
||||
* @throws IOException never thrown, unless there is a problem with the testing environment.
|
||||
*/
|
||||
@Test(enabled = true, dataProvider="alternativeInOutAbsenceCombinations")
|
||||
public void testInOutAbsenceException(final boolean useCsvFile, final boolean usePdfFile,
|
||||
final boolean useBQSRFile, final boolean useBeforeFile, final boolean useAfterFile)
|
||||
throws IOException {
|
||||
final WalkerTestSpec spec = new WalkerTestSpec(buildCommandLine(useCsvFile,usePdfFile,
|
||||
useBQSRFile,useBeforeFile,useAfterFile),0,UserException.class);
|
||||
executeTest("testInOutAbsencePresenceException", spec);
|
||||
}
|
||||
|
||||
/**
|
||||
* Test combinations of input and output inclusion exclusion of the
|
||||
* command line that won't cause an exception.
|
||||
*
|
||||
* @param useCsvFile whether to include the output csv file.
|
||||
* @param usePdfFile whether to include the output pdf file.
|
||||
* @param useBQSRFile whether to include the -BQSR input file.
|
||||
* @param useBeforeFile whether to include the -before input file.
|
||||
* @param useAfterFile whether to include the -after input file.
|
||||
* @throws IOException never thrown, unless there is a problem with the testing environment.
|
||||
*/
|
||||
@Test(enabled = true, dataProvider="alternativeInOutAbsenceCombinations")
|
||||
public void testInOutAbsence(final boolean useCsvFile, final boolean usePdfFile,
|
||||
final boolean useBQSRFile, final boolean useBeforeFile, final boolean useAfterFile)
|
||||
throws IOException {
|
||||
final List<String> md5 = Collections.emptyList();
|
||||
final WalkerTestSpec spec = new WalkerTestSpec(buildCommandLine(useCsvFile,usePdfFile,
|
||||
useBQSRFile,useBeforeFile,useAfterFile),md5);
|
||||
executeTest("testInOutAbsencePresence", spec);
|
||||
}
|
||||
|
||||
|
||||
|
||||
@DataProvider
|
||||
public Iterator<Object[]> alternativeInOutAbsenceCombinations(Method m) {
|
||||
List<Object[]> result = new LinkedList<Object[]>();
|
||||
if (m.getName().endsWith("Exception")) {
|
||||
result.add(new Object[] { false, false, true, true, true });
|
||||
result.add(new Object[] { true, true, false, false ,false});
|
||||
}
|
||||
else {
|
||||
result.add(new Object[] { true, true, true, false, false });
|
||||
result.add(new Object[] { true, true, false, true, false });
|
||||
result.add(new Object[] { true, true, false, false, true });
|
||||
result.add(new Object[] { true, false,false, true, false });
|
||||
result.add(new Object[] { false, true, true, false, false });
|
||||
|
||||
}
|
||||
return result.iterator();
|
||||
}
|
||||
|
||||
/**
|
||||
* Provide recalibration parameter change data to relevant tests.
|
||||
* @param m target test method.
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
@DataProvider
|
||||
public Iterator<Object[]> alternativeAfterFileProvider (Method m) {
|
||||
final boolean expectsException = m.getName().endsWith("Exception");
|
||||
final List<Object[]> result = new LinkedList<Object[]>();
|
||||
for (final Object[] data : DIFFERENT_PARAMETERS_AFTER_FILES) {
|
||||
if (data[1].equals(expectsException)) {
|
||||
result.add(new Object[] { data[0], data[2] });
|
||||
}
|
||||
}
|
||||
return result.iterator();
|
||||
}
|
||||
|
||||
/**
|
||||
* Triplets < alfter-grp-file, whether it should fail, what is different >
|
||||
*/
|
||||
private final Object[][] DIFFERENT_PARAMETERS_AFTER_FILES = {
|
||||
{"after-cov.table", true, "Adds additional covariate: repeat-length" },
|
||||
{"after-dpSOLID.table", true, "Change the default platform to SOLID" },
|
||||
{"after-noDp.table",true, "Unset the default platform" },
|
||||
{"after-mcs4.table", true, "Changed -mcs parameter from 2 to 4" }
|
||||
};
|
||||
|
||||
/**
|
||||
* Build the AC command line given what combinations of input and output files should be included.
|
||||
*
|
||||
* @param useCsvFile whether to include the output csv file.
|
||||
* @param usePdfFile whether to include the output pdf file.
|
||||
* @param useBQSRFile whether to include the -BQSR input file.
|
||||
* @param useBeforeFile whether to include the -before input file.
|
||||
* @param useAfterFile whether to include the -after input file.
|
||||
* @return never <code>null</code>.
|
||||
* @throws IOException never thrown, unless there is a problem with the testing environment.
|
||||
*/
|
||||
private String buildCommandLine(final boolean useCsvFile, final boolean usePdfFile,
|
||||
final boolean useBQSRFile, final boolean useBeforeFile, final boolean useAfterFile)
|
||||
throws IOException {
|
||||
|
||||
final File csvFile = useCsvFile ? File.createTempFile("ACTest",".csv") : null;
|
||||
final File pdfFile = usePdfFile ? File.createTempFile("ACTest",".pdf") : null;
|
||||
|
||||
if (csvFile != null) {
|
||||
csvFile.deleteOnExit();
|
||||
}
|
||||
|
||||
if (pdfFile != null) {
|
||||
pdfFile.deleteOnExit();
|
||||
}
|
||||
|
||||
return buildCommandLine(csvFile == null ? null : csvFile.toString(),
|
||||
pdfFile == null ? null : pdfFile.toString(),
|
||||
useBQSRFile,useBeforeFile,useAfterFile);
|
||||
}
|
||||
|
||||
/**
|
||||
* Build the AC command line given the output file names explicitly and what test input files to use.
|
||||
* <p/>
|
||||
*
|
||||
* @param csvFileName the csv output file, <code>null</code> if none should be provided.
|
||||
* @param pdfFileName the plots output file, <code>null</code> if none should be provided.
|
||||
* @param useBQSRFile whether to include the -BQSR input file.
|
||||
* @param useBeforeFile whether to include the -before input file.
|
||||
* @param useAfterFile whether to include the -after input file.
|
||||
*
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
private String buildCommandLine(final String csvFileName, final String pdfFileName, final boolean useBQSRFile,
|
||||
final boolean useBeforeFile, final boolean useAfterFile) {
|
||||
return buildCommandLine(csvFileName,pdfFileName,useBQSRFile ? BQSR_FILE : null,
|
||||
useBeforeFile ? BEFORE_FILE : null,
|
||||
useAfterFile ? AFTER_FILE : null);
|
||||
}
|
||||
|
||||
/**
|
||||
* Build the AC command line given the output file names and the after file name explicitly and what other
|
||||
* test input files to use.
|
||||
* <p/>
|
||||
*
|
||||
* @param csvFileName the csv output file, <code>null</code> if none should be provided.
|
||||
* @param pdfFileName the plots output file, <code>null</code> if none should be provided.
|
||||
* @param useBQSRFile whether to include the -BQSR input file.
|
||||
* @param useBeforeFile whether to include the -before input file.
|
||||
* @param afterFile the after input report file, <code>null</code> if none should be provided.
|
||||
*
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
private String buildCommandLine(final String csvFileName, final String pdfFileName, final boolean useBQSRFile,
|
||||
final boolean useBeforeFile, final File afterFile) {
|
||||
return buildCommandLine(csvFileName,pdfFileName,useBQSRFile ? BQSR_FILE : null,
|
||||
useBeforeFile ? BEFORE_FILE : null,
|
||||
afterFile);
|
||||
}
|
||||
|
||||
/**
|
||||
* Build the AC command line given the output file names and the after file name explicitly and what other
|
||||
* test input files to use.
|
||||
* <p/>
|
||||
*
|
||||
* @param csvFileName the csv output file, <code>null</code> if none should be provided.
|
||||
* @param pdfFileName the plots output file, <code>null</code> if none should be provided.
|
||||
* @param bqsrFile the BQSR input report file, <code>null</code> if none should be provided.
|
||||
* @param beforeFile the before input report file, <code>null</code> if non should be provided.
|
||||
* @param afterFile the after input report file, <code>null</code> if none should be provided.
|
||||
*
|
||||
* @return never <code>null</code>.
|
||||
*/
|
||||
private String buildCommandLine(final String csvFileName, final String pdfFileName, final File bqsrFile,
|
||||
final File beforeFile, final File afterFile) {
|
||||
|
||||
final List<String> args = new LinkedList<String>();
|
||||
args.add("-T");
|
||||
args.add(TOOL_NAME);
|
||||
args.add("-R");
|
||||
args.add(hg19Reference);
|
||||
args.add("-ignoreLMT");
|
||||
|
||||
if (csvFileName != null) {
|
||||
args.add("-" + AnalyzeCovariates.CSV_ARG_SHORT_NAME);
|
||||
args.add("'" + csvFileName + "'");
|
||||
}
|
||||
if (pdfFileName != null) {
|
||||
args.add("-" + AnalyzeCovariates.PDF_ARG_SHORT_NAME);
|
||||
args.add("'" + pdfFileName + "'");
|
||||
}
|
||||
if (bqsrFile != null) {
|
||||
args.add("-BQSR");
|
||||
args.add("'" + bqsrFile.getAbsoluteFile().toString() + "'");
|
||||
}
|
||||
if (beforeFile != null) {
|
||||
args.add("-" + AnalyzeCovariates.BEFORE_ARG_SHORT_NAME);
|
||||
args.add("'" + beforeFile.getAbsolutePath().toString() + "'");
|
||||
}
|
||||
if (afterFile != null) {
|
||||
args.add("-" + AnalyzeCovariates.AFTER_ARG_SHORT_NAME);
|
||||
args.add("'" + afterFile.getAbsolutePath().toString() + "'");
|
||||
}
|
||||
return Utils.join(" ", args);
|
||||
|
||||
}
|
||||
}
|
||||
|
|
@ -100,23 +100,23 @@ public class BQSRIntegrationTest extends WalkerTest {
|
|||
@DataProvider(name = "BQSRTest")
|
||||
public Object[][] createBQSRTestData() {
|
||||
return new Object[][]{
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, "", "61fd466b5e94d2d67e116f6f67c9f939")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --no_standard_covs -cov ContextCovariate", "e08b5bcdb64f4beea03730e5631a14ca")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --no_standard_covs -cov CycleCovariate", "448a45dc154c95d1387cb5cdddb67071")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --indels_context_size 4", "c1e7999e445d51bbe2e775dac5325643")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --low_quality_tail 5", "a57c16918cdfe12d55a89c21bf195279")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --quantizing_levels 6", "836dccacf48ccda6b2843d07e8f1ef4d")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --mismatches_context_size 4", "0fb2aedc2f8d66b5821cb570f15a8c4d")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", "", "c9953f020a65c1603a6d71aeeb1b95f3")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA19240.chr1.BFAST.SOLID.bam", "1:10,000,000-10,200,000", "", "85a120b7d86b61597b86b9e93decbdfc")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12873.454.SRP000031.2009_06.chr1.10_20mb.1RG.bam", "1:10,000,000-10,200,000", "", "5248dc49aec0323c74b496bb4928c73c")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "originalQuals.1kg.chr1.1-1K.1RG.bam", "1:1-1,000", " -OQ", "cb52f267e0010f849f50b0bf1de474a1")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA19240.chr1.BFAST.SOLID.bam", "1:10,000,000-20,000,000", " --solid_recal_mode REMOVE_REF_BIAS", "fb372d0a8fc41b01ced1adab31546850")},
|
||||
{new BQSRTest(b36KGReference, privateTestDir + "NA19240.chr1.BFAST.SOLID.hasCSNoCall.bam", "1:50,000-80,000", " --solid_nocall_strategy LEAVE_READ_UNRECALIBRATED", "c1c3cda8caceed619d3d439c3990cd26")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", " -knownSites:anyNameABCD,VCF " + privateTestDir + "vcfexample3.vcf", "c9953f020a65c1603a6d71aeeb1b95f3")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", " -knownSites:bed " + validationDataLocation + "bqsrKnownTest.bed", "5bfff0c699345cca12a9b33acf95588f")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, "", "f805a0020eea987b79f314fa99913806")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --no_standard_covs -cov ContextCovariate", "86075d3856eb06816a0dd81af55e421f")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --no_standard_covs -cov CycleCovariate", "155802237e1fc7a001398b8f4bcf4b72")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --indels_context_size 4", "38c7916cc019fe8d134df67639422b42")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --low_quality_tail 5", "b74e75f3c5aa90bd21af1e20f2ac8c40")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --quantizing_levels 6", "e564505aea11464de8ed72890d9ea89a")},
|
||||
{new BQSRTest(hg18Reference, HiSeqBam, HiSeqInterval, " --mismatches_context_size 4", "380d8be121ffaddd3461ee0ac3d1a76f")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", "", "0b5a8e259e997e4c7b5836d4c28e6f4d")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA19240.chr1.BFAST.SOLID.bam", "1:10,000,000-10,200,000", "", "281682124584ab384f23359934df0c3b")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12873.454.SRP000031.2009_06.chr1.10_20mb.1RG.bam", "1:10,000,000-10,200,000", "", "0a92fdff5fd26227c29d34eda5a32f49")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "originalQuals.1kg.chr1.1-1K.1RG.bam", "1:1-1,000", " -OQ", "90d8c24077e8ae9a0037a9aad5f09e31")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA19240.chr1.BFAST.SOLID.bam", "1:10,000,000-20,000,000", " --solid_recal_mode REMOVE_REF_BIAS", "c41ef02c640ef1fed4bfc03b9b33b616")},
|
||||
{new BQSRTest(b36KGReference, privateTestDir + "NA19240.chr1.BFAST.SOLID.hasCSNoCall.bam", "1:50,000-80,000", " --solid_nocall_strategy LEAVE_READ_UNRECALIBRATED", "b577cd1d529425f66db49620db09fdca")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", " -knownSites:anyNameABCD,VCF " + privateTestDir + "vcfexample3.vcf", "0b5a8e259e997e4c7b5836d4c28e6f4d")},
|
||||
{new BQSRTest(b36KGReference, validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.1Mb.1RG.bam", "1:10,000,000-10,200,000", " -knownSites:bed " + validationDataLocation + "bqsrKnownTest.bed", "9ad49269c0156f8ab1173261bf23e600")},
|
||||
// make sure we work with ION torrent bam
|
||||
{new BQSRTest(b37KGReference, privateTestDir + "iontorrent.bam", "20:10,000,000-10,200,000", "", "7375c7b692e76b651c278a9fb478fa1c")},
|
||||
{new BQSRTest(b37KGReference, privateTestDir + "iontorrent.bam", "20:10,000,000-10,200,000", "", "04bfa4760767022e7f5252e6e4432cc1")},
|
||||
};
|
||||
}
|
||||
|
||||
|
|
@ -141,22 +141,6 @@ public class BQSRIntegrationTest extends WalkerTest {
|
|||
executeTest("testBQSRFailWithoutDBSNP", spec);
|
||||
}
|
||||
|
||||
@Test
|
||||
public void testBQSRCSV() {
|
||||
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
|
||||
" -T BaseRecalibrator" +
|
||||
" -R " + b36KGReference +
|
||||
" -I " + validationDataLocation + "NA12892.SLX.SRP000031.2009_06.selected.bam" +
|
||||
" -knownSites " + b36dbSNP129 +
|
||||
" -L 1:10,000,000-10,200,000" +
|
||||
" -o /dev/null" +
|
||||
" -sortAllCols" +
|
||||
" --plot_pdf_file /dev/null" +
|
||||
" --intermediate_csv_file %s",
|
||||
Arrays.asList("90ad19143024684e3c4410dc8fd2bd9d"));
|
||||
executeTest("testBQSR-CSVfile", spec);
|
||||
}
|
||||
|
||||
@Test
|
||||
public void testBQSRFailWithSolidNoCall() {
|
||||
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
|
||||
|
|
|
|||
|
|
@ -12,7 +12,27 @@ if ( interactive() ) {
|
|||
args <- commandArgs(TRUE)
|
||||
}
|
||||
data <- read.csv(args[1])
|
||||
|
||||
data$Recalibration = as.factor(sapply(as.character(data$Recalibration),function(x) {
|
||||
xu = toupper(x);
|
||||
if (xu == "ORIGINAL") "BEFORE" else
|
||||
if (xu == "RECALIBRATED") "AFTER" else
|
||||
if (xu == "RECALIBRATION") "BQSR" else
|
||||
xu }));
|
||||
|
||||
gsa.report <- gsa.read.gatkreport(args[2])
|
||||
|
||||
gsa.report$Arguments$Value = as.character(gsa.report$Arguments$Value);
|
||||
gsa.report$Arguments = subset(gsa.report$Arguments,subset= Argument != "plot_pdf_file");
|
||||
if (length(levels(data$Recalibration)) > 1) {
|
||||
gsa.report$Arguments = subset(gsa.report$Arguments,subset= Argument != "recalibration_report");
|
||||
}
|
||||
gsa.report$Arguments$Value[gsa.report$Argument$Value == "null"] = "None";
|
||||
|
||||
gsa.report.covariate.argnum = gsa.report$Arguments$Argument == "covariate";
|
||||
gsa.report$Arguments$Value[gsa.report.covariate.argnum] = sapply(strsplit(gsa.report$Arguments$Value[gsa.report.covariate.argnum],","),function(x) {
|
||||
y = sub("(^.+)Covariate","\\1",x); paste(y,collapse=",") } );
|
||||
|
||||
data <- within(data, EventType <- factor(EventType, levels = rev(levels(EventType))))
|
||||
|
||||
numRG = length(unique(data$ReadGroup))
|
||||
|
|
@ -54,31 +74,31 @@ for(cov in levels(data$CovariateName)) { # for each covariate in turn
|
|||
d=rbind(dSub, dIns, dDel)
|
||||
|
||||
if( cov != "QualityScore" ) {
|
||||
p <- ggplot(d, aes(x=CovariateValue,y=Accuracy,alpha=log10(Observations))) +
|
||||
p <- ggplot(d, aes(x=CovariateValue,y=Accuracy,alpha=log10(Observations))) + ylim(min(-10,d$Accuracy),max(10,d$Accuracy)) +
|
||||
geom_abline(intercept=0, slope=0, linetype=2) +
|
||||
xlab(paste(cov,"Covariate")) +
|
||||
ylab("Quality Score Accuracy") +
|
||||
blankTheme
|
||||
if(cov == "Cycle") {
|
||||
b <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("maroon1","blue")) + facet_grid(.~EventType) +
|
||||
b <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black")) + facet_grid(.~EventType) +
|
||||
opts(axis.text.x=theme_text(angle=90, hjust=0))
|
||||
|
||||
p <- ggplot(d, aes(x=CovariateValue,y=AverageReportedQuality,alpha=log10(Observations))) +
|
||||
xlab(paste(cov,"Covariate")) +
|
||||
ylab("Mean Quality Score") +
|
||||
ylab("Mean Quality Score") + ylim(0,max(42,d$AverageReportedQuality)) +
|
||||
blankTheme
|
||||
e <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("maroon1","blue")) + facet_grid(.~EventType) +
|
||||
e <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black")) + facet_grid(.~EventType) +
|
||||
opts(axis.text.x=theme_text(angle=90, hjust=0))
|
||||
|
||||
|
||||
} else {
|
||||
c <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("maroon1","blue")) + facet_grid(.~EventType) +
|
||||
c <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black")) + facet_grid(.~EventType) +
|
||||
opts(axis.text.x=theme_text(angle=90, hjust=0)) + xlab(paste(cov,"Covariate (3 base suffix)"))
|
||||
p <- ggplot(d, aes(x=CovariateValue,y=AverageReportedQuality,alpha=log10(Observations))) +
|
||||
xlab(paste(cov,"Covariate (3 base suffix)")) +
|
||||
ylab("Mean Quality Score") +
|
||||
blankTheme
|
||||
f <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("maroon1","blue")) + facet_grid(.~EventType) +
|
||||
f <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black")) + facet_grid(.~EventType) +
|
||||
opts(axis.text.x=theme_text(angle=90, hjust=0))
|
||||
|
||||
}
|
||||
|
|
@ -88,14 +108,14 @@ for(cov in levels(data$CovariateName)) { # for each covariate in turn
|
|||
xlab("Reported Quality Score") +
|
||||
ylab("Empirical Quality Score") +
|
||||
blankTheme
|
||||
a <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("maroon1","blue")) + facet_grid(.~EventType)
|
||||
a <- p + geom_point(aes(color=Recalibration)) + scale_color_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black")) + facet_grid(.~EventType)
|
||||
|
||||
p <- ggplot(d, aes(x=CovariateValue)) +
|
||||
xlab(paste(cov,"Covariate")) +
|
||||
ylab("No. of Observations (area normalized)") +
|
||||
blankTheme
|
||||
d <- p + geom_histogram(aes(fill=Recalibration,weight=Observations,y=..ndensity..),alpha=0.6,binwidth=1,position="identity")
|
||||
d <- d + scale_fill_manual(values=c("maroon1","blue"))
|
||||
d <- d + scale_fill_manual(values=c("BEFORE"="maroon1","AFTER"="blue","BQSR"="black"))
|
||||
d <- d + facet_grid(.~EventType)
|
||||
# d <- d + scale_y_continuous(formatter="comma")
|
||||
}
|
||||
|
|
|
|||
|
|
@ -683,6 +683,36 @@ public class Utils {
|
|||
return denom == 0 ? "NA" : String.format("%.2f", num / (1.0 * denom));
|
||||
}
|
||||
|
||||
/**
|
||||
* Adds element from an array into a collection.
|
||||
*
|
||||
* In the event of exception being throw due to some element, <code>dest</code> might have been modified by
|
||||
* the successful addition of element before that one.
|
||||
*
|
||||
* @param dest the destination collection which cannot be <code>null</code> and should be able to accept
|
||||
* the input elements.
|
||||
* @param elements the element to add to <code>dest</code>
|
||||
* @param <T> collection type element.
|
||||
* @throws UnsupportedOperationException if the <tt>add</tt> operation
|
||||
* is not supported by <code>dest</code>.
|
||||
* @throws ClassCastException if the class of any of the elements
|
||||
* prevents it from being added to <code>dest</code>.
|
||||
* @throws NullPointerException if any of the elements is <code>null</code> and <code>dest</code>
|
||||
* does not permit <code>null</code> elements
|
||||
* @throws IllegalArgumentException if some property of any of the elements
|
||||
* prevents it from being added to this collection
|
||||
* @throws IllegalStateException if any of the elements cannot be added at this
|
||||
* time due to insertion restrictions.
|
||||
* @return <code>true</code> if the collection was modified as a result.
|
||||
*/
|
||||
public static <T> boolean addAll(Collection<T> dest, T ... elements) {
|
||||
boolean result = false;
|
||||
for (final T e : elements) {
|
||||
result = dest.add(e) | result;
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
/**
|
||||
* Create a constant map that maps each value in values to itself
|
||||
*/
|
||||
|
|
|
|||
|
|
@ -471,4 +471,10 @@ public class UserException extends ReviewedStingException {
|
|||
super(message,innerException);
|
||||
}
|
||||
}
|
||||
|
||||
public static class IncompatibleRecalibrationTableParameters extends UserException {
|
||||
public IncompatibleRecalibrationTableParameters(String s) {
|
||||
super(s);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in New Issue