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

VCF headers now include 'contig' lines describing the name, length, and assembly (when easily parsable) for each contig in the reference.

This commit is contained in:
Eric Banks 2011-08-09 11:24:48 -04:00
parent 7afb5c9f1c
commit ec76bf6d4a
1 changed files with 38 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
public/java/src/org/broadinstitute/sting/gatk/io/stubs/VCFWriterStub.java
View File

@ -25,6 +25,7 @@
package org.broadinstitute.sting.gatk.io.stubs; package org.broadinstitute.sting.gatk.io.stubs;
import net.sf.samtools.SAMSequenceRecord;
import org.broadinstitute.sting.gatk.CommandLineExecutable; import org.broadinstitute.sting.gatk.CommandLineExecutable;
import org.broadinstitute.sting.gatk.GenomeAnalysisEngine; import org.broadinstitute.sting.gatk.GenomeAnalysisEngine;
import org.broadinstitute.sting.gatk.io.OutputTracker; import org.broadinstitute.sting.gatk.io.OutputTracker;
@ -177,14 +178,21 @@ public class VCFWriterStub implements Stub<VCFWriter>, VCFWriter {
vcfHeader = header; vcfHeader = header;
// Check for the command-line argument header line. If not present, add it in. // Check for the command-line argument header line. If not present, add it in.
VCFHeaderLine commandLineArgHeaderLine = getCommandLineArgumentHeaderLine(); if ( !skipWritingHeader ) {
boolean foundCommandLineHeaderLine = false; VCFHeaderLine commandLineArgHeaderLine = getCommandLineArgumentHeaderLine();
for(VCFHeaderLine line: vcfHeader.getMetaData()) { boolean foundCommandLineHeaderLine = false;
if(line.getKey().equals(commandLineArgHeaderLine.getKey())) for (VCFHeaderLine line: vcfHeader.getMetaData()) {
foundCommandLineHeaderLine = true; if ( line.getKey().equals(commandLineArgHeaderLine.getKey()) )
foundCommandLineHeaderLine = true;
}
if ( !foundCommandLineHeaderLine )
vcfHeader.addMetaDataLine(commandLineArgHeaderLine);
// also put in the reference contig header lines
String assembly = getReferenceAssembly(engine.getArguments().referenceFile.getName());
for ( SAMSequenceRecord contig : engine.getReferenceDataSource().getReference().getSequenceDictionary().getSequences() )
vcfHeader.addMetaDataLine(getContigHeaderLine(contig, assembly));
} }
if(!foundCommandLineHeaderLine && !skipWritingHeader)
vcfHeader.addMetaDataLine(commandLineArgHeaderLine);
outputTracker.getStorage(this).writeHeader(vcfHeader); outputTracker.getStorage(this).writeHeader(vcfHeader);
} }
@ -220,4 +228,27 @@ public class VCFWriterStub implements Stub<VCFWriter>, VCFWriter {
CommandLineExecutable executable = JVMUtils.getObjectOfType(argumentSources,CommandLineExecutable.class); CommandLineExecutable executable = JVMUtils.getObjectOfType(argumentSources,CommandLineExecutable.class);
return new VCFHeaderLine(executable.getAnalysisName(), "\"" + engine.createApproximateCommandLineArgumentString(argumentSources.toArray()) + "\""); return new VCFHeaderLine(executable.getAnalysisName(), "\"" + engine.createApproximateCommandLineArgumentString(argumentSources.toArray()) + "\"");
} }
private VCFHeaderLine getContigHeaderLine(SAMSequenceRecord contig, String assembly) {
String val;
if ( assembly != null )
val = String.format("<ID=%s,length=%d,assembly=%s>", contig.getSequenceName(), contig.getSequenceLength(), assembly);
else
val = String.format("<ID=%s,length=%d>", contig.getSequenceName(), contig.getSequenceLength());
return new VCFHeaderLine("contig", val);
}
private String getReferenceAssembly(String refPath) {
// This doesn't need to be perfect as it's not a required VCF header line, but we might as well give it a shot
String assembly = null;
if ( refPath.indexOf("b37") != -1 || refPath.indexOf("v37") != -1 )
assembly = "b37";
else if ( refPath.indexOf("b36") != -1 )
assembly = "b36";
else if ( refPath.indexOf("hg18") != -1 )
assembly = "hg18";
else if ( refPath.indexOf("hg19") != -1 )
assembly = "hg19";
return assembly;
}
} }