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

Cleanup of old VCFRecord code 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3883 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
ebanks 2010-07-27 02:56:47 +00:00
parent 7dd55fbf13
commit 1d9ed1e214
2 changed files with 0 additions and 120 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

100
java/src/org/broadinstitute/sting/utils/genotype/vcf/VCFUtils.java
View File

@ -30,9 +30,6 @@ import org.broadinstitute.sting.gatk.GenomeAnalysisEngine;
import org.broadinstitute.sting.gatk.contexts.variantcontext.VariantContext;
import org.broadinstitute.sting.gatk.datasources.simpleDataSources.ReferenceOrderedDataSource;
import org.broadinstitute.sting.gatk.refdata.tracks.RMDTrack;
import org.broadinstitute.sting.utils.GenomeLocParser;
import org.broadinstitute.sting.utils.collections.Pair;
import org.broadinstitute.sting.utils.Utils;
import org.apache.log4j.Logger;
import java.util.*;
@ -70,64 +67,8 @@ public class VCFUtils {
return fields;
}
/**
* Merges various vcf records into a single one using the mapping from rodNamesToSampleNames to get unique sample names
*
* @param rods the vcf rods
* @param rodNamesToSampleNames mapping of rod/sample pairs to new uniquified sample names
* @return the new merged vcf record
*/
public static VCFRecord mergeRecords(Map<VCFRecord,String> rods, Map<Pair<String, String>, String> rodNamesToSampleNames) {
VCFParameters params = new VCFParameters();
params.addFormatItem(VCFConstants.GENOTYPE_KEY);
// keep track of the data so we can merge them intelligently
double maxConfidence = 0.0;
String id = null;
Map<String, String> infoFields = new HashMap<String, String>();
List<String> filters = new ArrayList<String>();
for ( VCFRecord rod : rods.keySet() ) {
List<VCFGenotypeRecord> myGenotypes = rod.getVCFGenotypeRecords();
for ( VCFGenotypeRecord call : myGenotypes ) {
// set the name to be the new uniquified name and add it to the list of genotypes
call.setSampleName(rodNamesToSampleNames.get(new Pair<String, String>(rods.get(rod), call.getSampleName())));
if ( params.getPosition() < 1 )
params.setLocations(GenomeLocParser.createGenomeLoc(rod.getChr(), rod.getStart()), call.getReference());
params.addGenotypeRecord(createVCFGenotypeRecord(params, call, rod));
}
// set the overall confidence to be the max entry we see
double confidence = 10.0 * rod.getNegLog10PError();
if ( confidence > maxConfidence )
maxConfidence = confidence;
if ( rod.getID() != null )
id = rod.getID();
if ( rod.isFiltered() )
filters.add(rod.getFilterString());
// just take the last value we see for a given key
infoFields.putAll(rod.getInfoValues());
}
return new VCFRecord(params.getReferenceBases(),
params.getContig(),
params.getPosition(),
(id != null ? id : "."),
params.getAlternateBases(),
maxConfidence,
filters.size() == 0 ? "0" : Utils.join(";", filters),
infoFields,
params.getFormatString(),
params.getGenotypeRecords());
}
public static Set<VCFHeaderLine> smartMergeHeaders(Collection<VCFHeader> headers, Logger logger) throws IllegalStateException {
HashMap<String, VCFHeaderLine> map = new HashMap<String, VCFHeaderLine>(); // from KEY.NAME -> line
HashSet<VCFHeaderLine> lines = new HashSet<VCFHeaderLine>();
// todo -- needs to remove all version headers from sources and add its own VCF version line
for ( VCFHeader source : headers ) {
@ -182,45 +123,4 @@ public class VCFUtils {
return new HashSet<VCFHeaderLine>(map.values());
}
/**
* create the VCF genotype record
*
* @param params the VCF parameters object
* @param gtype the genotype
* @param vcfrecord the VCF record
*
* @return a VCFGenotypeRecord
*/
public static VCFGenotypeRecord createVCFGenotypeRecord(VCFParameters params, VCFGenotypeRecord gtype, VCFRecord vcfrecord) {
List<VCFGenotypeEncoding> alleles = createAlleleArray(gtype);
for (VCFGenotypeEncoding allele : alleles) {
params.addAlternateBase(allele);
}
VCFGenotypeRecord record = new VCFGenotypeRecord(gtype.getSampleName(), alleles, VCFGenotypeRecord.PHASE.UNPHASED);
for ( Map.Entry<String, String> entry : gtype.getFields().entrySet() ) {
record.setField(entry.getKey(), entry.getValue());
params.addFormatItem(entry.getKey());
}
record.setVCFRecord(vcfrecord);
return record;
}
/**
* create the allele array?
*
* @param gtype the gentoype object
*
* @return a list of string representing the string array of alleles
*/
private static List<VCFGenotypeEncoding> createAlleleArray(VCFGenotypeRecord gtype) {
List<VCFGenotypeEncoding> alleles = new ArrayList<VCFGenotypeEncoding>();
for (char allele : gtype.getBases().toCharArray()) {
alleles.add(new VCFGenotypeEncoding(String.valueOf(allele)));
}
return alleles;
}
}

20
java/src/org/broadinstitute/sting/utils/genotype/vcf/VCFWriter.java
View File

@ -101,26 +101,6 @@ public class VCFWriter {
}
}
/**
* output a record to the VCF file
*
* @param record the record to output
*/
@Deprecated
public void addRecord(VCFRecord record) {
if ( mHeader == null )
throw new IllegalStateException("The VCF Header must be written before records can be added");
String vcfString = record.toStringEncoding(mHeader);
try {
mWriter.write(vcfString + "\n");
mWriter.flush(); // necessary so that writing to an output stream will work
} catch (IOException e) {
throw new RuntimeException("Unable to write the VCF object to a file");
}
}
/**
* attempt to close the VCF file
*/