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

Active Region Traversal should use GATKSAMRecords everywhere instead of SAMRecords. misc cleanup.

This commit is contained in:
Ryan Poplin 2012-02-03 17:13:52 -05:00
parent 5af3999a2d
commit 894d3340be
5 changed files with 16 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
public/java/src/org/broadinstitute/sting/analyzecovariates/AnalyzeCovariates.java
View File

@ -139,11 +139,11 @@ public class AnalyzeCovariates extends CommandLineProgram {
*/
@Argument(fullName="max_histogram_value", shortName="maxHist", required = false, doc="If supplied, this value will be the max value of the histogram plots")
private int MAX_HISTOGRAM_VALUE = 0;
@Hidden
@Argument(fullName="do_indel_quality", shortName="indels", required = false, doc="If supplied, do indel quality plotting")
private boolean DO_INDEL_QUALITY = false;
/////////////////////////////
// Private Member Variables
/////////////////////////////
@ -274,7 +274,6 @@ public class AnalyzeCovariates extends CommandLineProgram {
RecalDatum datum = new RecalDatum( Long.parseLong( vals[iii] ), Long.parseLong( vals[iii + 1] ), Double.parseDouble( vals[1] ), 0.0 );
// Add that datum to all the collapsed tables which will be used in the sequential calculation
dataManager.addToAllTables( key, datum, IGNORE_QSCORES_LESS_THAN );
}
private void writeDataTables() {
@ -341,7 +340,7 @@ public class AnalyzeCovariates extends CommandLineProgram {
// for each covariate
for( int iii = 1; iii < requestedCovariates.size(); iii++ ) {
Covariate cov = requestedCovariates.get(iii);
final Covariate cov = requestedCovariates.get(iii);
final File outputFile = new File(OUTPUT_DIR, readGroup + "." + cov.getClass().getSimpleName()+ ".dat");
if (DO_INDEL_QUALITY) {
RScriptExecutor executor = new RScriptExecutor();
@ -349,7 +348,7 @@ public class AnalyzeCovariates extends CommandLineProgram {
// The second argument is the name of the covariate in order to make the plots look nice
executor.addArgs(outputFile, cov.getClass().getSimpleName().split("Covariate")[0]);
executor.exec();
} else {
} else {
if( iii == 1 ) {
// Analyze reported quality
RScriptExecutor executor = new RScriptExecutor();

21
public/java/src/org/broadinstitute/sting/gatk/traversals/TraverseActiveRegions.java
View File

@ -1,6 +1,5 @@
package org.broadinstitute.sting.gatk.traversals;
import net.sf.samtools.SAMRecord;
import org.apache.log4j.Logger;
import org.broadinstitute.sting.gatk.WalkerManager;
import org.broadinstitute.sting.gatk.contexts.AlignmentContext;
@ -29,7 +28,7 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
protected static Logger logger = Logger.getLogger(TraversalEngine.class);
private final Queue<ActiveRegion> workQueue = new LinkedList<ActiveRegion>();
private final LinkedHashSet<SAMRecord> myReads = new LinkedHashSet<SAMRecord>();
private final LinkedHashSet<GATKSAMRecord> myReads = new LinkedHashSet<GATKSAMRecord>();
@Override
protected String getTraversalType() {
@ -101,7 +100,7 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
// Grab all the previously unseen reads from this pileup and add them to the massive read list
for( final PileupElement p : locus.getBasePileup() ) {
final SAMRecord read = p.getRead();
final GATKSAMRecord read = p.getRead();
if( !myReads.contains(read) ) {
myReads.add(read);
}
@ -111,7 +110,7 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
// which active regions in the work queue are now safe to process
if( !locusView.hasNext() ) {
for( final PileupElement p : locus.getBasePileup() ) {
final SAMRecord read = p.getRead();
final GATKSAMRecord read = p.getRead();
if( !myReads.contains(read) ) {
myReads.add(read);
}
@ -156,9 +155,9 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
return sum;
}
private T processActiveRegion( final ActiveRegion activeRegion, final LinkedHashSet<SAMRecord> reads, final Queue<ActiveRegion> workQueue, final T sum, final ActiveRegionWalker<M,T> walker ) {
final ArrayList<SAMRecord> placedReads = new ArrayList<SAMRecord>();
for( final SAMRecord read : reads ) {
private T processActiveRegion( final ActiveRegion activeRegion, final LinkedHashSet<GATKSAMRecord> reads, final Queue<ActiveRegion> workQueue, final T sum, final ActiveRegionWalker<M,T> walker ) {
final ArrayList<GATKSAMRecord> placedReads = new ArrayList<GATKSAMRecord>();
for( final GATKSAMRecord read : reads ) {
final GenomeLoc readLoc = this.engine.getGenomeLocParser().createGenomeLoc( read );
if( activeRegion.getLocation().overlapsP( readLoc ) ) {
// The region which the highest amount of overlap is chosen as the primary region for the read (tie breaking is done as right most region)
@ -170,22 +169,22 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
bestRegion = otherRegionToTest;
}
}
bestRegion.add( (GATKSAMRecord) read );
bestRegion.add( read );
// The read is also added to all other regions in which it overlaps but marked as non-primary
if( walker.wantsNonPrimaryReads() ) {
if( !bestRegion.equals(activeRegion) ) {
activeRegion.add( (GATKSAMRecord) read );
activeRegion.add( read );
}
for( final ActiveRegion otherRegionToTest : workQueue ) {
if( !bestRegion.equals(otherRegionToTest) && otherRegionToTest.getExtendedLoc().overlapsP( readLoc ) ) {
otherRegionToTest.add( (GATKSAMRecord) read );
otherRegionToTest.add( read );
}
}
}
placedReads.add( read );
} else if( activeRegion.getExtendedLoc().overlapsP( readLoc ) && walker.wantsNonPrimaryReads() ) {
activeRegion.add( (GATKSAMRecord) read );
activeRegion.add( read );
}
}
reads.removeAll( placedReads ); // remove all the reads which have been placed into their active region

2
public/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/CountCovariatesWalker.java
View File

@ -354,7 +354,7 @@ public class CountCovariatesWalker extends LocusWalker<CountCovariatesWalker.Cou
if( tracker.getValues(knownSites).size() == 0 ) { // If something here is in one of the knownSites tracks then skip over it, otherwise proceed
// For each read at this locus
for( final PileupElement p : context.getBasePileup() ) {
final GATKSAMRecord gatkRead = (GATKSAMRecord) p.getRead();
final GATKSAMRecord gatkRead = p.getRead();
int offset = p.getOffset();
if( gatkRead.containsTemporaryAttribute( SKIP_RECORD_ATTRIBUTE ) ) {

2
public/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/TableRecalibrationWalker.java
View File

@ -399,7 +399,7 @@ public class TableRecalibrationWalker extends ReadWalker<SAMRecord, SAMFileWrite
//compute all covariate values for this read
final Comparable[][] covariateValues_offset_x_covar =
RecalDataManager.computeCovariates((GATKSAMRecord) read, requestedCovariates);
RecalDataManager.computeCovariates(read, requestedCovariates);
// For each base in the read
for( int offset = 0; offset < read.getReadLength(); offset++ ) {

2
public/java/src/org/broadinstitute/sting/utils/baq/BAQ.java
View File

@ -673,7 +673,7 @@ public class BAQ {
}
/**
* Returns true if we don't think this read is eligable for the BAQ calculation. Examples include non-PF reads,
* Returns true if we don't think this read is eligible for the BAQ calculation. Examples include non-PF reads,
* duplicates, or unmapped reads. Used by baqRead to determine if a read should fall through the calculation.
*
* @param read