Brought all of ReduceReads to fastutils
-- Added unit tests to ReduceReads name compression -- Updated reduce reads walker for unit testing GSATDG-83
This commit is contained in:
Mauricio Carneiro
parent
2ad559cf58
commit
9e5a31b595
2
ivy.xml
2
ivy.xml
|
|
@ -41,6 +41,8 @@
|
|||
<dependency org="log4j" name="log4j" rev="1.2.15"/>
|
||||
<dependency org="javax.mail" name="mail" rev="1.4.4"/>
|
||||
<dependency org="colt" name="colt" rev="1.2.0"/>
|
||||
<dependency org="it.unimi.dsi" name="fastutil" rev="6.5.3" />
|
||||
|
||||
<!-- <dependency org="jboss" name="javassist" rev="3.7.ga"/> -->
|
||||
<dependency org="org.simpleframework" name="simple-xml" rev="2.0.4"/>
|
||||
<dependency org="org.apache.bcel" name="bcel" rev="5.2"/>
|
||||
|
|
|
|||
|
|
@ -46,10 +46,12 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.ObjectAVLTreeSet;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectSortedSet;
|
||||
import org.broadinstitute.sting.utils.*;
|
||||
|
||||
import java.util.Collection;
|
||||
import java.util.TreeSet;
|
||||
|
||||
|
||||
/**
|
||||
* A stash of regions that must be kept uncompressed in all samples
|
||||
|
|
@ -61,7 +63,7 @@ import java.util.TreeSet;
|
|||
* Date: 10/15/12
|
||||
* Time: 4:08 PM
|
||||
*/
|
||||
public class CompressionStash extends TreeSet<FinishedGenomeLoc> {
|
||||
public class CompressionStash extends ObjectAVLTreeSet<FinishedGenomeLoc> {
|
||||
public CompressionStash() {
|
||||
super();
|
||||
}
|
||||
|
|
@ -75,7 +77,7 @@ public class CompressionStash extends TreeSet<FinishedGenomeLoc> {
|
|||
*/
|
||||
@Override
|
||||
public boolean add(final FinishedGenomeLoc insertLoc) {
|
||||
TreeSet<FinishedGenomeLoc> removedLocs = new TreeSet<FinishedGenomeLoc>();
|
||||
ObjectSortedSet<FinishedGenomeLoc> removedLocs = new ObjectAVLTreeSet<FinishedGenomeLoc>();
|
||||
for (FinishedGenomeLoc existingLoc : this) {
|
||||
if (existingLoc.isPast(insertLoc)) {
|
||||
break; // if we're past the loc we're done looking for overlaps.
|
||||
|
|
|
|||
|
|
@ -46,10 +46,10 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.ints.IntArrayList;
|
||||
import org.broadinstitute.sting.utils.MathUtils;
|
||||
import org.broadinstitute.sting.utils.exceptions.ReviewedStingException;
|
||||
|
||||
import java.util.LinkedList;
|
||||
|
||||
/**
|
||||
* The element that describes the header of the sliding window.
|
||||
|
|
@ -64,7 +64,7 @@ public class HeaderElement {
|
|||
private int insertionsToTheRight; // How many reads in this site had insertions to the immediate right
|
||||
private int nSoftClippedBases; // How many bases in this site came from soft clipped bases
|
||||
private int location; // Genome location of this site (the sliding window knows which contig we're at
|
||||
private LinkedList<Integer> mappingQuality; // keeps the mapping quality of each read that contributed to this element (site)
|
||||
private IntArrayList mappingQuality; // keeps the mapping quality of each read that contributed to this element (site)
|
||||
|
||||
public int getLocation() {
|
||||
return location;
|
||||
|
|
@ -85,7 +85,7 @@ public class HeaderElement {
|
|||
* @param location the reference location for the new element
|
||||
*/
|
||||
public HeaderElement(final int location) {
|
||||
this(new BaseAndQualsCounts(), new BaseAndQualsCounts(), 0, 0, location, new LinkedList<Integer>());
|
||||
this(new BaseAndQualsCounts(), new BaseAndQualsCounts(), 0, 0, location, new IntArrayList());
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
@ -95,7 +95,7 @@ public class HeaderElement {
|
|||
* @param location the reference location for the new element
|
||||
*/
|
||||
public HeaderElement(final int location, final int insertionsToTheRight) {
|
||||
this(new BaseAndQualsCounts(), new BaseAndQualsCounts(), insertionsToTheRight, 0, location, new LinkedList<Integer>());
|
||||
this(new BaseAndQualsCounts(), new BaseAndQualsCounts(), insertionsToTheRight, 0, location, new IntArrayList());
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
@ -109,7 +109,7 @@ public class HeaderElement {
|
|||
* @param mappingQuality the list of mapping quality values of all reads that contributed to this
|
||||
* HeaderElement
|
||||
*/
|
||||
public HeaderElement(BaseAndQualsCounts consensusBaseCounts, BaseAndQualsCounts filteredBaseCounts, int insertionsToTheRight, int nSoftClippedBases, int location, LinkedList<Integer> mappingQuality) {
|
||||
public HeaderElement(BaseAndQualsCounts consensusBaseCounts, BaseAndQualsCounts filteredBaseCounts, int insertionsToTheRight, int nSoftClippedBases, int location, IntArrayList mappingQuality) {
|
||||
this.consensusBaseCounts = consensusBaseCounts;
|
||||
this.filteredBaseCounts = filteredBaseCounts;
|
||||
this.insertionsToTheRight = insertionsToTheRight;
|
||||
|
|
|
|||
|
|
@ -46,6 +46,7 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.*;
|
||||
import net.sf.samtools.SAMFileHeader;
|
||||
import org.apache.log4j.Logger;
|
||||
import org.broadinstitute.sting.utils.SampleUtils;
|
||||
|
|
@ -54,10 +55,6 @@ import org.broadinstitute.sting.utils.exceptions.ReviewedStingException;
|
|||
import org.broadinstitute.sting.utils.sam.AlignmentStartWithNoTiesComparator;
|
||||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
|
||||
import java.util.HashMap;
|
||||
import java.util.Map;
|
||||
import java.util.Set;
|
||||
import java.util.TreeSet;
|
||||
|
||||
/*
|
||||
* Copyright (c) 2009 The Broad Institute
|
||||
|
|
@ -91,7 +88,7 @@ import java.util.TreeSet;
|
|||
public class MultiSampleCompressor {
|
||||
protected static final Logger logger = Logger.getLogger(MultiSampleCompressor.class);
|
||||
|
||||
protected Map<String, SingleSampleCompressor> compressorsPerSample = new HashMap<String, SingleSampleCompressor>();
|
||||
protected Object2ObjectMap<String, SingleSampleCompressor> compressorsPerSample = new Object2ObjectOpenHashMap<String, SingleSampleCompressor>();
|
||||
|
||||
public MultiSampleCompressor(SAMFileHeader header,
|
||||
final int contextSize,
|
||||
|
|
@ -109,13 +106,13 @@ public class MultiSampleCompressor {
|
|||
}
|
||||
}
|
||||
|
||||
public Set<GATKSAMRecord> addAlignment(GATKSAMRecord read) {
|
||||
public ObjectSet<GATKSAMRecord> addAlignment(GATKSAMRecord read) {
|
||||
String sampleName = read.getReadGroup().getSample();
|
||||
SingleSampleCompressor compressor = compressorsPerSample.get(sampleName);
|
||||
if ( compressor == null )
|
||||
throw new ReviewedStingException("No compressor for sample " + sampleName);
|
||||
Pair<Set<GATKSAMRecord>, CompressionStash> readsAndStash = compressor.addAlignment(read);
|
||||
Set<GATKSAMRecord> reads = readsAndStash.getFirst();
|
||||
Pair<ObjectSet<GATKSAMRecord>, CompressionStash> readsAndStash = compressor.addAlignment(read);
|
||||
ObjectSet<GATKSAMRecord> reads = readsAndStash.getFirst();
|
||||
CompressionStash regions = readsAndStash.getSecond();
|
||||
|
||||
reads.addAll(closeVariantRegionsInAllSamples(regions));
|
||||
|
|
@ -123,17 +120,17 @@ public class MultiSampleCompressor {
|
|||
return reads;
|
||||
}
|
||||
|
||||
public Set<GATKSAMRecord> close() {
|
||||
Set<GATKSAMRecord> reads = new TreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
public ObjectSet<GATKSAMRecord> close() {
|
||||
ObjectSet<GATKSAMRecord> reads = new ObjectAVLTreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
for ( SingleSampleCompressor sample : compressorsPerSample.values() ) {
|
||||
Pair<Set<GATKSAMRecord>, CompressionStash> readsAndStash = sample.close();
|
||||
Pair<ObjectSet<GATKSAMRecord>, CompressionStash> readsAndStash = sample.close();
|
||||
reads = readsAndStash.getFirst();
|
||||
}
|
||||
return reads;
|
||||
}
|
||||
|
||||
private Set<GATKSAMRecord> closeVariantRegionsInAllSamples(CompressionStash regions) {
|
||||
Set<GATKSAMRecord> reads = new TreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
private ObjectSet<GATKSAMRecord> closeVariantRegionsInAllSamples(CompressionStash regions) {
|
||||
ObjectSet<GATKSAMRecord> reads = new ObjectAVLTreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
if (!regions.isEmpty()) {
|
||||
for (SingleSampleCompressor sample : compressorsPerSample.values()) {
|
||||
reads.addAll(sample.closeVariantRegions(regions));
|
||||
|
|
|
|||
|
|
@ -46,6 +46,10 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.Object2LongOpenHashMap;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectAVLTreeSet;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectSortedSet;
|
||||
import net.sf.samtools.SAMFileHeader;
|
||||
import net.sf.samtools.SAMFileWriter;
|
||||
import net.sf.samtools.SAMProgramRecord;
|
||||
|
|
@ -71,7 +75,6 @@ import org.broadinstitute.sting.utils.sam.BySampleSAMFileWriter;
|
|||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
import org.broadinstitute.sting.utils.sam.ReadUtils;
|
||||
|
||||
import java.util.*;
|
||||
|
||||
/**
|
||||
* Reduces the BAM file using read based compression that keeps only essential information for variant calling
|
||||
|
|
@ -107,7 +110,7 @@ import java.util.*;
|
|||
@PartitionBy(PartitionType.CONTIG)
|
||||
@ReadFilters({UnmappedReadFilter.class, NotPrimaryAlignmentFilter.class, DuplicateReadFilter.class, FailsVendorQualityCheckFilter.class, BadCigarFilter.class})
|
||||
@Downsample(by=DownsampleType.BY_SAMPLE, toCoverage=40)
|
||||
public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceReadsStash> {
|
||||
public class ReduceReads extends ReadWalker<ObjectArrayList<GATKSAMRecord>, ReduceReadsStash> {
|
||||
|
||||
@Output
|
||||
private StingSAMFileWriter out = null;
|
||||
|
|
@ -240,10 +243,10 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
|
||||
int nCompressedReads = 0;
|
||||
|
||||
HashMap<String, Long> readNameHash; // This hash will keep the name of the original read the new compressed name (a number).
|
||||
Object2LongOpenHashMap<String> readNameHash; // This hash will keep the name of the original read the new compressed name (a number).
|
||||
Long nextReadNumber = 1L; // The next number to use for the compressed read name.
|
||||
|
||||
SortedSet<GenomeLoc> intervalList;
|
||||
ObjectSortedSet<GenomeLoc> intervalList;
|
||||
|
||||
// IMPORTANT: DO NOT CHANGE THE VALUE OF THIS CONSTANT VARIABLE; IT IS NOW PERMANENTLY THE @PG NAME THAT EXTERNAL TOOLS LOOK FOR IN THE BAM HEADER
|
||||
public static final String PROGRAM_RECORD_NAME = "GATK ReduceReads"; // The name that will go in the @PG tag
|
||||
|
|
@ -257,8 +260,8 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
public void initialize() {
|
||||
super.initialize();
|
||||
GenomeAnalysisEngine toolkit = getToolkit();
|
||||
readNameHash = new HashMap<String, Long>(); // prepare the read name hash to keep track of what reads have had their read names compressed
|
||||
intervalList = new TreeSet<GenomeLoc>(); // get the interval list from the engine. If no interval list was provided, the walker will work in WGS mode
|
||||
readNameHash = new Object2LongOpenHashMap<String>(100000); // prepare the read name hash to keep track of what reads have had their read names compressed
|
||||
intervalList = new ObjectAVLTreeSet<GenomeLoc>(); // get the interval list from the engine. If no interval list was provided, the walker will work in WGS mode
|
||||
|
||||
if (toolkit.getIntervals() != null)
|
||||
intervalList.addAll(toolkit.getIntervals());
|
||||
|
|
@ -295,8 +298,8 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
* @return a linked list with all the reads produced by the clipping operations
|
||||
*/
|
||||
@Override
|
||||
public LinkedList<GATKSAMRecord> map(ReferenceContext ref, GATKSAMRecord read, RefMetaDataTracker metaDataTracker) {
|
||||
LinkedList<GATKSAMRecord> mappedReads;
|
||||
public ObjectArrayList<GATKSAMRecord> map(ReferenceContext ref, GATKSAMRecord read, RefMetaDataTracker metaDataTracker) {
|
||||
ObjectArrayList<GATKSAMRecord> mappedReads;
|
||||
if (!debugRead.isEmpty() && read.getReadName().contains(debugRead))
|
||||
System.out.println("Found debug read!");
|
||||
|
||||
|
|
@ -325,18 +328,18 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
if (HARD_CLIP_TO_INTERVAL)
|
||||
mappedReads = hardClipReadToInterval(read); // Hard clip the remainder of the read to the desired interval
|
||||
else {
|
||||
mappedReads = new LinkedList<GATKSAMRecord>();
|
||||
mappedReads = new ObjectArrayList<GATKSAMRecord>();
|
||||
mappedReads.add(read);
|
||||
}
|
||||
}
|
||||
else {
|
||||
mappedReads = new LinkedList<GATKSAMRecord>();
|
||||
mappedReads = new ObjectArrayList<GATKSAMRecord>();
|
||||
if (!read.isEmpty())
|
||||
mappedReads.add(read);
|
||||
}
|
||||
|
||||
if (!mappedReads.isEmpty() && !DONT_USE_SOFTCLIPPED_BASES) {
|
||||
LinkedList<GATKSAMRecord> tempList = new LinkedList<GATKSAMRecord>();
|
||||
ObjectArrayList<GATKSAMRecord> tempList = new ObjectArrayList<GATKSAMRecord>();
|
||||
for (GATKSAMRecord mRead : mappedReads) {
|
||||
GATKSAMRecord clippedRead = ReadClipper.hardClipLowQualitySoftClips(mRead, minBaseQual);
|
||||
if (!clippedRead.isEmpty())
|
||||
|
|
@ -375,7 +378,7 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
* @param stash the stash that keeps the reads in order for processing
|
||||
* @return the stash with all reads that have not been processed yet
|
||||
*/
|
||||
public ReduceReadsStash reduce(LinkedList<GATKSAMRecord> mappedReads, ReduceReadsStash stash) {
|
||||
public ReduceReadsStash reduce(ObjectArrayList<GATKSAMRecord> mappedReads, ReduceReadsStash stash) {
|
||||
if (debugLevel == 1)
|
||||
stash.print();
|
||||
|
||||
|
|
@ -387,7 +390,7 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
throw new ReviewedStingException("Empty read sent to reduce, this should never happen! " + read.getReadName() + " -- " + read.getCigar() + " -- " + read.getReferenceName() + ":" + read.getAlignmentStart() + "-" + read.getAlignmentEnd());
|
||||
|
||||
if (originalRead) {
|
||||
List<GATKSAMRecord> readsReady = new LinkedList<GATKSAMRecord>();
|
||||
ObjectArrayList<GATKSAMRecord> readsReady = new ObjectArrayList<GATKSAMRecord>();
|
||||
readsReady.addAll(stash.getAllReadsBefore(read));
|
||||
readsReady.add(read);
|
||||
|
||||
|
|
@ -433,8 +436,8 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
* @param read the read to be hard clipped to the interval.
|
||||
* @return a shallow copy of the read hard clipped to the interval
|
||||
*/
|
||||
private LinkedList<GATKSAMRecord> hardClipReadToInterval(GATKSAMRecord read) {
|
||||
LinkedList<GATKSAMRecord> clippedReads = new LinkedList<GATKSAMRecord>();
|
||||
private ObjectArrayList<GATKSAMRecord> hardClipReadToInterval(GATKSAMRecord read) {
|
||||
ObjectArrayList<GATKSAMRecord> clippedReads = new ObjectArrayList<GATKSAMRecord>();
|
||||
|
||||
GenomeLoc intervalOverlapped = null; // marks the interval to which the original read overlapped (so we can cut all previous intervals from the list)
|
||||
|
||||
|
|
@ -588,7 +591,7 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
System.out.println("BAM: " + read.getCigar() + " " + read.getAlignmentStart() + " " + read.getAlignmentEnd());
|
||||
|
||||
if (!DONT_COMPRESS_READ_NAMES)
|
||||
compressReadName(read);
|
||||
nextReadNumber = compressReadName(readNameHash, read, nextReadNumber);
|
||||
|
||||
writerToUse.addAlignment(read);
|
||||
}
|
||||
|
|
@ -625,19 +628,20 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
*
|
||||
* @param read any read
|
||||
*/
|
||||
private void compressReadName(GATKSAMRecord read) {
|
||||
String name = read.getReadName();
|
||||
protected static long compressReadName(Object2LongOpenHashMap<String> hash, GATKSAMRecord read, long nextReadNumber) {
|
||||
final String name = read.getReadName();
|
||||
long result = nextReadNumber;
|
||||
String compressedName = read.isReducedRead() ? "C" : "";
|
||||
final Long readNumber = readNameHash.get(name);
|
||||
final Long readNumber = hash.get(name);
|
||||
if (readNumber != null) {
|
||||
compressedName += readNumber.toString();
|
||||
} else {
|
||||
readNameHash.put(name, nextReadNumber);
|
||||
compressedName += nextReadNumber.toString();
|
||||
nextReadNumber++;
|
||||
hash.put(name, nextReadNumber);
|
||||
compressedName += "" + nextReadNumber;
|
||||
result++;
|
||||
}
|
||||
|
||||
read.setReadName(compressedName);
|
||||
return result;
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
@ -649,8 +653,8 @@ public class ReduceReads extends ReadWalker<LinkedList<GATKSAMRecord>, ReduceRea
|
|||
* @param read the read
|
||||
* @return Returns true if the read is the original read that went through map().
|
||||
*/
|
||||
private boolean isOriginalRead(LinkedList<GATKSAMRecord> list, GATKSAMRecord read) {
|
||||
return isWholeGenome() || list.getFirst().equals(read);
|
||||
private boolean isOriginalRead(ObjectArrayList<GATKSAMRecord> list, GATKSAMRecord read) {
|
||||
return isWholeGenome() || list.get(0).equals(read);
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
|
|||
|
|
@ -46,14 +46,11 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.*;
|
||||
import org.broadinstitute.sting.utils.collections.Pair;
|
||||
import org.broadinstitute.sting.utils.sam.AlignmentStartWithNoTiesComparator;
|
||||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
|
||||
import java.util.Collections;
|
||||
import java.util.Set;
|
||||
import java.util.TreeSet;
|
||||
|
||||
/**
|
||||
*
|
||||
* @author carneiro, depristo
|
||||
|
|
@ -72,7 +69,7 @@ public class SingleSampleCompressor {
|
|||
private SlidingWindow slidingWindow;
|
||||
private int slidingWindowCounter;
|
||||
|
||||
public static Pair<Set<GATKSAMRecord>, CompressionStash> emptyPair = new Pair<Set<GATKSAMRecord>,CompressionStash>(new TreeSet<GATKSAMRecord>(), new CompressionStash());
|
||||
public static Pair<ObjectSet<GATKSAMRecord>, CompressionStash> emptyPair = new Pair<ObjectSet<GATKSAMRecord>,CompressionStash>(new ObjectAVLTreeSet<GATKSAMRecord>(), new CompressionStash());
|
||||
|
||||
public SingleSampleCompressor(final int contextSize,
|
||||
final int downsampleCoverage,
|
||||
|
|
@ -93,8 +90,8 @@ public class SingleSampleCompressor {
|
|||
this.allowPolyploidReduction = allowPolyploidReduction;
|
||||
}
|
||||
|
||||
public Pair<Set<GATKSAMRecord>, CompressionStash> addAlignment( GATKSAMRecord read ) {
|
||||
Set<GATKSAMRecord> reads = new TreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
public Pair<ObjectSet<GATKSAMRecord>, CompressionStash> addAlignment( GATKSAMRecord read ) {
|
||||
ObjectSet<GATKSAMRecord> reads = new ObjectAVLTreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
CompressionStash stash = new CompressionStash();
|
||||
int readOriginalStart = read.getUnclippedStart();
|
||||
|
||||
|
|
@ -104,7 +101,7 @@ public class SingleSampleCompressor {
|
|||
(readOriginalStart - contextSize > slidingWindow.getStopLocation()))) { // this read is too far away from the end of the current sliding window
|
||||
|
||||
// close the current sliding window
|
||||
Pair<Set<GATKSAMRecord>, CompressionStash> readsAndStash = slidingWindow.close();
|
||||
Pair<ObjectSet<GATKSAMRecord>, CompressionStash> readsAndStash = slidingWindow.close();
|
||||
reads = readsAndStash.getFirst();
|
||||
stash = readsAndStash.getSecond();
|
||||
slidingWindow = null; // so we create a new one on the next if
|
||||
|
|
@ -116,15 +113,15 @@ public class SingleSampleCompressor {
|
|||
}
|
||||
|
||||
stash.addAll(slidingWindow.addRead(read));
|
||||
return new Pair<Set<GATKSAMRecord>, CompressionStash>(reads, stash);
|
||||
return new Pair<ObjectSet<GATKSAMRecord>, CompressionStash>(reads, stash);
|
||||
}
|
||||
|
||||
public Pair<Set<GATKSAMRecord>, CompressionStash> close() {
|
||||
public Pair<ObjectSet<GATKSAMRecord>, CompressionStash> close() {
|
||||
return (slidingWindow != null) ? slidingWindow.close() : emptyPair;
|
||||
}
|
||||
|
||||
public Set<GATKSAMRecord> closeVariantRegions(CompressionStash regions) {
|
||||
return slidingWindow == null ? Collections.<GATKSAMRecord>emptySet() : slidingWindow.closeVariantRegions(regions);
|
||||
public ObjectSet<GATKSAMRecord> closeVariantRegions(CompressionStash regions) {
|
||||
return slidingWindow == null ? ObjectSets.EMPTY_SET : slidingWindow.closeVariantRegions(regions);
|
||||
}
|
||||
|
||||
}
|
||||
|
|
|
|||
|
|
@ -48,6 +48,10 @@ package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
|||
|
||||
import com.google.java.contract.Ensures;
|
||||
import com.google.java.contract.Requires;
|
||||
import it.unimi.dsi.fastutil.bytes.Byte2IntArrayMap;
|
||||
import it.unimi.dsi.fastutil.bytes.Byte2IntMap;
|
||||
import it.unimi.dsi.fastutil.bytes.Byte2IntOpenHashMap;
|
||||
import it.unimi.dsi.fastutil.objects.*;
|
||||
import net.sf.samtools.Cigar;
|
||||
import net.sf.samtools.CigarElement;
|
||||
import net.sf.samtools.CigarOperator;
|
||||
|
|
@ -62,7 +66,11 @@ import org.broadinstitute.sting.utils.sam.GATKSAMReadGroupRecord;
|
|||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
import org.broadinstitute.sting.utils.sam.ReadUtils;
|
||||
|
||||
import java.util.*;
|
||||
import java.util.Comparator;
|
||||
import java.util.Iterator;
|
||||
import java.util.LinkedList;
|
||||
import java.util.ListIterator;
|
||||
|
||||
|
||||
/**
|
||||
* Created by IntelliJ IDEA.
|
||||
|
|
@ -73,7 +81,7 @@ import java.util.*;
|
|||
public class SlidingWindow {
|
||||
|
||||
// Sliding Window data
|
||||
final private TreeSet<GATKSAMRecord> readsInWindow;
|
||||
final private ObjectAVLTreeSet<GATKSAMRecord> readsInWindow;
|
||||
final private LinkedList<HeaderElement> windowHeader;
|
||||
protected int contextSize; // the largest context size (between mismatches and indels)
|
||||
protected String contig;
|
||||
|
|
@ -144,7 +152,7 @@ public class SlidingWindow {
|
|||
|
||||
this.windowHeader = new LinkedList<HeaderElement>();
|
||||
windowHeader.addFirst(new HeaderElement(startLocation));
|
||||
this.readsInWindow = new TreeSet<GATKSAMRecord>();
|
||||
this.readsInWindow = new ObjectAVLTreeSet<GATKSAMRecord>();
|
||||
}
|
||||
|
||||
public SlidingWindow(String contig, int contigIndex, int contextSize, SAMFileHeader samHeader, GATKSAMReadGroupRecord readGroupAttribute, int windowNumber, final double minAltProportionToTriggerVariant, final double minIndelProportionToTriggerVariant, int minBaseQual, int minMappingQuality, int downsampleCoverage, final ReduceReads.DownsampleStrategy downsampleStrategy, boolean hasIndelQualities, boolean allowPolyploidReduction) {
|
||||
|
|
@ -157,7 +165,7 @@ public class SlidingWindow {
|
|||
this.MIN_MAPPING_QUALITY = minMappingQuality;
|
||||
|
||||
this.windowHeader = new LinkedList<HeaderElement>();
|
||||
this.readsInWindow = new TreeSet<GATKSAMRecord>(new Comparator<GATKSAMRecord>() {
|
||||
this.readsInWindow = new ObjectAVLTreeSet<GATKSAMRecord>(new Comparator<GATKSAMRecord>() {
|
||||
@Override
|
||||
public int compare(GATKSAMRecord read1, GATKSAMRecord read2) {
|
||||
final int difference = read1.getSoftEnd() - read2.getSoftEnd();
|
||||
|
|
@ -287,7 +295,7 @@ public class SlidingWindow {
|
|||
}
|
||||
|
||||
while (!readsInWindow.isEmpty() && readsInWindow.first().getSoftEnd() < windowHeaderStartLocation) {
|
||||
readsInWindow.pollFirst();
|
||||
readsInWindow.remove(readsInWindow.first());
|
||||
}
|
||||
|
||||
return regions;
|
||||
|
|
@ -401,8 +409,8 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"start >= 0 && (end >= start || end == 0)"})
|
||||
@Ensures("result != null")
|
||||
protected List<GATKSAMRecord> addToSyntheticReads(LinkedList<HeaderElement> header, int start, int end, boolean isNegativeStrand) {
|
||||
LinkedList<GATKSAMRecord> reads = new LinkedList<GATKSAMRecord>();
|
||||
protected ObjectArrayList<GATKSAMRecord> addToSyntheticReads(LinkedList<HeaderElement> header, int start, int end, boolean isNegativeStrand) {
|
||||
ObjectArrayList<GATKSAMRecord> reads = new ObjectArrayList<GATKSAMRecord>();
|
||||
if (start < end) {
|
||||
ListIterator<HeaderElement> headerElementIterator = header.listIterator(start);
|
||||
|
||||
|
|
@ -454,9 +462,9 @@ public class SlidingWindow {
|
|||
* @param type the synthetic reads you want to close
|
||||
* @return a possibly null list of GATKSAMRecords generated by finalizing the synthetic reads
|
||||
*/
|
||||
private List<GATKSAMRecord> finalizeAndAdd(ConsensusType type) {
|
||||
private ObjectArrayList<GATKSAMRecord> finalizeAndAdd(ConsensusType type) {
|
||||
GATKSAMRecord read = null;
|
||||
List<GATKSAMRecord> list = new LinkedList<GATKSAMRecord>();
|
||||
ObjectArrayList<GATKSAMRecord> list = new ObjectArrayList<GATKSAMRecord>();
|
||||
|
||||
switch (type) {
|
||||
case CONSENSUS:
|
||||
|
|
@ -556,8 +564,8 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"start >= 0 && (end >= start || end == 0)"})
|
||||
@Ensures("result != null")
|
||||
private List<GATKSAMRecord> addToFilteredData(LinkedList<HeaderElement> header, int start, int end, boolean isNegativeStrand) {
|
||||
List<GATKSAMRecord> result = new ArrayList<GATKSAMRecord>(0);
|
||||
private ObjectArrayList<GATKSAMRecord> addToFilteredData(LinkedList<HeaderElement> header, int start, int end, boolean isNegativeStrand) {
|
||||
ObjectArrayList<GATKSAMRecord> result = new ObjectArrayList<GATKSAMRecord>();
|
||||
|
||||
if (filteredDataConsensus == null)
|
||||
filteredDataConsensus = new SyntheticRead(samHeader, readGroupAttribute, contig, contigIndex, filteredDataReadName + filteredDataConsensusCounter++, header.get(start).getLocation(), GATKSAMRecord.REDUCED_READ_CONSENSUS_TAG, hasIndelQualities, isNegativeStrand);
|
||||
|
|
@ -640,8 +648,8 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"start >= 0 && (stop >= start || stop == 0)"})
|
||||
@Ensures("result != null")
|
||||
protected List<GATKSAMRecord> compressVariantRegion(final int start, final int stop, final boolean disallowPolyploidReductionAtThisPosition) {
|
||||
List<GATKSAMRecord> allReads = new LinkedList<GATKSAMRecord>();
|
||||
protected ObjectList<GATKSAMRecord> compressVariantRegion(final int start, final int stop, final boolean disallowPolyploidReductionAtThisPosition) {
|
||||
ObjectList<GATKSAMRecord> allReads = new ObjectArrayList<GATKSAMRecord>();
|
||||
|
||||
// Try to compress into a polyploid consensus
|
||||
int nVariantPositions = 0;
|
||||
|
|
@ -685,7 +693,7 @@ public class SlidingWindow {
|
|||
final int refStart = windowHeader.get(start).getLocation();
|
||||
final int refStop = windowHeader.get(stop).getLocation();
|
||||
|
||||
LinkedList<GATKSAMRecord> toRemove = new LinkedList<GATKSAMRecord>();
|
||||
ObjectList<GATKSAMRecord> toRemove = new ObjectArrayList<GATKSAMRecord>();
|
||||
for (GATKSAMRecord read : readsInWindow) {
|
||||
if (read.getSoftStart() <= refStop) {
|
||||
if (read.getAlignmentEnd() >= refStart) {
|
||||
|
|
@ -710,24 +718,24 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"start >= 0 && (stop >= start || stop == 0)"})
|
||||
@Ensures("result != null")
|
||||
protected List<GATKSAMRecord> closeVariantRegion(final int start, final int stop, final boolean disallowPolyploidReductionAtThisPosition) {
|
||||
List<GATKSAMRecord> allReads = compressVariantRegion(start, stop, disallowPolyploidReductionAtThisPosition);
|
||||
protected ObjectList<GATKSAMRecord> closeVariantRegion(final int start, final int stop, final boolean disallowPolyploidReductionAtThisPosition) {
|
||||
ObjectList<GATKSAMRecord> allReads = compressVariantRegion(start, stop, disallowPolyploidReductionAtThisPosition);
|
||||
|
||||
List<GATKSAMRecord> result = (downsampleCoverage > 0) ? downsampleVariantRegion(allReads) : allReads;
|
||||
ObjectList<GATKSAMRecord> result = (downsampleCoverage > 0) ? downsampleVariantRegion(allReads) : allReads;
|
||||
result.addAll(addToSyntheticReads(windowHeader, 0, stop, false));
|
||||
result.addAll(finalizeAndAdd(ConsensusType.BOTH));
|
||||
|
||||
return result; // finalized reads will be downsampled if necessary
|
||||
}
|
||||
|
||||
public Set<GATKSAMRecord> closeVariantRegions(CompressionStash regions) {
|
||||
TreeSet<GATKSAMRecord> allReads = new TreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
public ObjectSet<GATKSAMRecord> closeVariantRegions(CompressionStash regions) {
|
||||
ObjectAVLTreeSet<GATKSAMRecord> allReads = new ObjectAVLTreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
if (!regions.isEmpty()) {
|
||||
int lastStop = -1;
|
||||
int windowHeaderStart = getStartLocation(windowHeader);
|
||||
|
||||
for (GenomeLoc region : regions) {
|
||||
if (((FinishedGenomeLoc)region).isFinished() && region.getContig() == contig && region.getStart() >= windowHeaderStart && region.getStop() < windowHeaderStart + windowHeader.size()) {
|
||||
if (((FinishedGenomeLoc)region).isFinished() && region.getContig().equals(contig) && region.getStart() >= windowHeaderStart && region.getStop() < windowHeaderStart + windowHeader.size()) {
|
||||
int start = region.getStart() - windowHeaderStart;
|
||||
int stop = region.getStop() - windowHeaderStart;
|
||||
|
||||
|
|
@ -759,7 +767,7 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"allReads != null"})
|
||||
@Ensures("result != null")
|
||||
protected List<GATKSAMRecord> downsampleVariantRegion(final List<GATKSAMRecord> allReads) {
|
||||
protected ObjectList<GATKSAMRecord> downsampleVariantRegion(final ObjectList<GATKSAMRecord> allReads) {
|
||||
int nReads = allReads.size();
|
||||
if (nReads == 0)
|
||||
return allReads;
|
||||
|
|
@ -769,7 +777,7 @@ public class SlidingWindow {
|
|||
|
||||
ReservoirDownsampler <GATKSAMRecord> downsampler = new ReservoirDownsampler<GATKSAMRecord>(downsampleCoverage);
|
||||
downsampler.submit(allReads);
|
||||
return downsampler.consumeFinalizedItems();
|
||||
return new ObjectArrayList<GATKSAMRecord>(downsampler.consumeFinalizedItems());
|
||||
}
|
||||
|
||||
|
||||
|
|
@ -781,9 +789,9 @@ public class SlidingWindow {
|
|||
* @return A non-null set/list of all reads generated
|
||||
*/
|
||||
@Ensures("result != null")
|
||||
public Pair<Set<GATKSAMRecord>, CompressionStash> close() {
|
||||
public Pair<ObjectSet<GATKSAMRecord>, CompressionStash> close() {
|
||||
// mark variant regions
|
||||
Set<GATKSAMRecord> finalizedReads = new TreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
ObjectSet<GATKSAMRecord> finalizedReads = new ObjectAVLTreeSet<GATKSAMRecord>(new AlignmentStartWithNoTiesComparator());
|
||||
CompressionStash regions = new CompressionStash();
|
||||
boolean forceCloseUnfinishedRegions = true;
|
||||
|
||||
|
|
@ -798,7 +806,7 @@ public class SlidingWindow {
|
|||
}
|
||||
}
|
||||
|
||||
return new Pair<Set<GATKSAMRecord>, CompressionStash>(finalizedReads, regions);
|
||||
return new Pair<ObjectSet<GATKSAMRecord>, CompressionStash>(finalizedReads, regions);
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
@ -847,16 +855,16 @@ public class SlidingWindow {
|
|||
*/
|
||||
@Requires({"start >= 0 && (stop >= start || stop == 0)"})
|
||||
@Ensures("result != null")
|
||||
private List<GATKSAMRecord> createPolyploidConsensus(final int start, final int stop, final int hetRefPosition) {
|
||||
private ObjectList<GATKSAMRecord> createPolyploidConsensus(final int start, final int stop, final int hetRefPosition) {
|
||||
// we will create two (positive strand, negative strand) headers for each contig
|
||||
List<LinkedList<HeaderElement>> headersPosStrand = new ArrayList<LinkedList<HeaderElement>>();
|
||||
List<LinkedList<HeaderElement>> headersNegStrand = new ArrayList<LinkedList<HeaderElement>>();
|
||||
List<GATKSAMRecord> hetReads = new LinkedList<GATKSAMRecord>();
|
||||
Map<Byte, Integer> haplotypeHeaderMap = new HashMap<Byte, Integer>(2);
|
||||
ObjectList<LinkedList<HeaderElement>> headersPosStrand = new ObjectArrayList<LinkedList<HeaderElement>>();
|
||||
ObjectList<LinkedList<HeaderElement>> headersNegStrand = new ObjectArrayList<LinkedList<HeaderElement>>();
|
||||
ObjectList<GATKSAMRecord> hetReads = new ObjectArrayList<GATKSAMRecord>();
|
||||
Byte2IntMap haplotypeHeaderMap = new Byte2IntArrayMap(2);
|
||||
int currentHaplotype = 0;
|
||||
int refStart = windowHeader.get(start).getLocation();
|
||||
int refStop = windowHeader.get(stop).getLocation();
|
||||
List<GATKSAMRecord> toRemove = new LinkedList<GATKSAMRecord>();
|
||||
ObjectList<GATKSAMRecord> toRemove = new ObjectArrayList<GATKSAMRecord>();
|
||||
for (GATKSAMRecord read : readsInWindow) {
|
||||
int haplotype;
|
||||
|
||||
|
|
@ -1031,7 +1039,7 @@ public class SlidingWindow {
|
|||
}
|
||||
}
|
||||
|
||||
private void removeReadsFromWindow (List<GATKSAMRecord> readsToRemove) {
|
||||
private void removeReadsFromWindow (ObjectList<GATKSAMRecord> readsToRemove) {
|
||||
for (GATKSAMRecord read : readsToRemove) {
|
||||
readsInWindow.remove(read);
|
||||
}
|
||||
|
|
|
|||
|
|
@ -47,6 +47,8 @@
|
|||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import com.google.java.contract.Requires;
|
||||
import it.unimi.dsi.fastutil.bytes.ByteArrayList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
|
||||
import net.sf.samtools.Cigar;
|
||||
import net.sf.samtools.CigarElement;
|
||||
import net.sf.samtools.CigarOperator;
|
||||
|
|
@ -57,10 +59,8 @@ import org.broadinstitute.sting.utils.recalibration.EventType;
|
|||
import org.broadinstitute.sting.utils.sam.GATKSAMReadGroupRecord;
|
||||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.Iterator;
|
||||
import java.util.LinkedList;
|
||||
import java.util.List;
|
||||
|
||||
|
||||
/**
|
||||
* Running Consensus is a read that is compressed as a sliding window travels over the reads
|
||||
|
|
@ -123,7 +123,7 @@ public class SyntheticRead {
|
|||
}
|
||||
|
||||
|
||||
private final List<SingleBaseInfo> basesCountsQuals;
|
||||
private final ObjectArrayList<SingleBaseInfo> basesCountsQuals;
|
||||
private double mappingQuality; // the average of the rms of the mapping qualities of all the reads that contributed to this consensus
|
||||
private String readTag;
|
||||
|
||||
|
|
@ -151,7 +151,7 @@ public class SyntheticRead {
|
|||
*/
|
||||
public SyntheticRead(SAMFileHeader header, GATKSAMReadGroupRecord readGroupRecord, String contig, int contigIndex, String readName, int refStart, String readTag, boolean hasIndelQualities, boolean isNegativeRead) {
|
||||
final int initialCapacity = 10000;
|
||||
basesCountsQuals = new ArrayList<SingleBaseInfo>(initialCapacity);
|
||||
basesCountsQuals = new ObjectArrayList<SingleBaseInfo>(initialCapacity);
|
||||
mappingQuality = 0.0;
|
||||
|
||||
this.readTag = readTag;
|
||||
|
|
@ -165,8 +165,8 @@ public class SyntheticRead {
|
|||
this.isNegativeStrand = isNegativeRead;
|
||||
}
|
||||
|
||||
public SyntheticRead(List<BaseIndex> bases, List<Byte> counts, List<Byte> quals, List<Byte> insertionQuals, List<Byte> deletionQuals, double mappingQuality, String readTag, SAMFileHeader header, GATKSAMReadGroupRecord readGroupRecord, String contig, int contigIndex, String readName, int refStart, boolean hasIndelQualities, boolean isNegativeRead) {
|
||||
basesCountsQuals = new ArrayList<SingleBaseInfo>(bases.size());
|
||||
public SyntheticRead(ObjectArrayList<BaseIndex> bases, ByteArrayList counts, ByteArrayList quals, ByteArrayList insertionQuals, ByteArrayList deletionQuals, double mappingQuality, String readTag, SAMFileHeader header, GATKSAMReadGroupRecord readGroupRecord, String contig, int contigIndex, String readName, int refStart, boolean hasIndelQualities, boolean isNegativeRead) {
|
||||
basesCountsQuals = new ObjectArrayList<SingleBaseInfo>(bases.size());
|
||||
for (int i = 0; i < bases.size(); ++i) {
|
||||
basesCountsQuals.add(new SingleBaseInfo(bases.get(i).getOrdinalByte(), counts.get(i), quals.get(i), insertionQuals.get(i), deletionQuals.get(i)));
|
||||
}
|
||||
|
|
@ -316,7 +316,7 @@ public class SyntheticRead {
|
|||
* @return the cigar string for the synthetic read
|
||||
*/
|
||||
private Cigar buildCigar() {
|
||||
LinkedList<CigarElement> cigarElements = new LinkedList<CigarElement>();
|
||||
ObjectArrayList<CigarElement> cigarElements = new ObjectArrayList<CigarElement>();
|
||||
CigarOperator cigarOperator = null;
|
||||
int length = 0;
|
||||
for (final SingleBaseInfo singleBaseInfo : basesCountsQuals) {
|
||||
|
|
|
|||
|
|
@ -0,0 +1,111 @@
|
|||
/*
|
||||
* 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.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.Object2LongOpenHashMap;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectOpenHashSet;
|
||||
import org.broadinstitute.sting.BaseTest;
|
||||
import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
||||
import org.testng.Assert;
|
||||
import org.testng.annotations.DataProvider;
|
||||
import org.testng.annotations.Test;
|
||||
|
||||
import java.util.Random;
|
||||
|
||||
|
||||
public class ReduceReadsUnitTest extends BaseTest {
|
||||
|
||||
Random random = new Random(987743);
|
||||
Object2LongOpenHashMap<String> hash = new Object2LongOpenHashMap<String>();
|
||||
long nextNumber = 0L;
|
||||
|
||||
/**
|
||||
* Combinatorial unit test data provider example.
|
||||
*
|
||||
* Creates data for testMyData test function, containing two arguments, start and size at each value
|
||||
*
|
||||
* @return Object[][] for testng DataProvider
|
||||
*/
|
||||
@DataProvider(name = "ReadNameProvider")
|
||||
public Object[][] readNameProvider() {
|
||||
final int readNameLength = 4;
|
||||
final int nReads = 100000;
|
||||
final int charVariety = 20;
|
||||
ObjectArrayList<Object[]> tests = new ObjectArrayList<Object[]>();
|
||||
ObjectOpenHashSet<String> truthSet = new ObjectOpenHashSet<String>();
|
||||
byte[] bytes = new byte[readNameLength];
|
||||
for ( int i = 0; i<nReads; i++) {
|
||||
random.nextBytes(bytes);
|
||||
StringBuilder readNameBuilder = new StringBuilder(readNameLength);
|
||||
for (byte b : bytes) {
|
||||
readNameBuilder.append((char) ('a' + Math.abs(b) % charVariety));
|
||||
}
|
||||
String readName = readNameBuilder.toString();
|
||||
tests.add(new Object[]{readName, truthSet.contains(readName)});
|
||||
truthSet.add(readName);
|
||||
}
|
||||
|
||||
return tests.toArray(new Object[][]{});
|
||||
}
|
||||
|
||||
/**
|
||||
* Test the read name compression functionality
|
||||
*/
|
||||
@Test(dataProvider = "ReadNameProvider")
|
||||
public void testReadNameCompression(final String name, final boolean alreadySeen) {
|
||||
GATKSAMRecord read = GATKSAMRecord.createRandomRead(1);
|
||||
read.setReadName(name);
|
||||
final int previousHashSize = hash.keySet().size();
|
||||
final long previousNumber = nextNumber;
|
||||
nextNumber = ReduceReads.compressReadName(hash, read, nextNumber);
|
||||
Assert.assertEquals(hash.keySet().size(), alreadySeen ? previousHashSize : previousHashSize + 1);
|
||||
Assert.assertEquals(nextNumber, alreadySeen ? previousNumber : previousNumber + 1);
|
||||
Assert.assertTrue(hash.containsKey(name));
|
||||
}
|
||||
|
||||
}
|
||||
|
|
@ -46,6 +46,9 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectSet;
|
||||
import net.sf.picard.reference.IndexedFastaSequenceFile;
|
||||
import net.sf.samtools.Cigar;
|
||||
import net.sf.samtools.CigarElement;
|
||||
|
|
@ -227,7 +230,7 @@ public class SlidingWindowUnitTest extends BaseTest {
|
|||
|
||||
private static final int readLength = 100;
|
||||
private static final int testRegionSize = 1000;
|
||||
private final List<GATKSAMRecord> basicReads = new ArrayList<GATKSAMRecord>(20);
|
||||
private final ObjectList<GATKSAMRecord> basicReads = new ObjectArrayList<GATKSAMRecord>(20);
|
||||
private IndexedFastaSequenceFile seq;
|
||||
private SAMFileHeader header;
|
||||
|
||||
|
|
@ -364,7 +367,7 @@ public class SlidingWindowUnitTest extends BaseTest {
|
|||
SlidingWindow slidingWindow = new SlidingWindow("1", 0, 10, header, new GATKSAMReadGroupRecord("test"), 0, 0.05, 0.05, 20, 20, 100, ReduceReads.DownsampleStrategy.Normal, false, false);
|
||||
for ( final GATKSAMRecord read : test.myReads )
|
||||
slidingWindow.addRead(read);
|
||||
Pair<Set<GATKSAMRecord>, CompressionStash> result = slidingWindow.close();
|
||||
Pair<ObjectSet<GATKSAMRecord>, CompressionStash> result = slidingWindow.close();
|
||||
|
||||
Assert.assertEquals(result.getFirst().size(), test.expectedNumberOfReads);
|
||||
|
||||
|
|
@ -403,7 +406,7 @@ public class SlidingWindowUnitTest extends BaseTest {
|
|||
@Test(dataProvider = "Downsampling", enabled = true)
|
||||
public void testDownsamplingTest(DSTest test) {
|
||||
final SlidingWindow slidingWindow = new SlidingWindow("1", 0, 10, header, new GATKSAMReadGroupRecord("test"), 0, 0.05, 0.05, 20, 20, test.dcov, ReduceReads.DownsampleStrategy.Normal, false, false);
|
||||
final List<GATKSAMRecord> result = slidingWindow.downsampleVariantRegion(basicReads);
|
||||
final ObjectList<GATKSAMRecord> result = slidingWindow.downsampleVariantRegion(basicReads);
|
||||
|
||||
Assert.assertEquals(result.size(), Math.min(test.dcov, basicReads.size()));
|
||||
}
|
||||
|
|
@ -453,7 +456,7 @@ public class SlidingWindowUnitTest extends BaseTest {
|
|||
final SlidingWindow slidingWindow = new SlidingWindow("1", 0, 10, header, new GATKSAMReadGroupRecord("test"), 0, 0.05, 0.05, minUsableConsensusQual, 20, 100, ReduceReads.DownsampleStrategy.Normal, false, false);
|
||||
for ( final GATKSAMRecord read : test.myReads )
|
||||
slidingWindow.addRead(read);
|
||||
final Pair<Set<GATKSAMRecord>, CompressionStash> result = slidingWindow.close();
|
||||
final Pair<ObjectSet<GATKSAMRecord>, CompressionStash> result = slidingWindow.close();
|
||||
|
||||
Assert.assertEquals(result.getFirst().size(), 1);
|
||||
final GATKSAMRecord read = result.getFirst().iterator().next();
|
||||
|
|
|
|||
|
|
@ -46,6 +46,8 @@
|
|||
|
||||
package org.broadinstitute.sting.gatk.walkers.compression.reducereads;
|
||||
|
||||
import it.unimi.dsi.fastutil.bytes.ByteArrayList;
|
||||
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
|
||||
import net.sf.samtools.SAMFileHeader;
|
||||
import org.broadinstitute.sting.BaseTest;
|
||||
import org.broadinstitute.sting.utils.sam.ArtificialSAMUtils;
|
||||
|
|
@ -54,9 +56,6 @@ import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
|
|||
import org.testng.Assert;
|
||||
import org.testng.annotations.Test;
|
||||
|
||||
import java.util.Arrays;
|
||||
import java.util.Random;
|
||||
|
||||
public class SyntheticReadUnitTest extends BaseTest {
|
||||
final SAMFileHeader artificialSAMHeader = ArtificialSAMUtils.createArtificialSamHeader(1, 1, 1);
|
||||
final GATKSAMReadGroupRecord artificialGATKRG = new GATKSAMReadGroupRecord("synthetic");
|
||||
|
|
@ -66,35 +65,32 @@ public class SyntheticReadUnitTest extends BaseTest {
|
|||
final int artificialRefStart = 1;
|
||||
final double artificialMappingQuality = 60;
|
||||
|
||||
final Random random = new Random(8854875);
|
||||
|
||||
|
||||
@Test
|
||||
public void testBaseCounts() {
|
||||
BaseIndex [] bases = new BaseIndex[] {BaseIndex.A,BaseIndex.A,BaseIndex.A,BaseIndex.A};
|
||||
Byte[] quals = new Byte[] {20, 20, 20, 20 };
|
||||
byte[] quals = new byte[] {20, 20, 20, 20 };
|
||||
|
||||
TestRead [] testReads = new TestRead [] {
|
||||
new TestRead(bases, quals, new Byte[] {100, 100, 100, 101}, new byte [] {100, 0, 0, 1}),
|
||||
new TestRead(bases, quals, new Byte[] {1, 100, 100, 0}, new byte [] {1, 99, 99, -1}),
|
||||
new TestRead(bases, quals, new Byte[] {127, 100, 0, 1}, new byte [] {127, -27, -127, -126}),
|
||||
new TestRead(bases, quals, new Byte[] {1, 127, 51, 126}, new byte [] {1, 126, 50, 125})};
|
||||
new TestRead(bases, quals, new byte[] {100, 100, 100, 101}, new byte [] {100, 0, 0, 1}),
|
||||
new TestRead(bases, quals, new byte[] {1, 100, 100, 0}, new byte [] {1, 99, 99, -1}),
|
||||
new TestRead(bases, quals, new byte[] {127, 100, 0, 1}, new byte [] {127, -27, -127, -126}),
|
||||
new TestRead(bases, quals, new byte[] {1, 127, 51, 126}, new byte [] {1, 126, 50, 125})};
|
||||
|
||||
for (TestRead testRead : testReads) {
|
||||
SyntheticRead syntheticRead = new SyntheticRead(Arrays.asList(testRead.getBases()), Arrays.asList(testRead.getCounts()), Arrays.asList(testRead.getQuals()), Arrays.asList(testRead.getInsQuals()), Arrays.asList(testRead.getDelQuals()), artificialMappingQuality, GATKSAMRecord.REDUCED_READ_CONSENSUS_TAG, artificialSAMHeader, artificialGATKRG, artificialContig, artificialContigIndex, artificialReadName, artificialRefStart, false, false);
|
||||
SyntheticRead syntheticRead = new SyntheticRead(new ObjectArrayList<BaseIndex>(testRead.getBases()), new ByteArrayList(testRead.getCounts()), new ByteArrayList(testRead.getQuals()), new ByteArrayList(testRead.getInsQuals()), new ByteArrayList(testRead.getDelQuals()), artificialMappingQuality, GATKSAMRecord.REDUCED_READ_CONSENSUS_TAG, artificialSAMHeader, artificialGATKRG, artificialContig, artificialContigIndex, artificialReadName, artificialRefStart, false, false);
|
||||
Assert.assertEquals(syntheticRead.convertBaseCounts(), testRead.getExpectedCounts());
|
||||
}
|
||||
}
|
||||
|
||||
private class TestRead {
|
||||
BaseIndex[] bases;
|
||||
Byte[] quals;
|
||||
Byte[] insQuals;
|
||||
Byte[] delQuals;
|
||||
Byte[] counts;
|
||||
byte [] expectedCounts;
|
||||
byte[] quals;
|
||||
byte[] insQuals;
|
||||
byte[] delQuals;
|
||||
byte[] counts;
|
||||
byte[] expectedCounts;
|
||||
|
||||
private TestRead(BaseIndex[] bases, Byte[] quals, Byte[] counts, byte[] expectedCounts) {
|
||||
private TestRead(BaseIndex[] bases, byte[] quals, byte[] counts, byte[] expectedCounts) {
|
||||
this.bases = bases;
|
||||
this.quals = quals;
|
||||
this.insQuals = quals;
|
||||
|
|
@ -107,19 +103,19 @@ private class TestRead {
|
|||
return bases;
|
||||
}
|
||||
|
||||
public Byte[] getQuals() {
|
||||
public byte[] getQuals() {
|
||||
return quals;
|
||||
}
|
||||
|
||||
public Byte[] getInsQuals() {
|
||||
public byte[] getInsQuals() {
|
||||
return insQuals;
|
||||
}
|
||||
|
||||
public Byte[] getDelQuals() {
|
||||
public byte[] getDelQuals() {
|
||||
return delQuals;
|
||||
}
|
||||
|
||||
public Byte[] getCounts() {
|
||||
public byte[] getCounts() {
|
||||
return counts;
|
||||
}
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue