org.broadinstitute.hellbender.engine

Class AssemblyRegion

java.lang.Object

org.broadinstitute.hellbender.engine.AssemblyRegion

All Implemented Interfaces:

htsjdk.samtools.util.Locatable

public final class AssemblyRegion
extends java.lang.Object
implements htsjdk.samtools.util.Locatable

Region of the genome that gets assembled by the local assembly engine. As AssemblyRegion is defined by two intervals -- a primary interval containing a territory for variant calling and a second, padded, interval for assembly -- as well as the reads overlapping the padded interval. Although we do not call variants in the padded interval, assembling over a larger territory improves calls in the primary territory. This concept is complicated somewhat by the fact that these intervals are mutable and the fact that the AssemblyRegion onject lives on after assembly during local realignment during PairHMM. Here is an example of the life cycle of an AssemblyRegion: Suppose that the HaplotypeCaller engine finds an evidence for a het in a pileup at locus 400 -- that is, it produces an ActivityProfileState with non-zero probability at site 400 and passes it to its ActivityProfile. The ActivityProfile eventually produces an AssemblyRegion based on the AssemblyRegionArgumentCollection parameters. Let's suppose that this initial region has primary span 350-450 and padded span 100 - 700. Next, the assembly engine assembles all reads that overlap the padded interval to find variant haplotypes and the variants they contain. The AssemblyRegion is then trimmed down to a new primary interval bound by all assembled variants within the original primary interval and a new padded interval. The amount of padding of the new padded interval around the variants depends on the needs of local realignment and as such need not equal the original padding that was used for assembly.

Constructor Summary

Constructors

Constructor and Description
AssemblyRegion(SimpleInterval activeSpan, boolean isActive, int padding, htsjdk.samtools.SAMFileHeader header) Create a new AssemblyRegion containing no reads
AssemblyRegion(SimpleInterval activeSpan, int padding, htsjdk.samtools.SAMFileHeader header) Simple interface to create an assembly region that isActive without any profile state
AssemblyRegion(SimpleInterval activeSpan, SimpleInterval paddedSpan, boolean isActive, htsjdk.samtools.SAMFileHeader header) Create a new AssemblyRegion containing no reads

Method Summary

Modifier and Type	Method and Description
void	add(GATKRead read) Add read to this region Read must have alignment start >= than the last read currently in this active region.
void	addAll(java.util.Collection<GATKRead> readsToAdd) Add all readsToAdd to this region
void	clearReads() Clear all of the reads currently in this region
byte[]	getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile referenceReader) See getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile) with padding == 0
byte[]	getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile referenceReader, int padding) Get the reference bases from referenceReader spanned by the padded span of this active region, including additional padding bp on either side.
java.lang.String	getContig()
int	getEnd()
htsjdk.samtools.SAMFileHeader	getHeader() Returns the header for the reads in this region.
SimpleInterval	getPaddedSpan() Get the span of this assembly region including the padding value
java.util.List<GATKRead>	getReads() Get an unmodifiable copy of the list of reads currently in this assembly region.
SimpleInterval	getSpan() Get the raw span of this assembly region (excluding the padding)
int	getStart()
boolean	isActive() Does this region represent an active region (all isActiveProbs above threshold) or an inactive region (all isActiveProbs below threshold)?
boolean	isFinalized()
void	removeAll(java.util.Collection<GATKRead> readsToRemove) Remove all of the reads in readsToRemove from this region
void	setFinalized(boolean value)
int	size() Get the number of reads currently in this region
java.lang.String	toString()
AssemblyRegion	trim(SimpleInterval span, int padding) Trim this region to just the span, producing a new assembly region without any reads that has only the extent of newExtend intersected with the current extent
AssemblyRegion	trim(SimpleInterval span, SimpleInterval paddedSpan) Trim this region to no more than the span, producing a new assembly region with properly trimmed reads that attempts to provide the best possible representation of this region covering the span.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface htsjdk.samtools.util.Locatable

contains, contigsMatch, getLengthOnReference, overlaps, withinDistanceOf

Constructor Detail

AssemblyRegion

public AssemblyRegion(SimpleInterval activeSpan,
                      boolean isActive,
                      int padding,
                      htsjdk.samtools.SAMFileHeader header)

Create a new AssemblyRegion containing no reads

Parameters:

activeSpan - the span of this active region

isActive - indicates whether this is an active region, or an inactive one

padding - the active region padding to use for this active region

AssemblyRegion

public AssemblyRegion(SimpleInterval activeSpan,
                      SimpleInterval paddedSpan,
                      boolean isActive,
                      htsjdk.samtools.SAMFileHeader header)

Create a new AssemblyRegion containing no reads

Parameters:

activeSpan - the span of this active region

paddedSpan - the padded span of this active region

isActive - indicates whether this is an active region, or an inactive one

AssemblyRegion

public AssemblyRegion(SimpleInterval activeSpan,
                      int padding,
                      htsjdk.samtools.SAMFileHeader header)

Simple interface to create an assembly region that isActive without any profile state

Method Detail

getContig

public java.lang.String getContig()

Specified by:

getContig in interface htsjdk.samtools.util.Locatable

getStart

public int getStart()

Specified by:

getStart in interface htsjdk.samtools.util.Locatable

getEnd

public int getEnd()

Specified by:

getEnd in interface htsjdk.samtools.util.Locatable

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

isActive

public boolean isActive()

Does this region represent an active region (all isActiveProbs above threshold) or an inactive region (all isActiveProbs below threshold)?

getPaddedSpan

public SimpleInterval getPaddedSpan()

Get the span of this assembly region including the padding value

Returns:

a non-null SimpleInterval

getSpan

public SimpleInterval getSpan()

Get the raw span of this assembly region (excluding the padding)

Returns:

a non-null SimpleInterval

getReads

public java.util.List<GATKRead> getReads()

Get an unmodifiable copy of the list of reads currently in this assembly region. The reads are sorted by their coordinate position.

Returns:

an unmodifiable and inmutable copy of the reads in the assembly region.

getHeader

public htsjdk.samtools.SAMFileHeader getHeader()

Returns the header for the reads in this region.

trim

public AssemblyRegion trim(SimpleInterval span,
                           int padding)

Trim this region to just the span, producing a new assembly region without any reads that has only the extent of newExtend intersected with the current extent

Parameters:

span - the new extend of the active region we want

padding - the padding size we want for the newly trimmed active region

Returns:

a non-null, empty assembly region

trim

public AssemblyRegion trim(SimpleInterval span,
                           SimpleInterval paddedSpan)

Trim this region to no more than the span, producing a new assembly region with properly trimmed reads that attempts to provide the best possible representation of this region covering the span. The challenge here is that span may (1) be larger than can be represented by this assembly region + its original padding and (2) the padding must be symmetric on both sides. This algorithm therefore determines how best to represent span as a subset of the span of this region with a padding value that captures as much of the span as possible. For example, suppose this active region is Active: 100-200 with padding of 50, so that the true span is 50-250 NewExtent: 150-225 saying that we'd ideally like to just have bases 150-225 Here we represent the assembly region as a region from 150-200 with 25 bp of padding. The overall constraint is that the region can never exceed the original region, and the padding is chosen to maximize overlap with the desired region

Parameters:

span - the new extend of the active region we want

Returns:

a non-null, empty active region

add

public void add(GATKRead read)

Add read to this region Read must have alignment start >= than the last read currently in this active region.

Parameters:

read - a non-null GATKRead

Throws:

java.lang.IllegalArgumentException - if read doesn't overlap the padded region of this active region

size

public int size()

Get the number of reads currently in this region

Returns:

an integer >= 0

clearReads

public void clearReads()

Clear all of the reads currently in this region

removeAll

public void removeAll(java.util.Collection<GATKRead> readsToRemove)

Remove all of the reads in readsToRemove from this region

Parameters:

readsToRemove - the set of reads we want to remove

addAll

public void addAll(java.util.Collection<GATKRead> readsToAdd)

Add all readsToAdd to this region

Parameters:

readsToAdd - a collection of readsToAdd to add to this active region

getAssemblyRegionReference

public byte[] getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile referenceReader)

See getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile) with padding == 0

getAssemblyRegionReference

public byte[] getAssemblyRegionReference(htsjdk.samtools.reference.ReferenceSequenceFile referenceReader,
                                         int padding)

Get the reference bases from referenceReader spanned by the padded span of this active region, including additional padding bp on either side. If this expanded region would exceed the boundaries of the active region's contig, the returned result will be truncated to only include on-genome reference bases

Parameters:

referenceReader - the source of the reference genome bases

padding - the padding, in BP, we want to add to either side of this active region padded region

Returns:

a non-null array of bytes holding the reference bases in referenceReader

setFinalized

public void setFinalized(boolean value)

isFinalized

public boolean isFinalized()