org.broadinstitute.hellbender.utils.pairhmm

Class PairHMM

java.lang.Object

org.broadinstitute.hellbender.utils.pairhmm.PairHMM

All Implemented Interfaces:

java.io.Closeable, java.lang.AutoCloseable

Direct Known Subclasses:

Log10PairHMM, LoglessPairHMM

public abstract class PairHMM
extends java.lang.Object
implements java.io.Closeable

Class for performing the pair HMM for local alignment. Figure 4.3 in Durbin 1998 book.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	PairHMM.Implementation

Field Summary

Fields

Modifier and Type	Field and Description
static byte	BASE_QUALITY_SCORE_THRESHOLD
protected boolean	constantsAreInitialized
protected boolean	doNotUseTristateCorrection
protected static java.lang.Boolean	doProfiling
protected int	hapStartIndex
protected boolean	initialized
protected static org.apache.logging.log4j.Logger	logger
protected int	maxHaplotypeLength
protected int	maxReadLength
protected double[]	mLogLikelihoodArray
protected int	paddedHaplotypeLength
protected int	paddedMaxHaplotypeLength
protected int	paddedMaxReadLength
protected int	paddedReadLength
protected static long	pairHMMComputeTime
protected byte[]	previousHaplotypeBases
protected long	startTime
protected long	threadLocalPairHMMComputeTimeDiff

Constructor Summary

Constructors

Constructor and Description
PairHMM()

Method Summary

Modifier and Type	Method and Description
void	close() Called at the end of the program to close files, print profiling information etc
void	computeLog10Likelihoods(LikelihoodMatrix<Haplotype> logLikelihoods, java.util.List<GATKRead> processedReads, java.util.Map<GATKRead,byte[]> gcp) Given a list of reads and haplotypes, for every read compute the total probability of said read arising from each haplotype given base substitution, insertion, and deletion probabilities.
protected void	doNotUseTristateCorrection()
static int	findFirstPositionWhereHaplotypesDiffer(byte[] haplotype1, byte[] haplotype2) Compute the first position at which two haplotypes differ If the haplotypes are exact copies of each other, returns the min length of the two haplotypes.
double[]	getLogLikelihoodArray() Return the results of the computeLogLikelihoods function
void	initialize(int readMaxLength, int haplotypeMaxLength) Initialize this PairHMM, making it suitable to run against a read and haplotype with given lengths Note: Do not worry about padding, just provide the true max length of the read and haplotype.
void	initialize(java.util.List<Haplotype> haplotypes, java.util.Map<java.lang.String,java.util.List<GATKRead>> perSampleReadList, int readMaxLength, int haplotypeMaxLength) Initialize this PairHMM, making it suitable to run against a read and haplotype with given lengths This function is used by the JNI implementations to transfer all data once to the native code
protected abstract double	subComputeReadLikelihoodGivenHaplotypeLog10(byte[] haplotypeBases, byte[] readBases, byte[] readQuals, byte[] insertionGOP, byte[] deletionGOP, byte[] overallGCP, int hapStartIndex, boolean recacheReadValues, int nextHapStartIndex) To be implemented by subclasses to do calculation for #computeReadLikelihoodGivenHaplotypeLog10

Methods inherited from class java.lang.Object

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

Field Detail

logger

protected static final org.apache.logging.log4j.Logger logger

BASE_QUALITY_SCORE_THRESHOLD

public static final byte BASE_QUALITY_SCORE_THRESHOLD

See Also:

Constant Field Values

constantsAreInitialized

protected boolean constantsAreInitialized

previousHaplotypeBases

protected byte[] previousHaplotypeBases

hapStartIndex

protected int hapStartIndex

maxHaplotypeLength

protected int maxHaplotypeLength

maxReadLength

protected int maxReadLength

paddedMaxReadLength

protected int paddedMaxReadLength

paddedMaxHaplotypeLength

protected int paddedMaxHaplotypeLength

paddedReadLength

protected int paddedReadLength

paddedHaplotypeLength

protected int paddedHaplotypeLength

initialized

protected boolean initialized

doNotUseTristateCorrection

protected boolean doNotUseTristateCorrection

mLogLikelihoodArray

protected double[] mLogLikelihoodArray

doProfiling

protected static java.lang.Boolean doProfiling

pairHMMComputeTime

protected static long pairHMMComputeTime

threadLocalPairHMMComputeTimeDiff

protected long threadLocalPairHMMComputeTimeDiff

startTime

protected long startTime

Constructor Detail

PairHMM

public PairHMM()

Method Detail

doNotUseTristateCorrection

protected void doNotUseTristateCorrection()

initialize

public void initialize(int readMaxLength,
                       int haplotypeMaxLength)
                throws java.lang.IllegalArgumentException

Initialize this PairHMM, making it suitable to run against a read and haplotype with given lengths Note: Do not worry about padding, just provide the true max length of the read and haplotype. The HMM will take care of the padding.

Parameters:

haplotypeMaxLength - the max length of haplotypes we want to use with this PairHMM

readMaxLength - the max length of reads we want to use with this PairHMM

Throws:

java.lang.IllegalArgumentException - if readMaxLength or haplotypeMaxLength is less than or equal to zero

initialize

public void initialize(java.util.List<Haplotype> haplotypes,
                       java.util.Map<java.lang.String,java.util.List<GATKRead>> perSampleReadList,
                       int readMaxLength,
                       int haplotypeMaxLength)

Initialize this PairHMM, making it suitable to run against a read and haplotype with given lengths This function is used by the JNI implementations to transfer all data once to the native code

Parameters:

haplotypes - the list of haplotypes

perSampleReadList - map from sample name to list of reads

haplotypeMaxLength - the max length of haplotypes we want to use with this PairHMM

readMaxLength - the max length of reads we want to use with this PairHMM

computeLog10Likelihoods

public void computeLog10Likelihoods(LikelihoodMatrix<Haplotype> logLikelihoods,
                                    java.util.List<GATKRead> processedReads,
                                    java.util.Map<GATKRead,byte[]> gcp)

Given a list of reads and haplotypes, for every read compute the total probability of said read arising from each haplotype given base substitution, insertion, and deletion probabilities.

Parameters:

processedReads - reads to analyze instead of the ones present in the destination read-likelihoods.

logLikelihoods - where to store the log likelihoods where position [a][r] is reserved for the log likelihood of reads[r] conditional to alleles[a].

gcp - penalty for gap continuations base array map for processed reads.

subComputeReadLikelihoodGivenHaplotypeLog10

protected abstract double subComputeReadLikelihoodGivenHaplotypeLog10(byte[] haplotypeBases,
                                                                      byte[] readBases,
                                                                      byte[] readQuals,
                                                                      byte[] insertionGOP,
                                                                      byte[] deletionGOP,
                                                                      byte[] overallGCP,
                                                                      int hapStartIndex,
                                                                      boolean recacheReadValues,
                                                                      int nextHapStartIndex)

To be implemented by subclasses to do calculation for #computeReadLikelihoodGivenHaplotypeLog10

findFirstPositionWhereHaplotypesDiffer

public static int findFirstPositionWhereHaplotypesDiffer(byte[] haplotype1,
                                                         byte[] haplotype2)
                                                  throws java.lang.IllegalArgumentException

Compute the first position at which two haplotypes differ If the haplotypes are exact copies of each other, returns the min length of the two haplotypes.

Parameters:

haplotype1 - the first haplotype1

haplotype2 - the second haplotype1

Returns:

the index of the first position in haplotype1 and haplotype2 where the byte isn't the same

Throws:

java.lang.IllegalArgumentException - if haplotype1 or haplotype2 are null or zero length

getLogLikelihoodArray

public double[] getLogLikelihoodArray()

Return the results of the computeLogLikelihoods function

close

public void close()

Called at the end of the program to close files, print profiling information etc

Specified by:

close in interface java.io.Closeable

Specified by:

close in interface java.lang.AutoCloseable