GenomeSequence Class Reference

Create/Access/Modify/Load Genome Sequences stored as binary mapped files. More...

#include <GenomeSequence.h>

Inheritance diagram for GenomeSequence:

[legend]

Collaboration diagram for GenomeSequence:

[legend]

List of all members.

Public Member Functions
	GenomeSequence ()
	Simple constructor - no implicit file open.
void	constructorClear ()
	GenomeSequence (std::string &referenceFilename)
	attempt to open an existing sequence
	GenomeSequence (const char *referenceFilename)
	Smarter constructor - attempt to open an existing sequence.
	~GenomeSequence ()
	Close the file if open and destroy the object.
bool	open (bool isColorSpace=false, int flags=O_RDONLY)
	open the reference specified using GenomeSequence::setReferenceName
bool	open (const char *filename, int flags=O_RDONLY)
	open the given file as the genome (no filename munging occurs).
bool	create ()
bool	create (bool isColor)
void	setProgressStream (std::ostream &progressStream)
void	setColorSpace (bool colorSpace)
void	setSearchCommonFileSuffix (bool searchCommonFileSuffix)
void	setCreateOverwrite (bool createOverwrite)
bool	loadFastaData (const char *filename)
bool	setReferenceName (std::string referenceFilename)
	set the reference name that will be used in open()
void	setApplication (std::string application)
	set the application name in the binary file header
const std::string &	getFastaName () const
const std::string &	getReferenceName () const
bool	isColorSpace () const
	tell us if we are a color space reference or not
genomeIndex_t	getNumberBases () const
	return the number of bases represented in this reference
int	getChromosome (genomeIndex_t position) const
	given a whole genome index, get the chromosome it is located in
int	getChromosome (const char *chromosomeName) const
	given a chromosome name, return the chromosome index
int	getChromosomeCount () const
	Return the number of chromosomes in the genome.
genomeIndex_t	getChromosomeStart (int chromosomeIndex) const
	given a chromosome, return the genome base it starts in
genomeIndex_t	getChromosomeSize (int chromosomeIndex) const
	given a chromosome, return its size in bases
genomeIndex_t	getGenomePosition (const char *chromosomeName, unsigned int chromosomeIndex) const
	given a chromosome name and position, return the genome position
genomeIndex_t	getGenomePosition (int chromosome, unsigned int chromosomeIndex) const
	given a chromosome index and position, return the genome position
genomeIndex_t	getGenomePosition (const char *chromosomeName) const
	given the chromosome name, get the corresponding 0 based genome index for the start of that chromosome
genomeIndex_t	getGenomePosition (int chromosomeIndex) const
const std::string &	getBaseFilename () const
const char *	getChromosomeName (int chromosomeIndex) const
void	setDebugFlag (bool d)
genomeIndex_t	sequenceLength () const
const char *	chromosomeName (int chr) const
void	sanityCheck (MemoryMap &fasta) const
std::string	IntegerToSeq (unsigned int n, unsigned int wordsize) const
bool	wordMatch (unsigned int index, std::string &word) const
bool	printNearbyWords (unsigned int index, unsigned int variance, std::string &word) const
char	BasePair (char c) const
void	dumpSequenceSAMDictionary (std::ostream &) const
void	dumpHeaderTSV (std::ostream &) const
char	operator[] (genomeIndex_t index) const
	Return the bases in base space or color space for within range index, ot.
uint8_t	getInteger (genomeIndex_t index) const
void	set (genomeIndex_t index, char value)
uint8_t *	getDataPtr (genomeIndex_t index)
	obtain the pointer to the raw data for other access methods
void	getReverseRead (std::string &read)
void	getReverseRead (String &read)
int	debugPrintReadValidation (std::string &read, std::string &quality, char direction, genomeIndex_t readLocation, int sumQuality, int mismatchCount, bool recurse=true)
void	getString (std::string &str, int chromosome, genomeIndex_t index, int baseCount) const
void	getString (String &str, int chromosome, genomeIndex_t index, int baseCount) const
void	getString (std::string &str, genomeIndex_t index, int baseCount) const
void	getString (String &str, genomeIndex_t index, int baseCount) const
void	getHighLightedString (std::string &str, genomeIndex_t index, int baseCount, genomeIndex_t highLightStart, genomeIndex_t highLightEnd) const
void	print30 (genomeIndex_t) const
genomeIndex_t	simpleLocalAligner (std::string &read, std::string &quality, genomeIndex_t index, int windowSize) const
int	getMismatchCount (std::string &read, genomeIndex_t location, char exclude='\0') const
	Return the mismatch count, disregarding CIGAR strings.
int	getSumQ (std::string &read, std::string &qualities, genomeIndex_t location) const
	brute force sumQ - no sanity checking
void	getMismatchHatString (std::string &result, const std::string &read, genomeIndex_t location) const
void	getMismatchString (std::string &result, const std::string &read, genomeIndex_t location) const
void	getChromosomeAndIndex (std::string &, genomeIndex_t) const
void	getChromosomeAndIndex (String &, genomeIndex_t) const
bool	checkRead (std::string &read, std::string &qualities, std::string &cigar, int &sumQ, int &gapOpenCount, int &gapExtendCount, int &gapDeleteCount, std::string &result) const
	check a SAM format read, using phred quality scores and the CIGAR string to determine if it is correct.
bool	populateDBSNP (mmapArrayBool_t &dbSNP, std::ifstream &inputFile) const
bool	loadDBSNP (mmapArrayBool_t &dbSNP, const char *inputFileName) const
	user friendly dbSNP loader.
Static Public Attributes
static const int	baseAIndex = 000
static const int	baseTIndex = 001
static const int	baseCIndex = 002
static const int	baseGIndex = 003
static const int	baseNIndex = 004
static const int	baseXIndex = 005
static unsigned char	base2int []
static const char	int2base [] = "ACGTNMXXXXXXXXXX"
static const char	int2colorSpace [] = "0123NXXXXXXXXXXX"
static unsigned char	base2complement []

---

Detailed Description

Create/Access/Modify/Load Genome Sequences stored as binary mapped files.

GenomeSequence is designed to be a high performance shared access reference object.

It is implemented as a MemoryMapArray template object with unsigned 8 bit ints, each of which stores two bases. Although 2 bits could be used, most references have more than four symbols (usually at least including 'N', indicating an unknown or masked out base).

Normal use of this class follows these steps:

1. create the reference
   1. instantiate the GenomeSequence class object
   2. create the actual file (memory mapped) that is to hold the data
   3. populate the data using GenomeSequence::set
2. use the reference
   1. use the reference by instantiating a GenomeSequence object
   2. either use the constructor with the reference filename
   3. or use GenomeSequence::setReferenceName() followed by open
   4. access the bases via the overloaded array operator []
   5. check sequence length by using GenomeSequence::getNumberBases()
3. accessing chromosomes in the reference
   1. you typically will need to know about the chromosomes in the sequence
   2. see methods and docs with prefix 'getChromosome'

Sharing is accomplished using the mmap() function via the MemoryMap base class. This allows a potentially large genome reference to be shared among a number of simultaneously executing instances of one or more programs sharing the same reference.

Definition at line 238 of file GenomeSequence.h.

---

Constructor & Destructor Documentation

GenomeSequence::GenomeSequence

(

std::string &

referenceFilename

)

[inline]

attempt to open an existing sequence

Parameters:

	referenceFilename	the name of the reference fasta file to open
	debug	if true, additional debug information is printed

Definition at line 287 of file GenomeSequence.h.

    {
        constructorClear();
        setup(referenceFilename.c_str());
    }

GenomeSequence::GenomeSequence

(

const char *

referenceFilename

)

[inline]

Smarter constructor - attempt to open an existing sequence.

Parameters:

	referenceFilename	the name of the reference fasta file to open
	debug	if true, additional debug information is printed

Definition at line 297 of file GenomeSequence.h.

    {
        constructorClear();
        setup(referenceFilename);
    }

---

Member Function Documentation

bool GenomeSequence::checkRead	(	std::string &	read,
		std::string &	qualities,
		std::string &	cigar,
		int &	sumQ,
		int &	gapOpenCount,
		int &	gapExtendCount,
		int &	gapDeleteCount,
		std::string &	result
	)			const

check a SAM format read, using phred quality scores and the CIGAR string to determine if it is correct.

Parameters:

	read	the read in base space
	qualities	the phred encoded qualities (Sanger, not Illumina)
	cigar	the SAM file CIGAR column
	sumQ	if >0 on entry, is checked against the computed sumQ
	insertions	count of insertions found in

int GenomeSequence::getChromosome

(

const char *

chromosomeName

)

const

given a chromosome name, return the chromosome index

This is done via a linear search of the chromosome table in the header of the mapped file, so it is O(N)

Parameters:

chromosomeName

the name of the chromosome - exact match only

Returns:

0-based index into chromosome table - INVALID_CHROMOSOME_INDEX if error

Definition at line 797 of file GenomeSequence.cpp.

{
    unsigned int i;
    for (i=0; i<header->_chromosomeCount; i++)
    {
        if (strcmp(header->_chromosomes[i].name, chromosomeName)==0)
        {
            return i;
        }
    }
    return INVALID_CHROMOSOME_INDEX;
}

int GenomeSequence::getChromosome

(

genomeIndex_t

position

)

const

given a whole genome index, get the chromosome it is located in

This is done via a binary search of the chromosome table in the header of the mapped file, so it is O(log(N))

Parameters:

0-based

position the base in the genome

Returns:

0-based index into chromosome table - INVALID_CHROMOSOME_INDEX if error

Definition at line 720 of file GenomeSequence.cpp.

Referenced by getGenomePosition().

{
    if (position == INVALID_GENOME_INDEX) return INVALID_CHROMOSOME_INDEX;

    if (header->_chromosomeCount == 0)
        return INVALID_CHROMOSOME_INDEX;

    int start = 0;
    int stop = header->_chromosomeCount - 1;

    // eliminate case where position is in the last chromosome, since the loop
    // below falls off the end of the list if it in the last one.

    if (position > header->_chromosomes[stop].start)
        return (stop);

    while (start <= stop)
    {
        int middle = (start + stop) / 2;

        if (position >= header->_chromosomes[middle].start && position < header->_chromosomes[middle + 1].start)
            return middle;

        if (position == header->_chromosomes[middle + 1].start)
            return (middle + 1);

        if (position > header->_chromosomes[middle + 1].start)
            start = middle + 1;

        if (position < header->_chromosomes[middle].start)
            stop = middle - 1;
    }

    return -1;
}

int GenomeSequence::getChromosomeCount

(

)

const

Return the number of chromosomes in the genome.

Returns:

number of chromosomes in the genome

Definition at line 714 of file GenomeSequence.cpp.

{
    return header->_chromosomeCount;
}

genomeIndex_t GenomeSequence::getChromosomeSize

(

int

chromosomeIndex

)

const [inline]

given a chromosome, return its size in bases

Parameters:

0-based

chromosome index

Returns:

size of the chromosome in bases

Definition at line 417 of file GenomeSequence.h.

    {
        if (chromosomeIndex==INVALID_CHROMOSOME_INDEX) return 0;
        return header->_chromosomes[chromosomeIndex].size;
    }

genomeIndex_t GenomeSequence::getChromosomeStart

(

int

chromosomeIndex

)

const [inline]

given a chromosome, return the genome base it starts in

Parameters:

0-based

chromosome index

Returns:

0-based genome index of the base that starts the chromosome

Definition at line 407 of file GenomeSequence.h.

    {
        if (chromosomeIndex==INVALID_CHROMOSOME_INDEX) return INVALID_GENOME_INDEX;
        return header->_chromosomes[chromosomeIndex].start;
    }

uint8_t* GenomeSequence::getDataPtr

(

genomeIndex_t

index

)

[inline]

obtain the pointer to the raw data for other access methods

this is a fairly ugly hack to reach into the raw genome vector, get the byte that encodes two bases, and return it. This is used by karma ReadIndexer::getSumQ to compare genome matchines by byte (two bases at a time) to speed it up.

Definition at line 563 of file GenomeSequence.h.

    {
        return ((uint8_t *) data + index/2);
    }

genomeIndex_t GenomeSequence::getGenomePosition	(	int	chromosome,
		unsigned int	chromosomeIndex
	)			const

given a chromosome index and position, return the genome position

Parameters:

	chromosome	index of the chromosome
	chromosomeIndex	1-based chromosome position

Returns:

genome index of the above chromosome position

Definition at line 771 of file GenomeSequence.cpp.

{
    if (chromosome<0 || chromosome >= (int) header->_chromosomeCount) return INVALID_GENOME_INDEX;

    genomeIndex_t i = header->_chromosomes[chromosome].start;
    if (i == INVALID_GENOME_INDEX) return INVALID_GENOME_INDEX;
    return i + chromosomeIndex - 1;
}

genomeIndex_t GenomeSequence::getGenomePosition	(	const char *	chromosomeName,
		unsigned int	chromosomeIndex
	)			const

given a chromosome name and position, return the genome position

Parameters:

	chromosomeName	name of the chromosome - exact match only
	chromosomeIndex	1-based chromosome position

Returns:

genome index of the above chromosome position

Definition at line 762 of file GenomeSequence.cpp.

Referenced by SamTags::createMDTag(), SamQuerySeqWithRefIter::reset(), SamQuerySeqWithRef::seqWithEquals(), and SamQuerySeqWithRef::seqWithoutEquals().

{
    genomeIndex_t i = getGenomePosition(chromosomeName);
    if (i == INVALID_GENOME_INDEX) return INVALID_GENOME_INDEX;
    return i + chromosomeIndex - 1;
}

int GenomeSequence::getMismatchCount	(	std::string &	read,
		genomeIndex_t	location,
		char	exclude = '\0'
	)			const [inline]

Return the mismatch count, disregarding CIGAR strings.

Parameters:

	read	is the sequence we're counting mismatches in
	location	is where in the genmoe we start comparing
	exclude	is a wildcard character (e.g. '.' or 'N')

Returns:

number of bases that don't match the reference, except those that match exclude

Definition at line 625 of file GenomeSequence.h.

    {
        int mismatchCount = 0;
        for (uint32_t i=0; i<read.size(); i++)
            if (read[i]!=exclude) mismatchCount += read[i]!=(*this)[location + i];
        return mismatchCount;
    };

genomeIndex_t GenomeSequence::getNumberBases

(

)

const [inline]

return the number of bases represented in this reference

Returns:

count of bases

Definition at line 377 of file GenomeSequence.h.

Referenced by loadDBSNP(), and operator[]().

    {
        return getElementCount();
    }

int GenomeSequence::getSumQ	(	std::string &	read,
		std::string &	qualities,
		genomeIndex_t	location
	)			const [inline]

brute force sumQ - no sanity checking

Parameters:

	read	shotgun sequencer read string
	qualities	phred quality string of same length
	location	the alignment location to check sumQ

Definition at line 638 of file GenomeSequence.h.

    {
        int sumQ = 0;
        for (uint32_t i=0; i<read.size(); i++)
            sumQ += (read[i]!=(*this)[location + i] ? (qualities[i]-33) : 0);
        return sumQ;
    };

bool GenomeSequence::isColorSpace

(

)

const [inline]

tell us if we are a color space reference or not

Returns:

true if colorspace, false otherwise

Definition at line 370 of file GenomeSequence.h.

Referenced by operator[]().

    {
        return _colorSpace;
    }

bool GenomeSequence::loadDBSNP	(	mmapArrayBool_t &	dbSNP,
		const char *	inputFileName
	)			const

user friendly dbSNP loader.

Parameters:

inputFileName

may be empty, point to a text file or a dbSNP vector file

In all cases, dbSNP is returned the same length as this genome.

When no SNPs are loaded, all values are false.

When a text file is given, the file is parsed with two space separated columns - the first column is the chromosome name, and the second is the 1-based chromosome position of the SNP.

Returns:

false if a dbSNP file was correctly loaded, true otherwise

Definition at line 1253 of file GenomeSequence.cpp.

References MemoryMapArray< elementT, indexT, cookieVal, versionVal, accessorFunc, setterFunc, elementCount2BytesFunc, arrayHeaderClass >::create(), getNumberBases(), and MemoryMapArray< elementT, indexT, cookieVal, versionVal, accessorFunc, setterFunc, elementCount2BytesFunc, arrayHeaderClass >::open().

{
    //
    // the goal in this section of code is to allow the user
    // to either specify a valid binary version of the SNP file,
    // or the original text file that it gets created from.
    //
    // To do this, we basically open, sniff the error message,
    // and if it claims it is not a binary version of the file,
    // we go ahead and treat it as the text file and use the
    // GenomeSequence::populateDBSNP method to load it.
    //
    // Further checking is really needed to ensure users don't
    // mix a dbSNP file for a different reference, since it is really
    // easy to do.
    //
    if (strlen(inputFileName)!=0)
    {
        std::cerr << "Load dbSNP file '" << inputFileName << "': " << std::flush;

        if (dbSNP.open(inputFileName, O_RDONLY))
        {
            //
            // failed to open, possibly due to bad magic.
            //
            // this is really awful ... need to have a return
            // code that is smart enough to avoid this ugliness:
            //
            if (dbSNP.getErrorString().find("wrong type of file")==std::string::npos)
            {
                std::cerr << "Error: " << dbSNP.getErrorString() << std::endl;
                exit(1);
            }
            //
            // we have a file, assume we can load it as a text file
            //
            std::ifstream inputFile;
            inputFile.open(inputFileName);
            if (inputFile.fail())
            {
                std::cerr << "Error: failed to open " << inputFileName << std::endl;
                exit(1);
            }

            std::cerr << "(as text file) ";

            // anonymously (RAM resident only) create:
            dbSNP.create(getNumberBases());

            // now load it into RAM
            populateDBSNP(dbSNP, inputFile);
            inputFile.close();

        }
        else
        {
            std::cerr << "(as binary mapped file) ";
        }

        std::cerr << "DONE!" << std::endl;
        return false;
    }
    else
    {
        return true;
    }
}

bool GenomeSequence::open	(	const char *	filename,
		int	flags = O_RDONLY
	)			[inline, virtual]

open the given file as the genome (no filename munging occurs).

Parameters:

	filename	the name of the file to open
	flags	pass through to the open() call (O_RDWR lets you modify the contents)

Returns:

false for success, true otherwise

Reimplemented from MemoryMapArray< elementT, indexT, cookieVal, versionVal, accessorFunc, setterFunc, elementCount2BytesFunc, arrayHeaderClass >.

Definition at line 318 of file GenomeSequence.h.

References open().

    {
        _umfaFilename = filename;
        return genomeSequenceArray::open(filename, flags);
    }

bool GenomeSequence::open	(	bool	isColorSpace = false,
		int	flags = O_RDONLY
	)

open the reference specified using GenomeSequence::setReferenceName

Parameters:

	isColorSpace	open the color space reference
	flags	pass through to the open() call (O_RDWR lets you modify the contents)

Returns:

false for success, true otherwise

Definition at line 250 of file GenomeSequence.cpp.

Referenced by open().

{
    bool rc;

    if (isColorSpace)
    {
        _umfaFilename = _baseFilename + "-cs.umfa";
    }
    else
    {
        _umfaFilename = _baseFilename + "-bs.umfa";
    }

    rc = genomeSequenceArray::open(_umfaFilename.c_str(), flags);
    if (rc)
    {
        std::cerr << "GenomeSequence::open: failed to open file "
                  << _umfaFilename
                  << std::endl;
        return true;
    }

    _colorSpace = header->_colorSpace;

    return false;
}

char GenomeSequence::operator[]

(

genomeIndex_t

index

)

const [inline]

Return the bases in base space or color space for within range index, ot.

Parameters:

index

the array-like index (0 based).

Returns:

ACTGN in base space; 0123N for color space; and 'N' for invalid. For color space, index i represents the transition of base at position (i-1) to base at position i

NB: bounds checking here needs to be deprecated - do not assume it will exist - the call must clip reads so that this routine is never called with a index value larger than the genome.

The reason for this is simply that this routine gets called hundreds of billions of time in one run of karma, which will absolutely kill performance. Every single instruction here matters a great, great deal.

Definition at line 522 of file GenomeSequence.h.

References getNumberBases(), and isColorSpace().

    {
        uint8_t val;
        if (index < getNumberBases())
        {
            if ((index&1)==0)
            {
                val = ((uint8_t *) data)[index>>1] & 0xf;
            }
            else
            {
                val = (((uint8_t *) data)[index>>1] & 0xf0) >> 4;
            }
        }
        else
        {
            val = baseNIndex;
        }
        val = isColorSpace() ? int2colorSpace[val] : int2base[val];
        return val;
    }

void GenomeSequence::setApplication

(

std::string

application

)

[inline]

set the application name in the binary file header

Parameters:

application

name of the application

Definition at line 355 of file GenomeSequence.h.

    {
        _application = application;     // used in ::create() to set application name
    }

bool GenomeSequence::setReferenceName

(

std::string

referenceFilename

)

set the reference name that will be used in open()

Parameters:

referenceFilename

the name of the reference fasta file to open

Returns:

false for success, true otherwise

See also:

open()

Definition at line 315 of file GenomeSequence.cpp.

{

    if (HAS_SUFFIX(referenceFilename, ".fa"))
    {
        _referenceFilename = referenceFilename;
        _baseFilename = _referenceFilename.substr(0, referenceFilename.size() - 3);
    }
    else if (HAS_SUFFIX(referenceFilename, ".umfa"))
    {
        _baseFilename = referenceFilename.substr(0, referenceFilename.size() - 5);
    }
    else if (HAS_SUFFIX(referenceFilename, "-cs.umfa"))
    {
        _baseFilename = referenceFilename.substr(0, referenceFilename.size() - 8);
    }
    else if (HAS_SUFFIX(referenceFilename, "-bs.umfa"))
    {
        _baseFilename = referenceFilename.substr(0, referenceFilename.size() - 8);
    }
    else
    {
        _baseFilename = referenceFilename;
    }
    _fastaFilename = _baseFilename + ".fa";

    return false;
}

---

Member Data Documentation

unsigned char GenomeSequence::base2complement [static]

Initial value:

    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "0123NNNNNNNNNNNN"  
    "NTNGNNNCNNNNNNNN"  
    "NNNNANNNNNNNNNNN"  
    "NTNGNNNCNNNNNNNN"  
    "NNNNANNNNNNNNNNN"  

    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"  
    "NNNNNNNNNNNNNNNN"

Definition at line 258 of file GenomeSequence.h.

unsigned char GenomeSequence::base2int [static]

Initial value:

    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\004\005"  
    "\000\001\002\003\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\000\005\001\005\005\005\002\005\005\005\005\005\005\004\005"  
    "\005\005\005\005\003\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\000\005\001\005\005\005\002\005\005\005\005\005\005\004\005"  
    "\005\005\005\005\003\005\005\005\005\005\005\005\005\005\005\005"  

    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"  
    "\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005\005"

Definition at line 253 of file GenomeSequence.h.

---

The documentation for this class was generated from the following files:

• general/GenomeSequence.h
• general/GenomeSequence.cpp