Routines for creating protein BLAST lookup tables. More...

#include <algo/blast/core/ncbi_std.h>
#include <algo/blast/core/blast_def.h>
#include <algo/blast/core/blast_lookup.h>
#include <algo/blast/core/blast_options.h>
#include <algo/blast/core/blast_rps.h>
#include <algo/blast/core/blast_stat.h>

Include dependency graph for blast_aalookup.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Go to the SVN repository for this file.

Classes
struct	AaLookupBackboneCell
	structure defining one cell of the compacted lookup table More...

struct	AaLookupSmallboneCell
	structure defining one cell of the small (i.e., use short) lookup table More...

struct	BlastAaLookupTable
	The basic lookup table structure for blastp searches. More...

struct	CompressedOverflowCell
	cell in list for holding query offsets More...

struct	CompressedMixedOffsets
	"alternative" structure of CompressedLookupBackboneCell storage More...

struct	CompressedLookupBackboneCell
	structure for hashtable of indexed query offsets More...

struct	BlastCompressedAaLookupTable
	The lookup table structure for protein searches using a compressed alphabet. More...

struct	RPSBackboneCell
	structure defining one cell of the RPS lookup table More...

struct	RPSBucket
	structure used for bucket sorting offsets retrieved from the RPS blast lookup table. More...

struct	BlastRPSLookupTable
	The basic lookup table structure for RPS blast searches. More...

Macros
#define	AA_HITS_PER_CELL 3
	maximum number of hits in one lookup table cell More...

#define	COMPRESSED_HITS_PER_BACKBONE_CELL 4
	number of query offsets to store in a backbone cell More...

#define	COMPRESSED_HITS_CELL_MASK 0x03

#define	COMPRESSED_HITS_PER_OVERFLOW_CELL 4
	number of query offsets to store in an overflow cell More...

#define	COMPRESSED_OVERFLOW_CELLS_IN_BANK 209710
	number of cells in one bank of cells More...

#define	COMPRESSED_OVERFLOW_MAX_BANKS 1024
	The maximum number of banks (usually less than 10 are needed; memory will run out before this is insufficient) More...

#define	RPS_HITS_PER_CELL 3
	maximum number of hits in an RPS backbone cell; this may be redundant (have the same value as AA_HITS_PER_CELL) but must be separate to guarantee binary compatibility with existing RPS blast databases More...

#define	RPS_BUCKET_SIZE 2048
	The number of regions into which the concatenated RPS blast database is split via bucket sorting. More...

Typedefs
typedef struct AaLookupBackboneCell	AaLookupBackboneCell
	structure defining one cell of the compacted lookup table More...

typedef struct AaLookupSmallboneCell	AaLookupSmallboneCell
	structure defining one cell of the small (i.e., use short) lookup table More...

typedef struct BlastAaLookupTable	BlastAaLookupTable
	The basic lookup table structure for blastp searches. More...

typedef struct CompressedOverflowCell	CompressedOverflowCell
	cell in list for holding query offsets More...

typedef struct CompressedMixedOffsets	CompressedMixedOffsets
	"alternative" structure of CompressedLookupBackboneCell storage More...

typedef struct CompressedLookupBackboneCell	CompressedLookupBackboneCell
	structure for hashtable of indexed query offsets More...

typedef struct BlastCompressedAaLookupTable	BlastCompressedAaLookupTable
	The lookup table structure for protein searches using a compressed alphabet. More...

typedef struct RPSBackboneCell	RPSBackboneCell
	structure defining one cell of the RPS lookup table More...

typedef struct RPSBucket	RPSBucket
	structure used for bucket sorting offsets retrieved from the RPS blast lookup table. More...

typedef struct BlastRPSLookupTable	BlastRPSLookupTable
	The basic lookup table structure for RPS blast searches. More...

Enumerations
enum	EBoneType { eBackbone = 0 , eSmallbone = 1 }
	types of cells More...

Functions
Int4	BlastAaLookupFinalize (BlastAaLookupTable *lookup, EBoneType bone_type)
	Pack the data structures comprising a protein lookup table into their final form. More...

Int4	BlastAaLookupTableNew (const LookupTableOptions opt, BlastAaLookupTable *lut)
	Create a new protein lookup table. More...

BlastAaLookupTable *	BlastAaLookupTableDestruct (BlastAaLookupTable *lookup)
	Free the lookup table. More...

void	BlastAaLookupIndexQuery (BlastAaLookupTable lookup, Int4 matrix, BLAST_SequenceBlk query, BlastSeqLoc *unmasked_regions, Int4 query_bias)
	Index a protein query. More...

Int4	BlastCompressedAaLookupTableNew (BLAST_SequenceBlk query, BlastSeqLoc locations, BlastCompressedAaLookupTable *lut, const LookupTableOptions opt, BlastScoreBlk *sbp)
	Create a new compressed protein lookup table. More...

BlastCompressedAaLookupTable *	BlastCompressedAaLookupTableDestruct (BlastCompressedAaLookupTable *lookup)
	Free the compressed lookup table. More...

static NCBI_INLINE Int4	s_ComputeCompressedIndex (Int4 wordsize, const Uint1 word, Int4 compressed_alphabet_size, Int4 skip, BlastCompressedAaLookupTable *lookup)
	Convert a word to use a compressed alphabet. More...

Int2	RPSLookupTableNew (const BlastRPSInfo rps_info, BlastRPSLookupTable *lut)
	Create a new RPS blast lookup table. More...

BlastRPSLookupTable *	RPSLookupTableDestruct (BlastRPSLookupTable *lookup)
	Free the lookup table. More...

Detailed Description

Routines for creating protein BLAST lookup tables.

Contains definitions and prototypes for the lookup table construction phase of blastp and RPS blast.

Definition in file blast_aalookup.h.

Macro Definition Documentation

◆ AA_HITS_PER_CELL

#define AA_HITS_PER_CELL 3

maximum number of hits in one lookup table cell

Definition at line 50 of file blast_aalookup.h.

◆ COMPRESSED_HITS_CELL_MASK

#define COMPRESSED_HITS_CELL_MASK 0x03

Definition at line 190 of file blast_aalookup.h.

◆ COMPRESSED_HITS_PER_BACKBONE_CELL

#define COMPRESSED_HITS_PER_BACKBONE_CELL 4

number of query offsets to store in a backbone cell

Definition at line 189 of file blast_aalookup.h.

◆ COMPRESSED_HITS_PER_OVERFLOW_CELL

#define COMPRESSED_HITS_PER_OVERFLOW_CELL 4

number of query offsets to store in an overflow cell

Definition at line 193 of file blast_aalookup.h.

◆ COMPRESSED_OVERFLOW_CELLS_IN_BANK

#define COMPRESSED_OVERFLOW_CELLS_IN_BANK 209710

number of cells in one bank of cells

Definition at line 196 of file blast_aalookup.h.

◆ COMPRESSED_OVERFLOW_MAX_BANKS

#define COMPRESSED_OVERFLOW_MAX_BANKS 1024

The maximum number of banks (usually less than 10 are needed; memory will run out before this is insufficient)

Definition at line 200 of file blast_aalookup.h.

◆ RPS_BUCKET_SIZE

#define RPS_BUCKET_SIZE 2048

The number of regions into which the concatenated RPS blast database is split via bucket sorting.

Definition at line 351 of file blast_aalookup.h.

◆ RPS_HITS_PER_CELL

#define RPS_HITS_PER_CELL 3

maximum number of hits in an RPS backbone cell; this may be redundant (have the same value as AA_HITS_PER_CELL) but must be separate to guarantee binary compatibility with existing RPS blast databases

Definition at line 334 of file blast_aalookup.h.

Typedef Documentation

◆ AaLookupBackboneCell

typedef struct AaLookupBackboneCell AaLookupBackboneCell

structure defining one cell of the compacted lookup table

◆ AaLookupSmallboneCell

typedef struct AaLookupSmallboneCell AaLookupSmallboneCell

structure defining one cell of the small (i.e., use short) lookup table

◆ BlastAaLookupTable

typedef struct BlastAaLookupTable BlastAaLookupTable

The basic lookup table structure for blastp searches.

◆ BlastCompressedAaLookupTable

typedef struct BlastCompressedAaLookupTable BlastCompressedAaLookupTable

The lookup table structure for protein searches using a compressed alphabet.

◆ BlastRPSLookupTable

typedef struct BlastRPSLookupTable BlastRPSLookupTable

The basic lookup table structure for RPS blast searches.

◆ CompressedLookupBackboneCell

typedef struct CompressedLookupBackboneCell CompressedLookupBackboneCell

structure for hashtable of indexed query offsets

◆ CompressedMixedOffsets

typedef struct CompressedMixedOffsets CompressedMixedOffsets

"alternative" structure of CompressedLookupBackboneCell storage

◆ CompressedOverflowCell

typedef struct CompressedOverflowCell CompressedOverflowCell

cell in list for holding query offsets

◆ RPSBackboneCell

typedef struct RPSBackboneCell RPSBackboneCell

structure defining one cell of the RPS lookup table

◆ RPSBucket

typedef struct RPSBucket RPSBucket

structure used for bucket sorting offsets retrieved from the RPS blast lookup table.

Enumeration Type Documentation

◆ EBoneType

enum EBoneType

types of cells

Enumerator
eBackbone
eSmallbone

Definition at line 88 of file blast_aalookup.h.

Function Documentation

◆ BlastAaLookupFinalize()

Int4 BlastAaLookupFinalize	(	BlastAaLookupTable *	lookup,
		EBoneType	bone_type
	)

Pack the data structures comprising a protein lookup table into their final form.

Parameters

lookup the lookup table [in]

Returns: Zero.

Definition at line 267 of file blast_aalookup.c.

References AA_HITS_PER_CELL, ASSERT, calloc(), eBackbone, AaLookupBackboneCell::entries, AaLookupSmallboneCell::entries, i, lookup(), NULL, AaLookupBackboneCell::num_used, AaLookupSmallboneCell::num_used, AaLookupBackboneCell::overflow_cursor, AaLookupSmallboneCell::overflow_cursor, AaLookupBackboneCell::payload, AaLookupSmallboneCell::payload, PV_ARRAY_BTS, PV_ARRAY_TYPE, PV_SET, and sfree.

Referenced by LookupTableWrapInit_MT(), and CMakeProfileDBApp::x_RPS_DbClose().

◆ BlastAaLookupIndexQuery()

void BlastAaLookupIndexQuery	(	BlastAaLookupTable *	lookup,
		Int4 **	matrix,
		BLAST_SequenceBlk *	query,
		BlastSeqLoc *	unmasked_regions,
		Int4	query_bias
	)

Index a protein query.

Parameters

lookup	the lookup table [in/modified]
matrix	the substitution matrix [in]
query	the array of queries to index
unmasked_regions	a BlastSeqLoc* which points to a (list of) integer pair(s) which specify the unmasked region(s) of the query [in]
query_bias	number added to each offset put into lookup table (only used for RPS blast database creation, otherwise 0) [in]

Definition at line 413 of file blast_aalookup.c.

References ASSERT, location, lookup(), NULL, query, s_AddNeighboringWords(), and s_AddPSSMNeighboringWords().

Referenced by LookupTableWrapInit_MT(), and CMakeProfileDBApp::x_RPSUpdateLookup().

◆ BlastAaLookupTableDestruct()

BlastAaLookupTable* BlastAaLookupTableDestruct ( BlastAaLookupTable * lookup )

Free the lookup table.

Parameters

lookup The lookup table structure to be freed

Returns: NULL

Definition at line 257 of file blast_aalookup.c.

References lookup(), NULL, and sfree.

Referenced by LookupTableWrapFree(), CMakeProfileDBApp::x_RPS_DbClose(), and CMakeProfileDBApp::CRPS_DbInfo::~CRPS_DbInfo().

◆ BlastAaLookupTableNew()

Int4 BlastAaLookupTableNew	(	const LookupTableOptions *	opt,
		BlastAaLookupTable **	lut
	)

Create a new protein lookup table.

Parameters

opt	pointer to lookup table options structure [in]
lut	handle to lookup table structure [in/modified]

Returns: 0 if successful, nonzero on failure

Definition at line 227 of file blast_aalookup.c.

References ASSERT, BLASTAA_SIZE, calloc(), i, ilog2(), lookup(), NULL, LookupTableOptions::threshold, and LookupTableOptions::word_size.

Referenced by LookupTableWrapInit_MT(), and CMakeProfileDBApp::x_RPSAddFirstSequence().

◆ BlastCompressedAaLookupTableDestruct()

BlastCompressedAaLookupTable* BlastCompressedAaLookupTableDestruct ( BlastCompressedAaLookupTable * lookup )

Free the compressed lookup table.

Parameters

lookup The lookup table structure to be freed

Returns: NULL

Definition at line 1359 of file blast_aalookup.c.

References free(), i, lookup(), NULL, and sfree.

Referenced by LookupTableWrapFree().

◆ BlastCompressedAaLookupTableNew()

Int4 BlastCompressedAaLookupTableNew	(	BLAST_SequenceBlk *	query,
		BlastSeqLoc *	locations,
		BlastCompressedAaLookupTable **	lut,
		const LookupTableOptions *	opt,
		BlastScoreBlk *	sbp
	)

Create a new compressed protein lookup table.

Parameters

query	The query sequence block (if concatenated sequence, the individual strands/sequences must be separated by a sentinel byte)[in]
locations	The locations to be included in the lookup table, e.g. [0,length-1] for full sequence. NULL means no sequence. [in]
lut	Pointer to the lookup table to be created [out]
opt	Options for lookup table creation [in]
sbp	pointer to score matrix information [in]

Returns: 0 if successful, nonzero on failure

Definition at line 1270 of file blast_aalookup.c.

References ASSERT, BLASTAA_SIZE, calloc(), SCompressedAlphabet::compress_table, COMPRESSED_OVERFLOW_CELLS_IN_BANK, COMPRESSED_OVERFLOW_MAX_BANKS, SBlastScoreMatrix::data, i, iexp(), letter(), lookup(), malloc(), SCompressedAlphabet::matrix, NULL, query, s_CompressedAddNeighboringWords(), s_CompressedLookupFinalize(), SCompressedAlphabetFree(), SCompressedAlphabetNew(), LookupTableOptions::threshold, and LookupTableOptions::word_size.

Referenced by LookupTableWrapInit_MT().

◆ RPSLookupTableDestruct()

BlastRPSLookupTable* RPSLookupTableDestruct ( BlastRPSLookupTable * lookup )

Free the lookup table.

Parameters

lookup The lookup table structure to free; note that the rps_backbone and rps_seq_offsets fields are not freed by this call, since they may refer to memory-mapped arrays

Returns: NULL

Definition at line 212 of file blast_aalookup.c.

References i, lookup(), NULL, and sfree.

Referenced by LookupTableWrapFree().

◆ RPSLookupTableNew()

Int2 RPSLookupTableNew	(	const BlastRPSInfo *	rps_info,
		BlastRPSLookupTable **	lut
	)

Create a new RPS blast lookup table.

Parameters

rps_info	pointer to structure with RPS setup information [in]
lut	handle to lookup table [in/modified]

Returns: 0 if successful, nonzero on failure

Definition at line 124 of file blast_aalookup.c.

References ASSERT, BLAST_WORDSIZE_PROT, calloc(), i, ilog2(), info, lookup(), BlastRPSLookupFileHeader::magic_number, BlastRPSProfileHeader::magic_number, malloc(), NULL, RPSBucket::num_alloc, RPSBucket::num_filled, BlastRPSProfileHeader::num_profiles, RPSBucket::offset_pairs, BlastRPSLookupFileHeader::overflow_hits, PV_ARRAY_BTS, PV_ARRAY_TYPE, PV_SET, RPS_BUCKET_SIZE, RPS_MAGIC_NUM, RPS_MAGIC_NUM_28, BlastRPSLookupFileHeader::start_of_backbone, and BlastRPSProfileHeader::start_offsets.

Referenced by LookupTableWrapInit_MT().

◆ s_ComputeCompressedIndex()

static NCBI_INLINE Int4 s_ComputeCompressedIndex	(	Int4	wordsize,
		const Uint1 *	word,
		Int4	compressed_alphabet_size,
		Int4 *	skip,
		BlastCompressedAaLookupTable *	lookup
	)

static

Convert a word to use a compressed alphabet.

The letters in the word are reversed compared to the original order

Parameters

wordsize	Number of consecutive letters in a word [in]
compressed_alphabet_size	Number of letters in compressed alphabet [in]
word	Sequence in "regular" AA alphabet [in]
skip	If a letter is encountered that cannot be compressed, the offset from word[] where index computation can begin again [out]
lookup	Translation tables etc [in]

Returns: Index calculated from scratch [out]

Definition at line 304 of file blast_aalookup.h.

References i, and lookup().

Referenced by s_BlastCompressedAaScanSubject(), and s_CompressedLookupAddUnencoded().

Classes

Macros

Typedefs

Enumerations

Functions

Detailed Description

Macro Definition Documentation

◆ AA_HITS_PER_CELL

◆ COMPRESSED_HITS_CELL_MASK

◆ COMPRESSED_HITS_PER_BACKBONE_CELL

◆ COMPRESSED_HITS_PER_OVERFLOW_CELL

◆ COMPRESSED_OVERFLOW_CELLS_IN_BANK

◆ COMPRESSED_OVERFLOW_MAX_BANKS

◆ RPS_BUCKET_SIZE

◆ RPS_HITS_PER_CELL

Typedef Documentation

◆ AaLookupBackboneCell

◆ AaLookupSmallboneCell

◆ BlastAaLookupTable

◆ BlastCompressedAaLookupTable

◆ BlastRPSLookupTable

◆ CompressedLookupBackboneCell

◆ CompressedMixedOffsets

◆ CompressedOverflowCell

◆ RPSBackboneCell

◆ RPSBucket

Enumeration Type Documentation

◆ EBoneType

Function Documentation

◆ BlastAaLookupFinalize()

◆ BlastAaLookupIndexQuery()

◆ BlastAaLookupTableDestruct()

◆ BlastAaLookupTableNew()

◆ BlastCompressedAaLookupTableDestruct()

◆ BlastCompressedAaLookupTableNew()

◆ RPSLookupTableDestruct()

◆ RPSLookupTableNew()

◆ s_ComputeCompressedIndex()