Private interface for Position Iterated BLAST API, contains the PSSM generation engine. More...

#include <algo/blast/core/ncbi_std.h>
#include <algo/blast/core/blast_stat.h>
#include <algo/blast/core/blast_psi.h>
#include "matrix_freq_ratios.h"

Include dependency graph for blast_psi_priv.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	_PSIPackedMsaCell
	Compact version of the PSIMsaCell structure. More...

struct	_PSIPackedMsa
	Compact version of PSIMsa structure. More...

struct	_PSIMsaCell
	Internal data structure to represent a position in the multiple sequence alignment data structure. More...

struct	_PSIMsa
	Internal multiple alignment data structure used by the PSSM engine. More...

struct	_PSIInternalPssmData
	Internal representation of a PSSM in various stages of its creation and its dimensions. More...

struct	_PSIAlignedBlock
	This structure keeps track of the regions aligned between the query sequence and those that were not purged. More...

struct	_PSISequenceWeights
	Internal data structure to keep computed sequence weights. More...

Macros
#define	PSI_SUCCESS (0)
	Successful operation. More...

#define	PSIERR_BADPARAM (-1)
	Bad parameter used in function. More...

#define	PSIERR_OUTOFMEM (-2)
	Out of memory. More...

#define	PSIERR_BADSEQWEIGHTS (-3)
	Sequence weights do not add to 1. More...

#define	PSIERR_NOFREQRATIOS (-4)
	No frequency ratios were found for the given scoring matrix. More...

#define	PSIERR_POSITIVEAVGSCORE (-5)
	Positive average score found when scaling matrix. More...

#define	PSIERR_NOALIGNEDSEQS (-6)
	After purge stage of PSSM creation, no sequences are left. More...

#define	PSIERR_GAPINQUERY (-7)
	GAP residue found in query sequence. More...

#define	PSIERR_UNALIGNEDCOLUMN (-8)
	Found an entire column with no participating sequences. More...

#define	PSIERR_COLUMNOFGAPS (-9)
	Found an entire column full of GAP residues. More...

#define	PSIERR_STARTINGGAP (-10)
	Found flanking gap at start of alignment. More...

#define	PSIERR_ENDINGGAP (-11)
	Found flanking gap at end of alignment. More...

#define	PSIERR_BADPROFILE (-12)
	Errors in conserved domain profile. More...

#define	PSIERR_UNKNOWN (-255)
	Unknown error. More...

Typedefs
typedef struct _PSIPackedMsaCell	_PSIPackedMsaCell
	Compact version of the PSIMsaCell structure. More...

typedef struct _PSIPackedMsa	_PSIPackedMsa
	Compact version of PSIMsa structure. More...

typedef struct _PSIMsaCell	_PSIMsaCell
	Internal data structure to represent a position in the multiple sequence alignment data structure. More...

typedef struct _PSIMsa	_PSIMsa
	Internal multiple alignment data structure used by the PSSM engine. More...

typedef struct _PSIInternalPssmData	_PSIInternalPssmData
	Internal representation of a PSSM in various stages of its creation and its dimensions. More...

typedef struct _PSIAlignedBlock	_PSIAlignedBlock
	This structure keeps track of the regions aligned between the query sequence and those that were not purged. More...

typedef struct _PSISequenceWeights	_PSISequenceWeights
	Internal data structure to keep computed sequence weights. More...

Functions
void **	_PSIAllocateMatrix (unsigned int ncols, unsigned int nrows, unsigned int data_type_sz)
	Generic 2 dimensional matrix allocator. More...

void **	_PSIDeallocateMatrix (void **matrix, unsigned int ncols)
	Generic 2 dimensional matrix deallocator. More...

void	_PSICopyMatrix_int (int dest, int src, unsigned int ncols, unsigned int nrows)
	Copies src matrix into dest matrix, both of which must be int matrices with dimensions ncols by nrows. More...

void	_PSICopyMatrix_double (double dest, double src, unsigned int ncols, unsigned int nrows)
	Copies src matrix into dest matrix, both of which must be double matrices with dimensions ncols by nrows. More...

_PSIPackedMsa *	_PSIPackedMsaNew (const PSIMsa *msa)
	Allocates and initializes the compact version of the PSIMsa structure (makes a deep copy) for internal use by the PSSM engine. More...

_PSIPackedMsa *	_PSIPackedMsaFree (_PSIPackedMsa *msa)
	Deallocates the _PSIMsa data structure. More...

unsigned int	_PSIPackedMsaGetNumberOfAlignedSeqs (const _PSIPackedMsa *msa)
	Retrieve the number of aligned sequences in the compact multiple sequence alignment. More...

_PSIMsa *	_PSIMsaNew (const _PSIPackedMsa *packed_msa, Uint4 alphabet_size)
	Allocates and initializes the internal version of the PSIMsa structure (makes a deep copy) for internal use by the PSSM engine. More...

_PSIMsa *	_PSIMsaFree (_PSIMsa *msa)
	Deallocates the _PSIMsa data structure. More...

_PSIInternalPssmData *	_PSIInternalPssmDataNew (Uint4 query_length, Uint4 alphabet_size)
	Allocates a new _PSIInternalPssmData structure. More...

_PSIInternalPssmData *	_PSIInternalPssmDataFree (_PSIInternalPssmData *pssm)
	Deallocates the _PSIInternalPssmData structure. More...

_PSIAlignedBlock *	_PSIAlignedBlockNew (Uint4 query_length)
	Allocates and initializes the _PSIAlignedBlock structure. More...

_PSIAlignedBlock *	_PSIAlignedBlockFree (_PSIAlignedBlock *aligned_blocks)
	Deallocates the _PSIAlignedBlock structure. More...

_PSISequenceWeights *	_PSISequenceWeightsNew (const PSIMsaDimensions dims, const BlastScoreBlk sbp)
	Allocates and initializes the _PSISequenceWeights structure. More...

_PSISequenceWeights *	_PSISequenceWeightsFree (_PSISequenceWeights *seq_weights)
	Deallocates the _PSISequenceWeights structure. More...

int	_PSIPurgeBiasedSegments (_PSIPackedMsa *msa)
	Main function for keeping only those selected sequences for PSSM construction (stage 2). More...

int	_PSIValidateMSA (const _PSIMsa *msa, Boolean ignored_unaligned_positions)
	Main validation function for multiple sequence alignment structure. More...

int	_PSIComputeAlignmentBlocks (const _PSIMsa msa, _PSIAlignedBlock aligned_block)
	Main function to compute aligned blocks' properties for each position within multiple alignment (stage 3) Corresponds to posit.c:posComputeExtents. More...

int	_PSIComputeSequenceWeights (const _PSIMsa msa, const _PSIAlignedBlock aligned_blocks, Boolean nsg_compatibility_mode, _PSISequenceWeights *seq_weights)
	Main function to calculate the sequence weights. More...

int	_PSIComputeFrequenciesFromCDs (const PSICdMsa cd_msa, BlastScoreBlk sbp, const PSIBlastOptions options, _PSISequenceWeights seq_weights)
	Main function to calculate CD weights and combine weighted residue counts from matched CDs. More...

int	_PSIComputeFreqRatios (const _PSIMsa msa, const _PSISequenceWeights seq_weights, const BlastScoreBlk sbp, const _PSIAlignedBlock aligned_blocks, Int4 pseudo_count, Boolean nsg_compatibility_mode, _PSIInternalPssmData *internal_pssm)
	Main function to compute the PSSM's frequency ratios (stage 5). More...

int	_PSIComputeFreqRatiosFromCDs (const PSICdMsa cd_msa, const _PSISequenceWeights seq_weights, const BlastScoreBlk sbp, Int4 pseudo_count, _PSIInternalPssmData internal_pssm)
	Main function to compute CD-based PSSM's frequency ratios. More...

int	_PSIConvertFreqRatiosToPSSM (_PSIInternalPssmData internal_pssm, const Uint1 query, const BlastScoreBlk sbp, const double std_probs)
	Converts the PSSM's frequency ratios obtained in the previous stage to a PSSM of scores. More...

int	_PSIScaleMatrix (const Uint1 query, const double std_probs, _PSIInternalPssmData internal_pssm, BlastScoreBlk sbp)
	Scales the PSSM (stage 7) More...

void	_PSIUpdateLambdaK (const int *pssm, const Uint1 query, Uint4 query_length, const double std_probs, BlastScoreBlk sbp)
	Updates the Karlin-Altschul parameters based on the query sequence and PSSM's score frequencies. More...

int	_IMPALAScaleMatrix (const Uint1 query, const double std_probs, _PSIInternalPssmData internal_pssm, BlastScoreBlk sbp, double scaling_factor)
	Provides a similar function to _PSIScaleMatrix but it performs the scaling as IMPALA did, i.e. More...

int	_PSIPurgeAlignedRegion (_PSIPackedMsa *msa, unsigned int seq_index, unsigned int start, unsigned int stop)
	Marks the (start, stop] region corresponding to sequence seq_index in alignment so that it is not further considered for PSSM calculation. More...

void	_PSIUpdatePositionCounts (_PSIMsa *msa)
	Counts the number of sequences matching the query per query position (columns of the multiple alignment) as well as the number of residues present in each position of the query. More...

Uint4	_PSISequenceLengthWithoutX (const Uint1 *seq, Uint4 length)
	Calculates the length of the sequence without including any 'X' residues. More...

Blast_ScoreFreq *	_PSIComputeScoreProbabilities (const int *pssm, const Uint1 query, Uint4 query_length, const double std_probs, const BlastScoreBlk sbp)
	Compute the probabilities for each score in the PSSM. More...

int	_PSISaveDiagnostics (const _PSIMsa msa, const _PSIAlignedBlock aligned_block, const _PSISequenceWeights seq_weights, const _PSIInternalPssmData internal_pssm, PSIDiagnosticsResponse *diagnostics)
	Collects diagnostic information from the process of creating the PSSM. More...

int	_PSISaveCDDiagnostics (const PSICdMsa msa, const _PSISequenceWeights seq_weights, const _PSIInternalPssmData internal_pssm, PSIDiagnosticsResponse diagnostics)
	Collects diagnostic information from the process of creating the CDD-based PSSM. More...

double *	_PSICalculateInformationContentFromScoreMatrix (Int4 *score_mat, const double std_prob, const Uint1 *query, Uint4 query_length, Uint4 alphabet_sz, double lambda)
	Calculates the information content from the scoring matrix. More...

double *	_PSICalculateInformationContentFromFreqRatios (double *freq_ratios, const double std_prob, Uint4 query_length, Uint4 alphabet_sz)
	Calculates the information content from the residue frequencies calculated in stage 5 of the PSSM creation algorithm Corresponds to posit.c:posFreqsToInformation. More...

void	_PSIStructureGroupCustomization (_PSIMsa *msa)
	Enable NCBI structure group customization to discard the query sequence, as this really isn't the result of a PSI-BLAST iteration, but rather an artificial consensus sequence of the multiple sequence alignment constructed by them. More...

int	_PSIValidateMSA_StructureGroup (const _PSIMsa *msa)
	Structure group validation function for multiple sequence alignment structure. More...

int	_PSIValidateCdMSA (const PSICdMsa *cd_msa, Uint4 alphabet_size)
	Validation of multiple alignment of conserved domains structure. More...

Variables
const double	kPSINearIdentical
	Percent identity threshold for discarding near-identical matches. More...

const double	kPSIIdentical
	Percent identity threshold for discarding identical matches. More...

const unsigned int	kQueryIndex
	Index into multiple sequence alignment structure for the query sequence. More...

const double	kEpsilon
	Small constant to test against 0. More...

const int	kPSIScaleFactor
	Successor to POSIT_SCALE_FACTOR. More...

const double	kPositScalingPercent
	Constant used in scaling PSSM routines: Successor to POSIT_PERCENT. More...

const Uint4	kPositScalingNumIterations
	Constant used in scaling PSSM routines: Successor to POSIT_NUM_ITERATIONS. More...

Detailed Description

Private interface for Position Iterated BLAST API, contains the PSSM generation engine.

Calculating PSSMs from Seq-aligns is a multi-stage process. These stages
include:
1) Processing the Seq-align
     Examine alignment and extract information about aligned characters,
     performed at the API level
2) Purge biased sequences: construct M multiple sequence alignment as
     described in page 3395[1] - performed at the core level; custom
     selection of sequences should be performed at the API level.
3) Compute extents of the alignment: M sub C as described in page 3395[1]
4) Compute sequence weights
5) Compute residue frequencies
6) Convert residue frequencies to PSSM
7) Scale the resulting PSSM

Definition in file blast_psi_priv.h.

Macro Definition Documentation

◆ PSI_SUCCESS

#define PSI_SUCCESS (0)

Successful operation.

Definition at line 366 of file blast_psi_priv.h.

◆ PSIERR_BADPARAM

#define PSIERR_BADPARAM (-1)

Bad parameter used in function.

Definition at line 368 of file blast_psi_priv.h.

◆ PSIERR_BADPROFILE

#define PSIERR_BADPROFILE (-12)

Errors in conserved domain profile.

Definition at line 390 of file blast_psi_priv.h.

◆ PSIERR_BADSEQWEIGHTS

#define PSIERR_BADSEQWEIGHTS (-3)

Sequence weights do not add to 1.

Definition at line 372 of file blast_psi_priv.h.

◆ PSIERR_COLUMNOFGAPS

#define PSIERR_COLUMNOFGAPS (-9)

Found an entire column full of GAP residues.

Definition at line 384 of file blast_psi_priv.h.

◆ PSIERR_ENDINGGAP

#define PSIERR_ENDINGGAP (-11)

Found flanking gap at end of alignment.

Definition at line 388 of file blast_psi_priv.h.

◆ PSIERR_GAPINQUERY

#define PSIERR_GAPINQUERY (-7)

GAP residue found in query sequence.

Definition at line 380 of file blast_psi_priv.h.

◆ PSIERR_NOALIGNEDSEQS

#define PSIERR_NOALIGNEDSEQS (-6)

After purge stage of PSSM creation, no sequences are left.

Definition at line 378 of file blast_psi_priv.h.

◆ PSIERR_NOFREQRATIOS

#define PSIERR_NOFREQRATIOS (-4)

No frequency ratios were found for the given scoring matrix.

Definition at line 374 of file blast_psi_priv.h.

◆ PSIERR_OUTOFMEM

#define PSIERR_OUTOFMEM (-2)

Out of memory.

Definition at line 370 of file blast_psi_priv.h.

◆ PSIERR_POSITIVEAVGSCORE

#define PSIERR_POSITIVEAVGSCORE (-5)

Positive average score found when scaling matrix.

Definition at line 376 of file blast_psi_priv.h.

◆ PSIERR_STARTINGGAP

#define PSIERR_STARTINGGAP (-10)

Found flanking gap at start of alignment.

Definition at line 386 of file blast_psi_priv.h.

◆ PSIERR_UNALIGNEDCOLUMN

#define PSIERR_UNALIGNEDCOLUMN (-8)

Found an entire column with no participating sequences.

Definition at line 382 of file blast_psi_priv.h.

◆ PSIERR_UNKNOWN

#define PSIERR_UNKNOWN (-255)

Unknown error.

Definition at line 392 of file blast_psi_priv.h.

Typedef Documentation

◆ _PSIAlignedBlock

typedef struct _PSIAlignedBlock _PSIAlignedBlock

This structure keeps track of the regions aligned between the query sequence and those that were not purged.

It is used when calculating the sequence weights (replaces posExtents in old code)

◆ _PSIInternalPssmData

typedef struct _PSIInternalPssmData _PSIInternalPssmData

Internal representation of a PSSM in various stages of its creation and its dimensions.

◆ _PSIMsa

typedef struct _PSIMsa _PSIMsa

Internal multiple alignment data structure used by the PSSM engine.

◆ _PSIMsaCell

typedef struct _PSIMsaCell _PSIMsaCell

Internal data structure to represent a position in the multiple sequence alignment data structure.

See also: _PSIMsa

◆ _PSIPackedMsa

typedef struct _PSIPackedMsa _PSIPackedMsa

Compact version of PSIMsa structure.

◆ _PSIPackedMsaCell

typedef struct _PSIPackedMsaCell _PSIPackedMsaCell

Compact version of the PSIMsaCell structure.

◆ _PSISequenceWeights

typedef struct _PSISequenceWeights _PSISequenceWeights

Internal data structure to keep computed sequence weights.

Function Documentation

◆ _IMPALAScaleMatrix()

int _IMPALAScaleMatrix	(	const Uint1 *	query,
		const double *	std_probs,
		_PSIInternalPssmData *	internal_pssm,
		BlastScoreBlk *	sbp,
		double	scaling_factor
	)

Provides a similar function to _PSIScaleMatrix but it performs the scaling as IMPALA did, i.e.

: allowing the specification of a scaling factor and when calculating the score probabilities, the query length includes 'X' residues.

Definition at line 2599 of file blast_psi_priv.c.

References _PSICopyMatrix_int(), _PSIInternalPssmData::freq_ratios, Kappa_compactSearchItemsFree(), Kappa_compactSearchItemsNew(), Kappa_impalaScaling(), Kappa_posSearchItemsFree(), Kappa_posSearchItemsNew(), BlastScoreBlk::name, _PSIInternalPssmData::ncols, _PSIInternalPssmData::nrows, NULL, PSI_SUCCESS, _PSIInternalPssmData::pssm, query, _PSIInternalPssmData::scaled_pssm, and TRUE.

Referenced by _PSICreateAndScalePssmFromFrequencyRatios().

◆ _PSIAlignedBlockFree()

_PSIAlignedBlock* _PSIAlignedBlockFree ( _PSIAlignedBlock * aligned_blocks )

Deallocates the _PSIAlignedBlock structure.

Parameters

aligned_blocks data structure to deallocate [in]

Returns: NULL

Definition at line 532 of file blast_psi_priv.c.

References NULL, _PSIAlignedBlock::pos_extnt, sfree, and _PSIAlignedBlock::size.

Referenced by _PSIAlignedBlockNew(), Deleter< _PSIAlignedBlock >::Delete(), and s_PSICreatePssmCleanUp().

◆ _PSIAlignedBlockNew()

_PSIAlignedBlock* _PSIAlignedBlockNew ( Uint4 query_length )

Allocates and initializes the _PSIAlignedBlock structure.

Parameters

query_length length of the query sequence of the multiple sequence alignment [in]

Returns: newly allocated structure or NULL in case of memory allocation failure

Definition at line 501 of file blast_psi_priv.c.

References _PSIAlignedBlockFree(), calloc(), i, SSeqRange::left, malloc(), NULL, _PSIAlignedBlock::pos_extnt, SSeqRange::right, and _PSIAlignedBlock::size.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIAllocateMatrix()

void** _PSIAllocateMatrix	(	unsigned int	ncols,
		unsigned int	nrows,
		unsigned int	data_type_sz
	)

Generic 2 dimensional matrix allocator.

Allocates a ncols by nrows matrix with cells of size data_type_sz. Must be freed using x_DeallocateMatrix

Parameters

ncols	number of columns in matrix [in]
nrows	number of rows in matrix [in]
data_type_sz	size of the data type (in bytes) to allocate for each element in the matrix [in]

Returns: pointer to allocated memory or NULL in case of failure

Definition at line 66 of file blast_psi_priv.c.

References _PSIDeallocateMatrix(), calloc(), i, malloc(), and NULL.

Referenced by _PSIInternalPssmDataNew(), _PSIMatrixFrequencyRatiosNew(), _PSIMsaNew(), _PSIPackedMsaNew(), _PSISequenceWeightsNew(), Kappa_posSearchItemsNew(), CRedoAlignmentTestFixture::loadPssmFromFile(), PSIDiagnosticsResponseNew(), PSIMatrixNew(), PSIMsaNew(), RPSRescalePssm(), s_RPSComputeTraceback(), s_RPSFillFreqRatiosInPsiMatrix(), SBlastScoreMatrixNew(), and SPsiBlastScoreMatrixNew().

◆ _PSICalculateInformationContentFromFreqRatios()

double* _PSICalculateInformationContentFromFreqRatios	(	double **	freq_ratios,
		const double *	std_prob,
		Uint4	query_length,
		Uint4	alphabet_sz
	)

Calculates the information content from the residue frequencies calculated in stage 5 of the PSSM creation algorithm Corresponds to posit.c:posFreqsToInformation.

See also: _PSIComputeFreqRatios: stage 5

Parameters

freq_ratios	matrix of frequency ratios (dimensions: query_length x alphabet_sz) (const) [in]
std_prob	standard residue probabilities [in]
query_length	length of the query [in]
alphabet_sz	length of the alphabet used by the query [in]

Returns: array of length query_length containing the information content per query position or NULL on error (e.g.: out-of-memory or NULL parameters)

Definition at line 2341 of file blast_psi_priv.c.

References calloc(), kEpsilon, log, NCBIMATH_LN2, NULL, and r().

Referenced by _PSISaveCDDiagnostics(), and _PSISaveDiagnostics().

◆ _PSICalculateInformationContentFromScoreMatrix()

double* _PSICalculateInformationContentFromScoreMatrix	(	Int4 **	score_mat,
		const double *	std_prob,
		const Uint1 *	query,
		Uint4	query_length,
		Uint4	alphabet_sz,
		double	lambda
	)

Calculates the information content from the scoring matrix.

Parameters

score_mat	alphabet by alphabet_sz matrix of scores (const) [in]
std_prob	standard residue probabilities [in]
query	query sequence [in]
query_length	length of the query [in]
alphabet_sz	length of the alphabet used by the query [in]
lambda	lambda parameter [in] FIXME documentation

Returns: array of length query_length containing the information content per query position or NULL on error (e.g.: out-of-memory or NULL parameters)

Definition at line 2299 of file blast_psi_priv.c.

References calloc(), kEpsilon, lambda(), log, NCBIMATH_LN2, NULL, query, r(), and tmp.

◆ _PSIComputeAlignmentBlocks()

int _PSIComputeAlignmentBlocks	(	const _PSIMsa *	msa,
		_PSIAlignedBlock *	aligned_block
	)

Main function to compute aligned blocks' properties for each position within multiple alignment (stage 3) Corresponds to posit.c:posComputeExtents.

Parameters

msa	multiple sequence alignment data structure [in]
aligned_block	data structure describing the aligned blocks' properties for each position of the multiple sequence alignment [out]

Returns: PSIERR_BADPARAM if arguments are NULL PSI_SUCCESS otherwise

Definition at line 1297 of file blast_psi_priv.c.

References _PSIComputeAlignedRegionLengths(), _PSIComputePositionExtents(), _PSIGetLeftExtents(), _PSIGetRightExtents(), _PSIMsa::dimensions, kQueryIndex, PSIMsaDimensions::num_seqs, PSI_SUCCESS, and PSIERR_BADPARAM.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIComputeFreqRatios()

int _PSIComputeFreqRatios	(	const _PSIMsa *	msa,
		const _PSISequenceWeights *	seq_weights,
		const BlastScoreBlk *	sbp,
		const _PSIAlignedBlock *	aligned_blocks,
		Int4	pseudo_count,
		Boolean	nsg_compatibility_mode,
		_PSIInternalPssmData *	internal_pssm
	)

Main function to compute the PSSM's frequency ratios (stage 5).

Implements formula 2 in Nucleic Acids Research, 2001, Vol 29, No 14. Corresponds to posit.c:posComputePseudoFreqs

Parameters

msa	multiple sequence alignment data structure [in]
seq_weights	data structure containing the data needed to compute the sequence weights [in]
sbp	score block structure initialized for the scoring system used with the query sequence [in]
aligned_blocks	data structure describing the aligned blocks' properties for each position of the multiple sequence alignment [in]
pseudo_count	pseudo count constant [in]
nsg_compatibility_mode	set to true to emulate the structure group's use of PSSM engine in the cddumper application. By default should be FALSE
internal_pssm	PSSM being computed [out]

Returns: PSIERR_BADPARAM if arguments are NULL, PSI_SUCCESS otherwise

Definition at line 2071 of file blast_psi_priv.c.

References _PSIMatrixFrequencyRatiosFree(), _PSIMatrixFrequencyRatiosNew(), BlastScoreBlk::alphabet_size, _PSIMsa::alphabet_size, AMINOACID_TO_NCBISTDAA, ASSERT, Blast_GetMatrixBackgroundFreq(), BLAST_SCORE_MIN, _PSIMsa::cell, SBlastScoreMatrix::data, SFreqRatios::data, _PSIMsa::dimensions, _PSIInternalPssmData::freq_ratios, i, _PSISequenceWeights::independent_observations, kEpsilon, kQueryIndex, _PSIMsaCell::letter, _PSISequenceWeights::match_weights, BlastScoreBlk::matrix, MAX_IND_OBSERVATIONS, BlastScoreBlk::name, NULL, PSEUDO_MAX, _PSIInternalPssmData::pseudocounts, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_UNKNOWN, PSIMsaDimensions::query_length, r(), s_columnSpecificPseudocounts(), s_effectiveObservations(), s_initializeExpNumObservations(), and _PSISequenceWeights::std_prob.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIComputeFreqRatiosFromCDs()

int _PSIComputeFreqRatiosFromCDs	(	const PSICdMsa *	cd_msa,
		const _PSISequenceWeights *	seq_weights,
		const BlastScoreBlk *	sbp,
		Int4	pseudo_count,
		_PSIInternalPssmData *	internal_pssm
	)

Main function to compute CD-based PSSM's frequency ratios.

Parameters

cd_msa	multiple alignment of CDs [in]
seq_weights	contains weighted residue frequencies and effective number of observations [in]
sbp	initialized score block data structure [in]
pseudo_count	pseudo count constant [in]
internal_pssm	PSSM [out]

Returns: status

Definition at line 2185 of file blast_psi_priv.c.

References _PSIMatrixFrequencyRatiosFree(), _PSIMatrixFrequencyRatiosNew(), BlastScoreBlk::alphabet_size, AMINOACID_TO_NCBISTDAA, ASSERT, Blast_GetMatrixBackgroundFreq(), BLAST_SCORE_MIN, SBlastScoreMatrix::data, SFreqRatios::data, PSICdMsa::dimensions, _PSIInternalPssmData::freq_ratios, i, _PSISequenceWeights::independent_observations, kEpsilon, _PSISequenceWeights::match_weights, BlastScoreBlk::matrix, MAX, BlastScoreBlk::name, NULL, PSEUDO_MAX, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_OUTOFMEM, PSICdMsa::query, PSIMsaDimensions::query_length, r(), s_columnSpecificPseudocounts(), and _PSISequenceWeights::std_prob.

Referenced by BOOST_AUTO_TEST_CASE(), PSICreatePssmFromCDD(), and s_TestCreatePssmFromFreqs().

◆ _PSIComputeFrequenciesFromCDs()

int _PSIComputeFrequenciesFromCDs	(	const PSICdMsa *	cd_msa,
		BlastScoreBlk *	sbp,
		const PSIBlastOptions *	options,
		_PSISequenceWeights *	seq_weights
	)

Main function to calculate CD weights and combine weighted residue counts from matched CDs.

Parameters

cd_msa	multiple alignment of conserved domains data structure [in]
sbp	BLAST score block [in]
options	CDD-related options [in]
seq_weights	data structure with CD frequencies [out]

Definition at line 1651 of file blast_psi_priv.c.

References BlastScoreBlk::alphabet_size, AMINOACID_TO_NCBISTDAA, ASSERT, PSICdMsaCell::data, PSICdMsa::dimensions, fabs, _PSISequenceWeights::independent_observations, PSICdMsaCellData::iobsr, PSICdMsaCell::is_aligned, malloc(), _PSISequenceWeights::match_weights, MIN, PSICdMsa::msa, NULL, PSIMsaDimensions::num_seqs, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_OUTOFMEM, PSICdMsa::query, PSIMsaDimensions::query_length, s_PSIComputeFrequenciesFromCDsCleanup(), and PSICdMsaCellData::wfreqs.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmFromCDD().

◆ _PSIComputeScoreProbabilities()

Blast_ScoreFreq* _PSIComputeScoreProbabilities	(	const int **	pssm,
		const Uint1 *	query,
		Uint4	query_length,
		const double *	std_probs,
		const BlastScoreBlk *	sbp
	)

Compute the probabilities for each score in the PSSM.

This is only valid for protein sequences. FIXME: Should this be moved to blast_stat.[hc]? used in kappa.c in notposfillSfp()

Parameters

pssm	PSSM for which to compute the score probabilities [in]
query	query sequence for the PSSM above in ncbistdaa encoding [in]
query_length	length of the query sequence above [in]
std_probs	array containing the standard background residue probabilities [in]
sbp	score block structure initialized for the scoring system used with the query sequence [in]

Returns: structure containing the score frequencies, or NULL in case of error

Definition at line 2647 of file blast_psi_priv.c.

References _PSISequenceLengthWithoutX(), BlastScoreBlk::alphabet_code, AMINOACID_TO_NCBISTDAA, ASSERT, Blast_GetStdAlphabet(), BLAST_SCORE_MAX, BLAST_SCORE_MIN, Blast_ScoreFreqNew(), BLASTAA_SEQ_CODE, BLASTAA_SIZE, kScore, MAX, MIN, NULL, Blast_ScoreFreq::obs_max, Blast_ScoreFreq::obs_min, query, r(), Blast_ScoreFreq::score_avg, and Blast_ScoreFreq::sprob.

Referenced by _PSIUpdateLambdaK().

◆ _PSIComputeSequenceWeights()

int _PSIComputeSequenceWeights	(	const _PSIMsa *	msa,
		const _PSIAlignedBlock *	aligned_blocks,
		Boolean	nsg_compatibility_mode,
		_PSISequenceWeights *	seq_weights
	)

Main function to calculate the sequence weights.

Should be called with the return value of PSIComputeAlignmentBlocks (stage 4) Corresponds to posit.c:posComputeSequenceWeights

Parameters

msa	multiple sequence alignment data structure [in]
aligned_blocks	data structure describing the aligned blocks' properties for each position of the multiple sequence alignment [in]
nsg_compatibility_mode	set to true to emulate the structure group's use of PSSM engine in the cddumper application. By default should be FALSE [in]
seq_weights	data structure containing the data needed to compute the sequence weights [out]

Returns: PSIERR_BADPARAM if arguments are NULL, PSIERR_OUTOFMEM in case of memory allocation failure, PSIERR_BADSEQWEIGHTS if the sequence weights fail to add up to 1.0, PSI_SUCCESS otherwise

Definition at line 1552 of file blast_psi_priv.c.

References _PSICalculateMatchWeights(), _PSICalculateNormalizedSequenceWeights(), _PSICheckSequenceWeights(), _PSIGetAlignedSequencesForPosition(), _PSISpreadGapWeights(), ASSERT, _PSIMsa::dimensions, DynamicUint4Array_AreEqual(), DynamicUint4Array_Copy(), DynamicUint4Array_Dup(), DynamicUint4ArrayFree(), DynamicUint4ArrayNewEx(), EFFECTIVE_ALPHABET, _PSISequenceWeights::norm_seq_weights, _PSIMsa::num_matching_seqs, PSIMsaDimensions::num_seqs, SDynamicUint4Array::num_used, _PSISequenceWeights::posDistinctDistrib, _PSISequenceWeights::posNumParticipating, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_OUTOFMEM, PSIMsaDimensions::query_length, _PSISequenceWeights::row_sigma, _PSISequenceWeights::sigma, and _PSIAlignedBlock::size.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIConvertFreqRatiosToPSSM()

int _PSIConvertFreqRatiosToPSSM	(	_PSIInternalPssmData *	internal_pssm,
		const Uint1 *	query,
		const BlastScoreBlk *	sbp,
		const double *	std_probs
	)

Converts the PSSM's frequency ratios obtained in the previous stage to a PSSM of scores.

(stage 6)

Parameters

internal_pssm	PSSM being computed [in\|out]
query	query sequence in ncbistdaa encoding. The length of this sequence is read from internal_pssm->ncols [in]
sbp	score block structure initialized for the scoring system used with the query sequence [in]
std_probs	array containing the standard residue probabilities [in]

Returns: PSIERR_BADPARAM if arguments are NULL, PSI_SUCCESS otherwise

Definition at line 2392 of file blast_psi_priv.c.

References _PSIMatrixFrequencyRatiosFree(), _PSIMatrixFrequencyRatiosNew(), BlastScoreBlk::alphabet_size, AMINOACID_TO_NCBISTDAA, SFreqRatios::bit_scale_factor, BLAST_Nint(), BLAST_SCORE_MIN, SBlastScoreMatrix::data, SFreqRatios::data, FALSE, _PSIInternalPssmData::freq_ratios, i, int, BlastScoreBlk::kbp_ideal, kEpsilon, kPSIScaleFactor, Blast_KarlinBlk::Lambda, log, BlastScoreBlk::matrix, BlastScoreBlk::name, NCBIMATH_LN2, _PSIInternalPssmData::ncols, NULL, PSI_SUCCESS, PSIERR_BADPARAM, _PSIInternalPssmData::pssm, query, _PSIInternalPssmData::scaled_pssm, tmp, and TRUE.

Referenced by _PSICreateAndScalePssmFromFrequencyRatios(), BOOST_AUTO_TEST_CASE(), s_ScalePosMatrix(), and s_TestCreatePssmFromFreqs().

◆ _PSICopyMatrix_double()

void _PSICopyMatrix_double	(	double **	dest,
		double **	src,
		unsigned int	ncols,
		unsigned int	nrows
	)

Copies src matrix into dest matrix, both of which must be double matrices with dimensions ncols by nrows.

Parameters

dest	Destination matrix [out]
src	Source matrix [in]
ncols	Number of columns to copy [in]
nrows	Number of rows to copy [in]

Definition at line 124 of file blast_psi_priv.c.

Referenced by PSICreatePssmFromFrequencyRatios(), and s_ScalePosMatrix().

◆ _PSICopyMatrix_int()

void _PSICopyMatrix_int	(	int **	dest,
		int **	src,
		unsigned int	ncols,
		unsigned int	nrows
	)

Copies src matrix into dest matrix, both of which must be int matrices with dimensions ncols by nrows.

Parameters

dest	Destination matrix [out]
src	Source matrix [in]
ncols	Number of columns to copy [in]
nrows	Number of rows to copy [in]

Definition at line 123 of file blast_psi_priv.c.

Referenced by _IMPALAScaleMatrix(), s_PSISavePssm(), s_ScalePosMatrix(), and CRedoAlignmentTestFixture::setupPositionBasedBlastScoreBlk().

◆ _PSIDeallocateMatrix()

void** _PSIDeallocateMatrix	(	void **	matrix,
		unsigned int	ncols
	)

Generic 2 dimensional matrix deallocator.

Deallocates the memory allocated by x_AllocateMatrix

Parameters

matrix	matrix to deallocate [in]
ncols	number of columns in the matrix [in]

Returns: NULL

Definition at line 88 of file blast_psi_priv.c.

References i, NULL, and sfree.

Referenced by _PSIAllocateMatrix(), _PSIInternalPssmDataFree(), _PSIMatrixFrequencyRatiosFree(), _PSIMsaFree(), _PSIPackedMsaFree(), _PSISequenceWeightsFree(), Kappa_posSearchItemsFree(), CRedoAlignmentTestFixture::loadPssmFromFile(), PSIDiagnosticsResponseFree(), PSIMatrixFree(), PSIMsaFree(), s_RPSComputeTraceback(), SBlastScoreMatrixFree(), CRedoAlignmentTestFixture::setupPositionBasedBlastScoreBlk(), and SPsiBlastScoreMatrixFree().

◆ _PSIInternalPssmDataFree()

_PSIInternalPssmData* _PSIInternalPssmDataFree ( _PSIInternalPssmData * pssm )

Deallocates the _PSIInternalPssmData structure.

Parameters

pssm	data structure to deallocate [in]

Returns: NULL

Definition at line 468 of file blast_psi_priv.c.

References _PSIDeallocateMatrix(), _PSIInternalPssmData::freq_ratios, _PSIInternalPssmData::ncols, NULL, _PSIInternalPssmData::pseudocounts, _PSIInternalPssmData::pssm, _PSIInternalPssmData::scaled_pssm, and sfree.

Referenced by _PSIInternalPssmDataNew(), Deleter< _PSIInternalPssmData >::Delete(), s_PSICreatePssmCleanUp(), s_PSICreatePssmFromFrequencyRatiosCleanUp(), and s_ScalePosMatrix().

◆ _PSIInternalPssmDataNew()

_PSIInternalPssmData* _PSIInternalPssmDataNew	(	Uint4	query_length,
		Uint4	alphabet_size
	)

Allocates a new _PSIInternalPssmData structure.

Parameters

query_length	number of columns for the PSSM [in]
alphabet_size	number of rows for the PSSM [in]

Returns: newly allocated structure or NULL in case of memory allocation failure

Definition at line 425 of file blast_psi_priv.c.

References _PSIAllocateMatrix(), _PSIInternalPssmDataFree(), calloc(), _PSIInternalPssmData::freq_ratios, _PSIInternalPssmData::ncols, _PSIInternalPssmData::nrows, NULL, _PSIInternalPssmData::pseudocounts, _PSIInternalPssmData::pssm, and _PSIInternalPssmData::scaled_pssm.

Referenced by BOOST_AUTO_TEST_CASE(), PSICreatePssmFromCDD(), PSICreatePssmFromFrequencyRatios(), PSICreatePssmWithDiagnostics(), s_ScalePosMatrix(), and s_TestCreatePssmFromFreqs().

◆ _PSIMsaFree()

_PSIMsa* _PSIMsaFree ( _PSIMsa * msa )

Deallocates the _PSIMsa data structure.

Parameters

msa	multiple sequence alignment data structure to deallocate [in]

Returns: NULL

Definition at line 389 of file blast_psi_priv.c.

References _PSIDeallocateMatrix(), _PSIMsa::cell, _PSIMsa::dimensions, NULL, _PSIMsa::num_matching_seqs, PSIMsaDimensions::num_seqs, _PSIMsa::query, PSIMsaDimensions::query_length, _PSIMsa::residue_counts, and sfree.

Referenced by _PSIMsaNew(), Deleter< _PSIMsa >::Delete(), and s_PSICreatePssmCleanUp().

◆ _PSIMsaNew()

_PSIMsa* _PSIMsaNew	(	const _PSIPackedMsa *	packed_msa,
		Uint4	alphabet_size
	)

Allocates and initializes the internal version of the PSIMsa structure (makes a deep copy) for internal use by the PSSM engine.

Parameters

packed_msa	compact multiple sequence alignment data structure [in]
alphabet_size	number of elements in the alphabet that makes up the aligned characters in the multiple sequence alignment [in]

Returns: newly allocated structure or NULL in case of memory allocation failure

Definition at line 308 of file blast_psi_priv.c.

References _PSIAllocateMatrix(), _PSIMsaFree(), _PSIPackedMsaGetNumberOfAlignedSeqs(), _PSIUpdatePositionCounts(), _PSIMsa::alphabet_size, ASSERT, calloc(), _PSIMsa::cell, _PSIPackedMsa::data, _PSIPackedMsa::dimensions, _PSIMsa::dimensions, _PSIMsaCell::extents, _PSIPackedMsaCell::is_aligned, _PSIMsaCell::is_aligned, IS_residue, kQueryIndex, SSeqRange::left, _PSIPackedMsaCell::letter, _PSIMsaCell::letter, malloc(), NULL, _PSIMsa::num_matching_seqs, PSIMsaDimensions::num_seqs, _PSIMsa::query, PSIMsaDimensions::query_length, _PSIMsa::residue_counts, SSeqRange::right, and _PSIPackedMsa::use_sequence.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIPackedMsaFree()

_PSIPackedMsa* _PSIPackedMsaFree ( _PSIPackedMsa * msa )

Deallocates the _PSIMsa data structure.

Parameters

msa	multiple sequence alignment data structure to deallocate [in]

Returns: NULL

Definition at line 183 of file blast_psi_priv.c.

References _PSIDeallocateMatrix(), _PSIPackedMsa::data, _PSIPackedMsa::dimensions, NULL, PSIMsaDimensions::num_seqs, sfree, and _PSIPackedMsa::use_sequence.

Referenced by _PSIPackedMsaNew(), Deleter< _PSIPackedMsa >::Delete(), PSICreatePssmWithDiagnostics(), and s_PSICreatePssmCleanUp().

◆ _PSIPackedMsaGetNumberOfAlignedSeqs()

unsigned int _PSIPackedMsaGetNumberOfAlignedSeqs ( const _PSIPackedMsa * msa )

Retrieve the number of aligned sequences in the compact multiple sequence alignment.

Parameters

msa	multiple sequence alignment data structure to deallocate [in]

Definition at line 209 of file blast_psi_priv.c.

References _PSIPackedMsa::dimensions, i, PSIMsaDimensions::num_seqs, and _PSIPackedMsa::use_sequence.

Referenced by _PSIMsaNew().

◆ _PSIPackedMsaNew()

_PSIPackedMsa* _PSIPackedMsaNew ( const PSIMsa * msa )

Allocates and initializes the compact version of the PSIMsa structure (makes a deep copy) for internal use by the PSSM engine.

Parameters

msa	multiple sequence alignment data structure provided by the user [in]

Returns: newly allocated structure or NULL in case of memory allocation failure

Definition at line 129 of file blast_psi_priv.c.

References _PSIAllocateMatrix(), _PSIPackedMsaFree(), ASSERT, BLASTAA_SIZE, Boolean, calloc(), _PSIPackedMsa::data, _PSIPackedMsa::dimensions, _PSIPackedMsaCell::is_aligned, _PSIPackedMsaCell::letter, malloc(), NULL, TRUE, and _PSIPackedMsa::use_sequence.

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSIPurgeAlignedRegion()

int _PSIPurgeAlignedRegion	(	_PSIPackedMsa *	msa,
		unsigned int	seq_index,
		unsigned int	start,
		unsigned int	stop
	)

Marks the (start, stop] region corresponding to sequence seq_index in alignment so that it is not further considered for PSSM calculation.

Note that the query sequence cannot be purged.

Parameters

msa	multiple sequence alignment data [in\|out]
seq_index	index of the sequence of interested in alignment [in]
start	start of the region to remove [in]
stop	stop of the region to remove [in]

Returns: PSIERR_BADPARAM if no alignment is given, or if seq_index or stop are invalid, PSI_SUCCESS otherwise

Definition at line 2781 of file blast_psi_priv.c.

References _PSIPackedMsa::data, _PSIPackedMsa::dimensions, FALSE, i, _PSIPackedMsaCell::is_aligned, _PSIPackedMsaCell::letter, NULL, PSIMsaDimensions::num_seqs, PSI_SUCCESS, PSIERR_BADPARAM, PSIMsaDimensions::query_length, and s_PSIDiscardIfUnused().

Referenced by _handleNeitherAligned().

◆ _PSIPurgeBiasedSegments()

int _PSIPurgeBiasedSegments ( _PSIPackedMsa * msa )

Main function for keeping only those selected sequences for PSSM construction (stage 2).

After this function the multiple sequence alignment data will not be modified.

See also: implementation of PSICreatePssmWithDiagnostics

Parameters

msa	multiple sequence alignment data structure [in]

Returns: PSIERR_BADPARAM if alignment is NULL; PSI_SUCCESS otherwise

Definition at line 948 of file blast_psi_priv.c.

References PSI_SUCCESS, PSIERR_BADPARAM, s_PSIPurgeNearIdenticalAlignments(), and s_PSIPurgeSelfHits().

Referenced by BOOST_AUTO_TEST_CASE(), and PSICreatePssmWithDiagnostics().

◆ _PSISaveCDDiagnostics()

int _PSISaveCDDiagnostics	(	const PSICdMsa *	msa,
		const _PSISequenceWeights *	seq_weights,
		const _PSIInternalPssmData *	internal_pssm,
		PSIDiagnosticsResponse *	diagnostics
	)

Collects diagnostic information from the process of creating the CDD-based PSSM.

Parameters

cd_msa	multiple alignment of CDs data structure [in]
seq_weights	sequence weights data structure [in]
internal_pssm	structure containing PSSM's frequency ratios [in]
diagnostics	output parameter [out]

Returns: PSI_SUCCESS on success, PSIERR_OUTOFMEM if memory allocation fails or PSIERR_BADPARAM if any of its arguments is NULL

Definition at line 2912 of file blast_psi_priv.c.

References _PSICalculateInformationContentFromFreqRatios(), PSIDiagnosticsResponse::alphabet_size, ASSERT, PSICdMsa::dimensions, _PSIInternalPssmData::freq_ratios, PSIDiagnosticsResponse::frequency_ratios, PSIDiagnosticsResponse::independent_observations, _PSISequenceWeights::independent_observations, info, PSIDiagnosticsResponse::information_content, _PSISequenceWeights::match_weights, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_OUTOFMEM, PSIMsaDimensions::query_length, PSIDiagnosticsResponse::query_length, r(), sfree, _PSISequenceWeights::std_prob, and PSIDiagnosticsResponse::weighted_residue_freqs.

Referenced by PSICreatePssmFromCDD().

◆ _PSISaveDiagnostics()

int _PSISaveDiagnostics	(	const _PSIMsa *	msa,
		const _PSIAlignedBlock *	aligned_block,
		const _PSISequenceWeights *	seq_weights,
		const _PSIInternalPssmData *	internal_pssm,
		PSIDiagnosticsResponse *	diagnostics
	)

Collects diagnostic information from the process of creating the PSSM.

Parameters

msa	multiple sequence alignment data structure [in]
aligned_block	aligned regions' extents [in]
seq_weights	sequence weights data structure [in]
internal_pssm	structure containing PSSM's frequency ratios [in]
diagnostics	output parameter [out]

Returns: PSI_SUCCESS on success, PSIERR_OUTOFMEM if memory allocation fails or PSIERR_BADPARAM if any of its arguments is NULL

Definition at line 2809 of file blast_psi_priv.c.

Referenced by PSICreatePssmWithDiagnostics().

◆ _PSIScaleMatrix()

int _PSIScaleMatrix	(	const Uint1 *	query,
		const double *	std_probs,
		_PSIInternalPssmData *	internal_pssm,
		BlastScoreBlk *	sbp
	)

Scales the PSSM (stage 7)

Parameters

query	query sequence in ncbistdaa encoding. The length of this sequence is read from internal_pssm->ncols [in]
std_probs	array containing the standard background residue probabilities [in]
internal_pssm	PSSM being computed [in\|out]
sbp	score block structure initialized for the scoring system used with the query sequence [in\|out]

Returns: PSIERR_BADPARAM if arguments are NULL, PSIERR_POSITIVEAVGSCORE if the average score of the generated PSSM is positive, PSI_SUCCESS otherwise

Definition at line 2481 of file blast_psi_priv.c.

References _PSIUpdateLambdaK(), ASSERT, BLAST_Nint(), BLAST_SCORE_MIN, FALSE, i, int, BlastScoreBlk::kbp_ideal, BlastScoreBlk::kbp_psi, kPositScalingNumIterations, kPositScalingPercent, kPSIScaleFactor, Blast_KarlinBlk::Lambda, _PSIInternalPssmData::ncols, _PSIInternalPssmData::nrows, NULL, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_POSITIVEAVGSCORE, _PSIInternalPssmData::pssm, query, _PSIInternalPssmData::scaled_pssm, and TRUE.

Referenced by _PSICreateAndScalePssmFromFrequencyRatios(), and BOOST_AUTO_TEST_CASE().

◆ _PSISequenceLengthWithoutX()

Uint4 _PSISequenceLengthWithoutX	(	const Uint1 *	seq,
		Uint4	length
	)

Calculates the length of the sequence without including any 'X' residues.

used in kappa.c

Parameters

seq	sequence to examine [in]
length	length of the sequence above [in]

Returns: number of non-X residues in the sequence

Definition at line 2629 of file blast_psi_priv.c.

References AMINOACID_TO_NCBISTDAA, ASSERT, and i.

Referenced by _PSIComputeScoreProbabilities().

◆ _PSISequenceWeightsFree()

_PSISequenceWeights* _PSISequenceWeightsFree ( _PSISequenceWeights * seq_weights )

Deallocates the _PSISequenceWeights structure.

Parameters

seq_weights data structure to deallocate [in]

Returns: NULL

Definition at line 627 of file blast_psi_priv.c.

References _PSIDeallocateMatrix(), _PSISequenceWeights::gapless_column_weights, _PSISequenceWeights::independent_observations, _PSISequenceWeights::match_weights, _PSISequenceWeights::match_weights_size, _PSISequenceWeights::norm_seq_weights, NULL, _PSISequenceWeights::posDistinctDistrib, _PSISequenceWeights::posDistinctDistrib_size, _PSISequenceWeights::posNumParticipating, _PSISequenceWeights::row_sigma, sfree, _PSISequenceWeights::sigma, and _PSISequenceWeights::std_prob.

Referenced by _PSISequenceWeightsNew(), Deleter< _PSISequenceWeights >::Delete(), and s_PSICreatePssmCleanUp().

◆ _PSISequenceWeightsNew()

_PSISequenceWeights* _PSISequenceWeightsNew	(	const PSIMsaDimensions *	dims,
		const BlastScoreBlk *	sbp
	)

Allocates and initializes the _PSISequenceWeights structure.

Parameters

dims	structure containing the multiple sequence alignment dimensions [in]
sbp	score block structure initialized for the scoring system used with the query sequence [in]

Returns: newly allocated structure or NULL in case of memory allocation failure

Definition at line 552 of file blast_psi_priv.c.

Referenced by BOOST_AUTO_TEST_CASE(), PSICreatePssmFromCDD(), and PSICreatePssmWithDiagnostics().

◆ _PSIStructureGroupCustomization()

void _PSIStructureGroupCustomization ( _PSIMsa * msa )

Enable NCBI structure group customization to discard the query sequence, as this really isn't the result of a PSI-BLAST iteration, but rather an artificial consensus sequence of the multiple sequence alignment constructed by them.

This should be called after _PSIPurgeBiasedSegments.

Definition at line 800 of file blast_psi_priv.c.

References _PSIUpdatePositionCounts(), _PSIMsa::cell, _PSIMsa::dimensions, FALSE, i, _PSIMsaCell::is_aligned, kQueryIndex, _PSIMsaCell::letter, and PSIMsaDimensions::query_length.

Referenced by PSICreatePssmWithDiagnostics().

◆ _PSIUpdateLambdaK()

void _PSIUpdateLambdaK	(	const int **	pssm,
		const Uint1 *	query,
		Uint4	query_length,
		const double *	std_probs,
		BlastScoreBlk *	sbp
	)

Updates the Karlin-Altschul parameters based on the query sequence and PSSM's score frequencies.

Port of blastool.c's updateLambdaK

Parameters

pssm	PSSM [in]
query	query sequence in ncbistdaa encoding [in]
query_length	length of the query sequence above [in]
std_probs	array containing the standard background residue probabilities [in]
sbp	Score block structure where the calculated lambda and K will be returned [in\|out]

Definition at line 2732 of file blast_psi_priv.c.

References _PSIComputeScoreProbabilities(), ASSERT, Blast_KarlinBlkUngappedCalc(), Blast_ScoreFreqFree(), Blast_KarlinBlk::K, BlastScoreBlk::kbp_gap_psi, BlastScoreBlk::kbp_gap_std, BlastScoreBlk::kbp_ideal, BlastScoreBlk::kbp_psi, log, Blast_KarlinBlk::logK, and query.

Referenced by _PSIScaleMatrix(), and impalaScaleMatrix().

◆ _PSIUpdatePositionCounts()

void _PSIUpdatePositionCounts ( _PSIMsa * msa )

Counts the number of sequences matching the query per query position (columns of the multiple alignment) as well as the number of residues present in each position of the query.

Should be called after multiple alignment data has been purged from biased sequences.

Parameters

msa	multiple sequence alignment structure [in\|out]

Definition at line 991 of file blast_psi_priv.c.

References _PSIMsa::alphabet_size, ASSERT, _PSIMsa::cell, _PSIMsa::dimensions, _PSIMsaCell::is_aligned, _PSIMsaCell::letter, _PSIMsa::num_matching_seqs, PSIMsaDimensions::num_seqs, PSIMsaDimensions::query_length, and _PSIMsa::residue_counts.

Referenced by _PSIMsaNew(), and _PSIStructureGroupCustomization().

◆ _PSIValidateCdMSA()

int _PSIValidateCdMSA	(	const PSICdMsa *	cd_msa,
		Uint4	alphabet_size
	)

Validation of multiple alignment of conserved domains structure.

Parameters

cd_msa	multiple alignment of CDs [in]
alphabet_size	alphabet size [in]

Returns: One of the errors defined above if validation fails or bad parameter is passed in, else PSI_SUCCESS

Definition at line 862 of file blast_psi_priv.c.

References AMINOACID_TO_NCBISTDAA, PSICdMsaCell::data, PSICdMsa::dimensions, fabs, PSICdMsaCellData::iobsr, PSICdMsaCell::is_aligned, kEpsilon, PSICdMsa::msa, PSIMsaDimensions::num_seqs, PSI_SUCCESS, PSIERR_BADPARAM, PSIERR_BADPROFILE, PSIERR_GAPINQUERY, PSICdMsa::query, PSIMsaDimensions::query_length, and PSICdMsaCellData::wfreqs.

Referenced by PSICreatePssmFromCDD().

◆ _PSIValidateMSA()

int _PSIValidateMSA	(	const _PSIMsa *	msa,
		Boolean	ignored_unaligned_positions
	)

Main validation function for multiple sequence alignment structure.

Should be called after _PSIPurgeBiasedSegments.

Parameters

msa	multiple sequence alignment data structure [in]
ignored_unaligned_positions	determines whether the unaligned positions test should be performend or not [in]

Returns: One of the errors defined above if validation fails or bad parameter is passed in, else PSI_SUCCESS

Definition at line 828 of file blast_psi_priv.c.

References PSI_SUCCESS, PSIERR_BADPARAM, s_PSIValidateAlignedColumns(), s_PSIValidateNoFlankingGaps(), s_PSIValidateNoGapsInQuery(), and s_PSIValidateParticipatingSequences().

Referenced by PSICreatePssmWithDiagnostics().

◆ _PSIValidateMSA_StructureGroup()

int _PSIValidateMSA_StructureGroup ( const _PSIMsa * msa )

Structure group validation function for multiple sequence alignment structure.

Should be called after _PSIStructureGroupCustomization.

Parameters

msa	multiple sequence alignment data structure [in]

Returns: One of the errors defined above if validation fails or bad parameter is passed in, else PSI_SUCCESS

Definition at line 811 of file blast_psi_priv.c.

References PSI_SUCCESS, PSIERR_BADPARAM, and s_PSIValidateParticipatingSequences().

Referenced by PSICreatePssmWithDiagnostics().

Variable Documentation

◆ kEpsilon

const double kEpsilon

extern

Small constant to test against 0.

Definition at line 58 of file blast_psi_priv.c.

Referenced by _PSICalculateInformationContentFromFreqRatios(), _PSICalculateInformationContentFromScoreMatrix(), _PSIComputeFreqRatios(), _PSIComputeFreqRatiosFromCDs(), _PSIConvertFreqRatiosToPSSM(), _PSISpreadGapWeights(), _PSIValidateCdMSA(), CalculateLinkHSPCutoffs(), PsiBlastSetupScoreBlock(), and CSeqGraphicRenderer::ZoomOnRange().

◆ kPositScalingNumIterations

const Uint4 kPositScalingNumIterations

extern

Constant used in scaling PSSM routines: Successor to POSIT_NUM_ITERATIONS.

Definition at line 61 of file blast_psi_priv.c.

Referenced by _PSIScaleMatrix(), and impalaScaleMatrix().

◆ kPositScalingPercent

const double kPositScalingPercent

extern

Constant used in scaling PSSM routines: Successor to POSIT_PERCENT.

Definition at line 60 of file blast_psi_priv.c.

Referenced by _PSIScaleMatrix(), and impalaScaleMatrix().

◆ kPSIIdentical

const double kPSIIdentical

extern

Percent identity threshold for discarding identical matches.

Definition at line 56 of file blast_psi_priv.c.

Referenced by s_PSIPurgeSelfHits().

◆ kPSINearIdentical

const double kPSINearIdentical

extern

Percent identity threshold for discarding near-identical matches.

Definition at line 55 of file blast_psi_priv.c.

Referenced by s_PSIPurgeNearIdenticalAlignments(), and CPssmInputTestData::SetupNearIdenticalHits().

◆ kPSIScaleFactor

const int kPSIScaleFactor

extern

Successor to POSIT_SCALE_FACTOR.

Definition at line 59 of file blast_psi_priv.c.

Referenced by _PSIConvertFreqRatiosToPSSM(), _PSIScaleMatrix(), and impalaScaleMatrix().

◆ kQueryIndex

const unsigned int kQueryIndex

extern

Index into multiple sequence alignment structure for the query sequence.

Definition at line 57 of file blast_psi_priv.c.

Referenced by _PSICheckSequenceWeights(), _PSIComputeAlignmentBlocks(), _PSIComputeFreqRatios(), _PSIMsaNew(), _PSISpreadGapWeights(), _PSIStructureGroupCustomization(), Blast_Message2TSearchMessages(), BOOST_AUTO_TEST_CASE(), s_ExtractSeqId(), s_PSIPurgeSelfHits(), s_PSIPurgeSimilarAlignments(), s_PSIValidateAlignedColumns(), s_PSIValidateNoGapsInQuery(), CPssmInputTestData::SetupDuplicateHit(), CPsiBlastInputClustalW::x_CopyQueryToMsa(), CPsiBlastInputData::x_CopyQueryToMsa(), CPsiBlastInputClustalW::x_ExtractAlignmentData(), and CPsiBlastInputData::x_ExtractAlignmentData().

Classes

Macros

Typedefs

Functions

Variables

Detailed Description

Macro Definition Documentation

◆ PSI_SUCCESS

◆ PSIERR_BADPARAM

◆ PSIERR_BADPROFILE

◆ PSIERR_BADSEQWEIGHTS

◆ PSIERR_COLUMNOFGAPS

◆ PSIERR_ENDINGGAP

◆ PSIERR_GAPINQUERY

◆ PSIERR_NOALIGNEDSEQS

◆ PSIERR_NOFREQRATIOS

◆ PSIERR_OUTOFMEM

◆ PSIERR_POSITIVEAVGSCORE

◆ PSIERR_STARTINGGAP

◆ PSIERR_UNALIGNEDCOLUMN

◆ PSIERR_UNKNOWN

Typedef Documentation

◆ _PSIAlignedBlock

◆ _PSIInternalPssmData

◆ _PSIMsa

◆ _PSIMsaCell

◆ _PSIPackedMsa

◆ _PSIPackedMsaCell

◆ _PSISequenceWeights

Function Documentation

◆ _IMPALAScaleMatrix()

◆ _PSIAlignedBlockFree()

◆ _PSIAlignedBlockNew()

◆ _PSIAllocateMatrix()

◆ _PSICalculateInformationContentFromFreqRatios()

◆ _PSICalculateInformationContentFromScoreMatrix()

◆ _PSIComputeAlignmentBlocks()

◆ _PSIComputeFreqRatios()

◆ _PSIComputeFreqRatiosFromCDs()

◆ _PSIComputeFrequenciesFromCDs()

◆ _PSIComputeScoreProbabilities()

◆ _PSIComputeSequenceWeights()

◆ _PSIConvertFreqRatiosToPSSM()

◆ _PSICopyMatrix_double()

◆ _PSICopyMatrix_int()

◆ _PSIDeallocateMatrix()

◆ _PSIInternalPssmDataFree()

◆ _PSIInternalPssmDataNew()

◆ _PSIMsaFree()

◆ _PSIMsaNew()

◆ _PSIPackedMsaFree()

◆ _PSIPackedMsaGetNumberOfAlignedSeqs()

◆ _PSIPackedMsaNew()

◆ _PSIPurgeAlignedRegion()

◆ _PSIPurgeBiasedSegments()

◆ _PSISaveCDDiagnostics()

◆ _PSISaveDiagnostics()

◆ _PSIScaleMatrix()

◆ _PSISequenceLengthWithoutX()

◆ _PSISequenceWeightsFree()

◆ _PSISequenceWeightsNew()

◆ _PSIStructureGroupCustomization()

◆ _PSIUpdateLambdaK()

◆ _PSIUpdatePositionCounts()

◆ _PSIValidateCdMSA()

◆ _PSIValidateMSA()

◆ _PSIValidateMSA_StructureGroup()

Variable Documentation

◆ kEpsilon

◆ kPositScalingNumIterations

◆ kPositScalingPercent

◆ kPSIIdentical

◆ kPSINearIdentical

◆ kPSIScaleFactor

◆ kQueryIndex