Tblastn Class Reference

#include <Match.h>

Inheritance diagram for Tblastn:

Tblastnmy Tblastnmy

List of all members.

Public Member Functions

 Tblastn ()
 Tblastn (const string &query, const string &target, int queryLen, int targetLen)
 Tblastn (const Tblastn &tbn)
Tblastn & operator= (const Tblastn &tbn)
 ~Tblastn ()
void clear ()
void addMatch (float identity, int alnlen, int mismatches, int gaps, int qbegin, int qend, int tbegin, int tend, double expect, double score)
void addMatch (M8Match *mptr)
void add (const Tblastn &tbn)
void setSortedByTarget (bool yes=true)
bool isSortedByTarget () const
ostream & output (ostream &ous, char delimiter[]="\t") const
void outputModel (const M8MatchChain &amodel, int &id, ostream &ouss, ostream &ousd, char delimiter[]="\t") const
void outputModel (const Boxchain &amodel, int &id, ostream &ouss, ostream &ousd, char delimiter[]="\t") const
void show (ostream &ous) const
int removeJunk (const Protein &qseq, const DNA &tseq, float factor=0.8)
void findFootprint (ostream &SUMM, ostream &DETA, float covcut=0.4)
vector< Match > findSections ()
bool empty ()
const string & query () const
const string & target () const
int numMatches () const
const vector< M8Match * > & getMatches () const
int queryLength () const
int targetLength () const
const Range & firstQueryRange () const
const Range & lastQueryRange () const
float qcov () const
bool isFirstQueryMatch (const M8Match *m) const
bool overlapFirstQueryMatch (const M8Match *m, int cutoff=9) const
bool isLastQueryMatch (const M8Match *m) const
bool overlapLastQueryMatch (const M8Match *m, int cutoff=9) const
int checkContain (ostream &ous, float factor=0.8) const
bool sameQTPair (const Tblastn &tbn)
void sortMatchByTarget ()
void directionalSortMatchByTarget ()

Static Public Member Functions

static void readConfig (const string &conf)
static double getIntronProbility (int len)

Private Member Functions

void erase (int b, int e)
int eraseMatch (const Range &r)
void removeNullMatch ()
void sortMatchByQuery () const
int removeAll (list< string > &rmlist, ostream &ous)
int removeAll ()

Private Attributes

string qid
string tid
int qlen
int tlen
vector< M8Match * > matches
bool sortedByTarget
vector< M8Match * > matchesQS
bool sortedByQuery

Static Private Attributes

static float qcovCutoff = 0.3
static vector< double > intronProb = vector<double>()

Friends

ostream & operator<< (ostream &ous, const Tblastn &b)

Detailed Description

this class contains a list of M8Matches

Constructor & Destructor Documentation

Tblastn::Tblastn (  )  [inline]

Tblastn::Tblastn ( const string &  query,
const string &  target,
int  queryLen,
int  targetLen 
) [inline]

Tblastn::Tblastn ( const Tblastn &  tbn  ) 

copy constructor for solving the boundary problem between sql LIMIT commands, need to save the result to be compared to the next trunk.

References matches.

Tblastn::~Tblastn (  ) 

References matches.

Member Function Documentation

Tblastn & Tblastn::operator= ( const Tblastn &  tbn  ) 

References matches, qid, qlen, sortedByQuery, sortedByTarget, tid, and tlen.

void Tblastn::clear (  )  [inline]

References removeAll().

Referenced by processOneBatch().

void Tblastn::addMatch ( float  identity,
int  alnlen,
int  mismatches,
int  gaps,
int  qbegin,
int  qend,
int  tbegin,
int  tend,
double  expect,
double  score 
)

add a new M8 match to this pair of queryxtarget The input field order is the same as the blast m8 output.

References matches.

Referenced by add(), Tblastnmy::addMatch(), and main().

void Tblastn::addMatch ( M8Match *  mptr  )  [inline]

References matches.

void Tblastn::add ( const Tblastn &  tbn  ) 

References addMatch(), matches, sortedByQuery, and sortedByTarget.

void Tblastn::setSortedByTarget ( bool  yes = true  )  [inline]

this method tell this object that the input data has already been sorted. If the input data came from a relational database, the input data can be easily sorted by the database engine. When taken data directly from the input file which is usually sorted by query,target you should not set this to true.

References sortedByTarget.

bool Tblastn::isSortedByTarget (  )  const [inline]

References sortedByTarget.

ostream & Tblastn::output ( ostream &  ous,
char  delimiter[] = "\t" 
) const

References matches, qid, and tid.

Referenced by removeAll().

void Tblastn::outputModel ( const M8MatchChain &  amodel,
int &  id,
ostream &  ouss,
ostream &  ousd,
char  delimiter[] = "\t" 
) const

ouss is the summary stream, ousd is the detailed stream

References M8MatchChain::_first, M8MatchChain::_last, M8MatchChain::avgIdentity(), M8MatchChain::chain, M8MatchChain::direction(), M8MatchChain::getQueryBegin(), M8MatchChain::getQueryEnd(), M8MatchChain::getTargetBegin(), M8MatchChain::getTargetEnd(), qid, qlen, M8MatchChain::queryLength(), M8MatchChain::queryOverlap(), M8MatchChain::size(), M8MatchChain::sumScore(), tid, and tlen.

Referenced by findFootprint(), and storeBestModels().

void Tblastn::outputModel ( const Boxchain &  amodel,
int &  id,
ostream &  ouss,
ostream &  ousd,
char  delimiter[] = "\t" 
) const

the Boxchain version to replace the old version

Summary output: fields: qid, tid, fpnum, qlen, tlen, qbegin, qend, tbegin, tend, num_exon, sumscore, avgiden, qcov, sumoverlap_fraction

fpnum is the id to count how may models are produced from the same q x t pair this is to make (q,t,fpnum) unique

Note: two fields fewer than the previous version. qcov <= 1, different from definition before.

Parameters:
ousd Detail of footprint output stream
Detail stream output field: qid, tid, fpnum, identity, alnlen, mismatches, gaps, qbegin, qend, tbegin, tend, E, score, exnum

The exnum is a counter from 1 to the number of exons of the footprint.

References Boxchain::first(), M8MatchEX::next(), M8Match::output(), qid, qlen, Boxchain::queryOverlap(), Boxchain::querySeqCovered(), Boxchain::summaryOutput(), tid, and tlen.

void Tblastn::show ( ostream &  ous  )  const

References matches, qid, qlen, tid, and tlen.

int Tblastn::removeJunk ( const Protein &  qseq,
const DNA &  tseq,
float  factor = 0.8 
)

this method will modify the matches vector. Junk rows will be removed.

Arguments: qseq query sequence (protein) tseq target sequence (genomic). This is passed to this function, but not used yet at this point. Could be useful for better algorithms. factor: the argument to check containment. ous: the stream to store the bad rows.

The removed results will be added to rows. The format is quoted comma separated string suitable for SQL insert statement.

This function uses the static qcovCutoff member for judgement.

return the number of matches removed. BAD features: (1) qcov < qcovCutoff

Not doing simplicity check. It is supposed to be done by previous steps. New version, not recording removed entries!

References matches, Range::merge(), Range::overlapFraction(), qcov(), qcovCutoff, removeAll(), removeNullMatch(), and sortMatchByTarget().

void Tblastn::findFootprint ( ostream &  SUMM,
ostream &  DETA,
float  covcut = 0.4 
)

locate all the matches of each protein on the genomic DNA. Write output to SUMM stream and the DETA stream Only good foot prints will be selected. covcut determines the coverage of the query

Parameters:
covcut cutoff for query coverage, default 0.4

References M8MatchChain::add(), directionalSortMatchByTarget(), firstQueryRange(), isFirstQueryMatch(), isLastQueryMatch(), matches, M8MatchChain::nofirst(), M8MatchChain::nolast(), outputModel(), overlapFirstQueryMatch(), overlapLastQueryMatch(), queryLength(), and M8MatchChain::queryLength().

Referenced by main().

vector< Match > Tblastn::findSections (  ) 

divide the matches into sections and remove bad sections

References directionalSortMatchByTarget(), eraseMatch(), getIntronProbility(), matches, queryLength(), sortedByTarget, and targetLength().

Referenced by Linkmatch::Linkmatch().

bool Tblastn::empty (  )  [inline]

References matches.

Referenced by processOneBatch().

const string& Tblastn::query (  )  const [inline]

References qid.

const string& Tblastn::target (  )  const [inline]

References tid.

int Tblastn::numMatches (  )  const [inline]

number of match segments in this pair of query x target

References matches.

Referenced by removeAll(), and storeBestModels().

const vector<M8Match*>& Tblastn::getMatches (  )  const [inline]

References matches.

Referenced by Linkmatch::Linkmatch().

int Tblastn::queryLength (  )  const [inline]

References qlen.

Referenced by findFootprint(), findSections(), and storeBestModels().

int Tblastn::targetLength (  )  const [inline]

References tlen.

Referenced by findSections().

const Range & Tblastn::firstQueryRange (  )  const

References matchesQS, and sortMatchByQuery().

Referenced by findFootprint().

const Range & Tblastn::lastQueryRange (  )  const

References matchesQS, and sortMatchByQuery().

float Tblastn::qcov (  )  const

merge all the matches on the query into a RangeChain then compute the sum of the Range over the length of the query

References RangeChain::add(), RangeChain::length(), matches, and qlen.

Referenced by removeJunk().

bool Tblastn::isFirstQueryMatch ( const M8Match *  m  )  const

to check a particular match is at the start of the query or at the end of the query

References matchesQS, Match::queryRange(), and sortMatchByQuery().

Referenced by findFootprint().

bool Tblastn::overlapFirstQueryMatch ( const M8Match *  m,
int  cutoff = 9 
) const

cutoff is the smallest overlap test M8Match* m overlap with the first query match. It uses the matcheQS vector to do the job.

References matchesQS, Range::overlap(), Match::queryRange(), and sortMatchByQuery().

Referenced by findFootprint().

bool Tblastn::isLastQueryMatch ( const M8Match *  m  )  const

References matchesQS, Match::queryRange(), and sortMatchByQuery().

Referenced by findFootprint().

bool Tblastn::overlapLastQueryMatch ( const M8Match *  m,
int  cutoff = 9 
) const

References matchesQS, Range::overlap(), Match::queryRange(), and sortMatchByQuery().

Referenced by findFootprint().

int Tblastn::checkContain ( ostream &  ous,
float  factor = 0.8 
) const

Check to see which match is contained in another. The entries that can be removed is written the the output stream ous. The factor is use to decide if the contained entry should be replaced by the formula: identity > factor*identity_of_larger_match or score > factor*score_of_larger_match

References matches, qid, and tid.

bool Tblastn::sameQTPair ( const Tblastn &  tbn  )  [inline]

References qid, and tid.

void Tblastn::sortMatchByTarget (  ) 

this will sort the matches in the order target,query,tbegin,tend,qbegin,qend Only carry out the sorting operation if the sortedByTarget is false. This sorting is different when two overlapping matches are in different directions, For our operation this operation is more correct, so we should not have used the SQL order. |--A--> |--A-> <-B-| <-B---| In the left case, SQL order A before B, our sorting would place B before A which is what we want. In the right case, SQL and our sorting would be the same.

This method uses the cmpTargetPtr functin object which in turn use the < operator of Range.

References matches, and sortedByTarget.

Referenced by removeJunk().

void Tblastn::directionalSortMatchByTarget (  ) 

this sorting is better for sections

References matches, and sortedByTarget.

Referenced by findFootprint(), and findSections().

void Tblastn::readConfig ( const string &  conf  )  [static]

read the config from file with path conf

References ifstream(), and intronProb.

Referenced by main().

double Tblastn::getIntronProbility ( int  len  )  [static]

References intronProb.

Referenced by findSections().

void Tblastn::erase ( int  b,
int  e 
) [private]

erase elements in matches from b to e [b, e] inclusive. use delete to deallocate memory, then set pointer to 0

References matches, and removeNullMatch().

int Tblastn::eraseMatch ( const Range &  r  )  [private]

Erase all matches inside r. use the Range.inside() function. return the number of matches removed

References matches, and removeNullMatch().

Referenced by findSections().

void Tblastn::removeNullMatch (  )  [private]

remove elements in matches whose value is 0 (null pointer)

References matches, and sortedByQuery.

Referenced by erase(), eraseMatch(), and removeJunk().

void Tblastn::sortMatchByQuery (  )  const [private]

References matches, matchesQS, and sortedByQuery.

Referenced by firstQueryRange(), isFirstQueryMatch(), isLastQueryMatch(), lastQueryRange(), overlapFirstQueryMatch(), and overlapLastQueryMatch().

int Tblastn::removeAll ( list< string > &  rmlist,
ostream &  ous 
) [private]

this version is for debuging

References matches, numMatches(), output(), qid, sortedByQuery, and tid.

int Tblastn::removeAll (  )  [private]

for production, we want to go faster Remove all matches, deallocate memorey. clear the matches vector.

References matches, numMatches(), and sortedByQuery.

Referenced by clear(), and removeJunk().

Friends And Related Function Documentation

ostream& operator<< ( ostream &  ous,
const Tblastn &  b 
) [friend]

Member Data Documentation

string Tblastn::qid [private]

Referenced by checkContain(), operator<<(), operator=(), output(), outputModel(), query(), removeAll(), sameQTPair(), and show().

string Tblastn::tid [private]

Referenced by checkContain(), operator<<(), operator=(), output(), outputModel(), removeAll(), sameQTPair(), show(), and target().

int Tblastn::qlen [private]

Referenced by operator<<(), operator=(), outputModel(), qcov(), queryLength(), and show().

int Tblastn::tlen [private]

Referenced by operator<<(), operator=(), outputModel(), show(), and targetLength().

vector<M8Match*> Tblastn::matches [private]

for sorting we need vector, for deletion we need list the matches are not sorted in any particlar order. However, the pfog program sorted the input as follows: target,query,tbegin,tend,qbegin,qend we should keep this information. with sortedByTarget

Referenced by add(), addMatch(), checkContain(), directionalSortMatchByTarget(), empty(), erase(), eraseMatch(), findFootprint(), findSections(), getMatches(), numMatches(), operator<<(), operator=(), output(), qcov(), removeAll(), removeJunk(), removeNullMatch(), show(), sortMatchByQuery(), sortMatchByTarget(), Tblastn(), and ~Tblastn().

bool Tblastn::sortedByTarget [private]

Referenced by add(), directionalSortMatchByTarget(), findSections(), isSortedByTarget(), operator=(), setSortedByTarget(), and sortMatchByTarget().

vector<M8Match*> Tblastn::matchesQS [mutable, private]

this is a copy of the matches, sorted by Query if any modification is made to matches, then the sortedByQuery is set to false Before using this vector, you must check the sortedByQuery value.

If the matches change then this should be re-created.

Referenced by firstQueryRange(), isFirstQueryMatch(), isLastQueryMatch(), lastQueryRange(), overlapFirstQueryMatch(), overlapLastQueryMatch(), and sortMatchByQuery().

bool Tblastn::sortedByQuery [mutable, private]

when false, then matchesQS needs to be recreated and sorted from matches

Referenced by add(), operator=(), removeAll(), removeNullMatch(), and sortMatchByQuery().

float Tblastn::qcovCutoff = 0.3 [static, private]

it stores the pointers to all the models created by the findFootPrint() method

Referenced by removeJunk().

vector< double > Tblastn::intronProb = vector<double>() [static, private]

store the intron probaility read from Match.conf file

Referenced by getIntronProbility(), and readConfig().

The documentation for this class was generated from the following files:
Generated on Wed Aug 10 11:57:15 2011 for Softwares from Orpara by  doxygen 1.5.6