org.strbio.mol
Class ThreadSet

java.lang.Object
  org.strbio.mol.ThreadSet

public class ThreadSet
extends java.lang.Object

Class to keep track of two sets of proteins that are being used in threading algorithms.

 Version 2.0, 11/12/02 - can load from 3DPSSM file
 Version 1.8, 1/30/1 - fixed bug where ASpc was printed as ASns
 Version 1.71, 8/28/00 - saveStrMatches follows .casp suffix
 Version 1.7, 4/11/00 - fixed bug where alignmentParameters2 was
   saying it was finished with a given sequence... had to be re-setup()
 Version 1.69, 12/7/99 - fixed bug in probabilityOfFindingMatch
 Version 1.68, 9/17/99 - added makeCATHStrMatches
 Version 1.67, 9/2/99 - added alignmentParameters2
 Version 1.66, 7/14/99 - fixed API for globalCompare
 Version 1.65, 7/8/99 - added getEstPMatch
 Version 1.64, 6/25/99 - added calcEstPMatch function, moved gapModel
   and scoreList into alignmentParameters
 Version 1.63, 6/9/99 - added reciprocal weighted average
 Version 1.62, 6/3/99 - added median functions, fixed bug in
   definition of 'same sequence'
 Version 1.61, 3/31/99 - saves avg score with/out gaps
 Version 1.6, 2/26/99 - fixed to conform to new GapModel api
 Version 1.5, 2/22/99 - fixed to conform to new AlignmentSet api
 Version 1.4, 10/20/98 - added new Alignment format to loadStrMatches
 Version 1.33, 10/15/98 - added naildown alignment to loadStrMatches
 Version 1.32, 10/7/98 - added minareaInfoAll
 Version 1.31, 9/30/98 - added minareaInfo, removed 1.3 change
 Version 1.3, 9/28/98 - added ratio score output option to
   alignStrMatchesMinarea
 Version 1.22, 9/1/98 - don't let somebody add scoring function with
   zero weight, to minimize stupid errors
 Version 1.21, 7/22/98 - alignStrMatchesMinarea auto-renumbers polymers
 Version 1.2, 4/23/98 - uses PolymerSet, made everything double precision
 Version 1.1, 4/6/98 - added all functions from thread_set.cpp
 Version 1.0, 4/3/98 - original version
 

Version:

2.0, 11/12/02

Author:

JMC

See Also:

PolymerSet, Alignment

Field Summary

AlignmentParameters	alignmentParameters Alignment parameters
AlignmentParameters	alignmentParameters2 Alignment parameters for evaluating alignment goodness in FR tests; Null means use same one as that used to generate the alignment.
PolymerSet	folds Folds being threaded.
PolymerSet	seqs Sequences being threaded.
int	tolerance margin for error in evaluating alignments.

Constructor Summary

ThreadSet()
Initialize everything to default values.

Method Summary

void	addStrMatch(AlignmentSet aSet) add a single structural match to the set, created by the user
AlignmentSet	addStrMatch(Polymer seq, Polymer fold) add a single structural match to the set.
void	addStrMatch(java.lang.String seq_name, java.lang.String fold_name, java.lang.String correct_name, java.lang.String nail_name) add a single structural match to the set.
void	alignmentStats(Printf outfile) compute stats on alignment.
void	alignStrMatches() Calculated alignment for all structural matches, without printing anything.
void	alignStrMatches(Printf outfile, boolean showAlignment, boolean showModeller) Aligns structural matches.
void	alignStrMatchesMinarea() redo all structural matches using minarea, and save the results to a single file
double[]	ASns(int tolerance) calculate alignment specificity, as in CASP2.
double[]	ASpc(int tolerance) calculate alignment specificity, as in CASP2.
double	averageRank() average rank of structural matches
double	averageRankRW() average rank with reciprocal weighting
double	averageZ() average z score of structural matches
DVector	calcMatchRank() Calculate vector of ranks, showing one rank for each match.
DVector	calcMatchZ() Calculate vector of z scores, showing one z score for each match.
void	clearSame() Forget that seq-fold pairs have the same name.
void	clearScores() forget about any alignments done so far.
void	clearStrMatches() Deny all knowledge of structural matches.
void	copyCorrectFrom(ThreadSet ts) Copy correct alignments from another threadSet to this one's calculated alignments.
void	findSameSeqs() Find seq-fold pairs with the same name.
DVector	getEstPMatch(int i) Get a DVector of estimated probabilities for one of the proteins in the seqs set, or null if this has not been calculated.
IVector	getSortedScores(int i) Get an IVector of sorted probabilities for one of the proteins in the seqs set, from most to least likely, or null if this has not been calculated.
AlignmentSet	getStrMatch(int i) Return a given structural match, or null if out of range.
AlignmentSet	getStrMatch(Polymer seq, Polymer fold) Get first structural match, if one exists, with a given seq and fold, or null if one doesn't exist.
void	globalCompare(Printf outfile, boolean optZ, boolean optN) do global comparison of all sequences, showing scores and z-scores.
boolean	isMatch(int n, int m) is sequence N a known structural match for fold M?
boolean	isSame(int n, int m) is sequence N the same as fold M?
int	load3DPSSM(java.lang.String filename) loads seqs, folds, and str matches from 3DPSSM output (mailbox of messages from server).
int	loadStrMatches(java.lang.String filename) get structural matches from a file, with or without correct and naildown alignments.
int	makeAllStrMatches() add all possible structural matches to the set.
void	makeCATHStrMatches() make a set of StrMatches using CATH data.
double	medianRank() median rank of structural matches
double	medianZ() median z score of structural matches
void	minareaInfo(Printf outfile) Show info on all structural matches.
void	minareaInfoAll(Printf outfile) Show info on all seq/fold pairs.
int	nMatches() How many structural matches do we know about?
int	nSame() How many seq-fold pairs are the same?
int	numberFound(int cutoff) number of structural matches in top N scores (cutoff)
void	oldAlignStrMatchesMinarea() Deprecated.
void	optimizeAlignment(Printf outfile) optimize the gap penalties for highest alignment accuracy.
void	optimizeASns(Printf outfile) optimize the gap penalties for highest asns(1).
void	optimizeMedianRank(Printf outfile) optimize the gap penalties for best median rank of known structural matches.
void	optimizeRank(Printf outfile) optimize the gap penalties for best average rank of known structural matches.
void	optimizeRankRW(Printf outfile) optimize the gap penalties for best reciprocal weighted average rank of known structural matches.
void	optimizeTopN(int n, Printf outfile) optimize the gap penalties for most correct structural matches in the top N.
void	optimizeZScore(Printf outfile) optimize the gap penalties for highest average Z score among structural matches.
double[]	pctRight(int tolerance) Returns pctRight for the structural matches.
DVector	probabilityOfFindingMatch() Calculates confidence vector... this is a DVector of the length of the number of folds.
void	removeStrMatchesOver(double score) Remove structural matches over a certain minarea score.
void	removeStrMatchesUnder(double score) Remove structural matches under a certain minarea score.
void	saveEstAAStats(java.lang.String filename) save the stats needed for alignment accuracy estimation to a file.
void	saveEstFRStats(java.lang.String filename) save the stats needed for FR accuracy estimation to a file.
void	saveStrMatches(java.lang.String filename) Save structural matches to a file.
void	saveStrMatchesModeller(java.lang.String filename) Save structural matches to a modeller format file.
void	setHookeParameters(double rho, double epsilon, int itermax, int iterstart) used to set optimizing parameters for optimize* functions; these are the Hooke parameters.
void	showFRAll(Printf outfile) show comparison results for all proteins.
void	showFRMatches(Printf outfile) show comparison results for str matches.
void	showPMatch(Printf outfile) shows probability of a structural match for various Z score ranges.

Methods inherited from class java.lang.Object

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

Field Detail

seqs

public PolymerSet seqs

Sequences being threaded.

folds

public PolymerSet folds

Folds being threaded.

alignmentParameters

public AlignmentParameters alignmentParameters

Alignment parameters

alignmentParameters2

public AlignmentParameters alignmentParameters2

Alignment parameters for evaluating alignment goodness in FR tests; Null means use same one as that used to generate the alignment.

tolerance

public int tolerance

margin for error in evaluating alignments.

Constructor Detail

ThreadSet

public ThreadSet()

Initialize everything to default values. Alignment paramters are set to null.

Method Detail

nMatches

public final int nMatches()

How many structural matches do we know about?

isMatch

public final boolean isMatch(int n,
                             int m)

is sequence N a known structural match for fold M?

isSame

public final boolean isSame(int n,
                            int m)

is sequence N the same as fold M?

copyCorrectFrom

public final void copyCorrectFrom(ThreadSet ts)

Copy correct alignments from another threadSet to this one's calculated alignments. Very big kludge. Does no error checking whatsoever.

load3DPSSM

public final int load3DPSSM(java.lang.String filename)

loads seqs, folds, and str matches from 3DPSSM output (mailbox of messages from server). E values stored in AlignmentStats' estPMatch field. returns number of structual alignments read.

loadStrMatches

public final int loadStrMatches(java.lang.String filename)

get structural matches from a file, with or without correct and naildown alignments. returns number found.

saveStrMatches

public final void saveStrMatches(java.lang.String filename)
                          throws java.io.IOException

Save structural matches to a file.

Throws:

java.io.IOException

saveStrMatchesModeller

public final void saveStrMatchesModeller(java.lang.String filename)
                                  throws java.io.IOException

Save structural matches to a modeller format file.

Throws:

java.io.IOException

alignStrMatchesMinarea

public final void alignStrMatchesMinarea()

redo all structural matches using minarea, and save the results to a single file

See Also:

AlignmentSet.minareaAlign()

oldAlignStrMatchesMinarea

public final void oldAlignStrMatchesMinarea()
                                     throws java.io.IOException

Deprecated.

redo all structural matches using minarea, and save the results to files called 'seqname.foldname.al_minarea'.

Throws:

java.io.IOException

See Also:

AlignmentSet.minareaAlign()

minareaInfo

public final void minareaInfo(Printf outfile)
                       throws java.io.IOException

Show info on all structural matches.

Throws:

java.io.IOException

See Also:

AlignmentSet.minareaInfo(org.strbio.io.Printf)

minareaInfoAll

public final void minareaInfoAll(Printf outfile)
                          throws java.io.IOException

Show info on all seq/fold pairs.

Throws:

java.io.IOException

See Also:

AlignmentSet.minareaInfo(org.strbio.io.Printf)

clearSame

public final void clearSame()

Forget that seq-fold pairs have the same name.

findSameSeqs

public final void findSameSeqs()

Find seq-fold pairs with the same name.

nSame

public final int nSame()

How many seq-fold pairs are the same?

addStrMatch

public final void addStrMatch(java.lang.String seq_name,
                              java.lang.String fold_name,
                              java.lang.String correct_name,
                              java.lang.String nail_name)

add a single structural match to the set.

addStrMatch

public final void addStrMatch(AlignmentSet aSet)

add a single structural match to the set, created by the user

addStrMatch

public final AlignmentSet addStrMatch(Polymer seq,
                                      Polymer fold)

add a single structural match to the set. Returns the AlignmentSet created.

makeAllStrMatches

public final int makeAllStrMatches()

add all possible structural matches to the set.

makeCATHStrMatches

public final void makeCATHStrMatches()

make a set of StrMatches using CATH data. Matches must agree at the H level or below. CATH data must be present for every seqs and folds which is considered.

clearStrMatches

public final void clearStrMatches()

Deny all knowledge of structural matches.

removeStrMatchesOver

public final void removeStrMatchesOver(double score)

Remove structural matches over a certain minarea score. This is probably not generally useful, so isn't well debugged.

removeStrMatchesUnder

public final void removeStrMatchesUnder(double score)

Remove structural matches under a certain minarea score. This is probably not generally useful, so isn't well debugged.

getStrMatch

public final AlignmentSet getStrMatch(int i)

Return a given structural match, or null if out of range.

getStrMatch

public final AlignmentSet getStrMatch(Polymer seq,
                                      Polymer fold)

Get first structural match, if one exists, with a given seq and fold, or null if one doesn't exist.

alignStrMatches

public final void alignStrMatches(Printf outfile,
                                  boolean showAlignment,
                                  boolean showModeller)
                           throws java.io.IOException

Aligns structural matches.

Throws:

java.io.IOException

alignStrMatches

public final void alignStrMatches()
                           throws java.io.IOException

Calculated alignment for all structural matches, without printing anything.

Throws:

java.io.IOException

pctRight

public final double[] pctRight(int tolerance)

Returns pctRight for the structural matches. Must have been already aligned. Return format is double[3] containing percent right, nCorrect, and nAligned.

ASpc

public final double[] ASpc(int tolerance)

calculate alignment specificity, as in CASP2. Returns a double[4] with the ASpc value, the ACrct value (correctly aligned positions), Na (aligned positions in this alignment), and NaC (aligned positions in the correct alignment) Must have been already aligned.

ASns

public final double[] ASns(int tolerance)

calculate alignment specificity, as in CASP2. Returns a double[4] with the ASpc value, the ACrct value (correctly aligned positions), Na (aligned positions in this alignment), and NaC (aligned positions in the correct alignment) Must have been already aligned.

alignmentStats

public final void alignmentStats(Printf outfile)

compute stats on alignment. For debugging purposes only; don't use it.

clearScores

public final void clearScores()

forget about any alignments done so far.

getEstPMatch

public final DVector getEstPMatch(int i)

Get a DVector of estimated probabilities for one of the proteins in the seqs set, or null if this has not been calculated.

getSortedScores

public final IVector getSortedScores(int i)

Get an IVector of sorted probabilities for one of the proteins in the seqs set, from most to least likely, or null if this has not been calculated.

saveEstAAStats

public final void saveEstAAStats(java.lang.String filename)
                          throws java.io.IOException

save the stats needed for alignment accuracy estimation to a file.

Throws:

java.io.IOException

saveEstFRStats

public final void saveEstFRStats(java.lang.String filename)
                          throws java.io.IOException

save the stats needed for FR accuracy estimation to a file.

Throws:

java.io.IOException

globalCompare

public final void globalCompare(Printf outfile,
                                boolean optZ,
                                boolean optN)
                         throws java.io.IOException

do global comparison of all sequences, showing scores and z-scores. Set optZ if optimizing for z scores (it saves time by not calculating sorted scores). Set optN if optimizing for top N scores (it saves time by not calculating z scores). If either is set, showAll, showMatch, outfile should all be off.

Parameters:

outfile - show results here

optZ - don't calculate top N scores

optN - don't calculate Z scores

Throws:

java.io.IOException

showFRAll

public final void showFRAll(Printf outfile)
                     throws java.io.IOException

show comparison results for all proteins.

Throws:

java.io.IOException

showFRMatches

public final void showFRMatches(Printf outfile)
                         throws java.io.IOException

show comparison results for str matches.

Throws:

java.io.IOException

numberFound

public final int numberFound(int cutoff)

number of structural matches in top N scores (cutoff)

calcMatchZ

public final DVector calcMatchZ()

Calculate vector of z scores, showing one z score for each match.

calcMatchRank

public final DVector calcMatchRank()

Calculate vector of ranks, showing one rank for each match. This is a DVector, for convenience in doing statistics.

averageZ

public final double averageZ()

average z score of structural matches

averageRank

public final double averageRank()

average rank of structural matches

medianRank

public final double medianRank()

median rank of structural matches

averageRankRW

public final double averageRankRW()

average rank with reciprocal weighting

medianZ

public final double medianZ()

median z score of structural matches

probabilityOfFindingMatch

public final DVector probabilityOfFindingMatch()

Calculates confidence vector... this is a DVector of the length of the number of folds. In every bin, it shows the probability that for a given seq there was a matching fold found in the top N-1 hits. Only sequences for which there is a matching fold are considered.

showPMatch

public final void showPMatch(Printf outfile)
                      throws java.io.IOException

shows probability of a structural match for various Z score ranges. The lowest range should contain no matches.

Throws:

java.io.IOException

setHookeParameters

public void setHookeParameters(double rho,
                               double epsilon,
                               int itermax,
                               int iterstart)

used to set optimizing parameters for optimize* functions; these are the Hooke parameters.

See Also:

Hooke

optimizeAlignment

public void optimizeAlignment(Printf outfile)
                       throws java.io.IOException

optimize the gap penalties for highest alignment accuracy.

Throws:

java.io.IOException

optimizeASns

public void optimizeASns(Printf outfile)
                  throws java.io.IOException

optimize the gap penalties for highest asns(1).

Throws:

java.io.IOException

optimizeZScore

public void optimizeZScore(Printf outfile)
                    throws java.io.IOException

optimize the gap penalties for highest average Z score among structural matches.

Throws:

java.io.IOException

optimizeRank

public void optimizeRank(Printf outfile)
                  throws java.io.IOException

optimize the gap penalties for best average rank of known structural matches.

Throws:

java.io.IOException

optimizeRankRW

public void optimizeRankRW(Printf outfile)
                    throws java.io.IOException

optimize the gap penalties for best reciprocal weighted average rank of known structural matches.

Throws:

java.io.IOException

optimizeMedianRank

public void optimizeMedianRank(Printf outfile)
                        throws java.io.IOException

optimize the gap penalties for best median rank of known structural matches.

Throws:

java.io.IOException

optimizeTopN

public void optimizeTopN(int n,
                         Printf outfile)
                  throws java.io.IOException

optimize the gap penalties for most correct structural matches in the top N.

Throws:

java.io.IOException