//#ifdef _MSC_VER
//#ifndef MSSINGLESCAN_MAIN
//int mssinglescan_main(int argc, char * argv[] ) {
//#else
//int main(int argc, char *argv[]){
//#endif
//#else
int main(int argc, char *argv[]){
//#endif
//Here are all the variable we are going to need
MSReader r;
Spectrum s;
int j;
if(argc==1){
printf("DESCRIPTION: Reads an MS/MS spectrum from any MSToolkit supported file type and outputs to screen in MS2 format.\n\n");
printf("USAGE: MSSingleScan [scan number] [file]\n");
exit(0);
};
r.readFile(argv[2],s,atoi(argv[1]));
if(s.getScanNumber()==0) exit(-1);
printf("S\t%d\t%d\t%.*f\n",s.getScanNumber(),s.getScanNumber(),2,s.getMZ());
if(s.getRTime()>0) printf("I\tRTime\t%.*f\n",4,s.getRTime());
for(j=0;j<s.sizeZ();j++){
printf("Z\t%d\t%.*f\n",s.atZ(j).z,2,s.atZ(j).mz);
};
for(j=0;j<s.size();j++){
printf("%.4f %.4f\n",s.at(j).mz,s.at(j).intensity);
};
return 0;
};
// assumes at least one spectrum in allRefs
void SearchLibrary::runSearch(Spectrum& s)
{
scoreMatches(s, cachedSpectra_, targetMatches_);
scoreMatches(s, cachedDecoySpectra_, decoyMatches_);
// keep scores from all target psms for estimating Weibull parameters
vector<double> allScores;
if(compute_pvalues_){
for(size_t i=0; i < targetMatches_.size(); i++){
double dotp = targetMatches_[i].getScore(DOTP);
allScores.push_back(dotp);
}
}
// there may have been spectra in cachedSpectra_ but none at the
// correct charge state. Check again
if( targetMatches_.size() == 0 ){
Verbosity::warn("No library spectra found for query %d "
"(precursor m/z %.2f).", s.getScanNumber(), s.getMz());
return;
}
if( compute_pvalues_ ){
addNullScores(s, allScores);
}
// sort the matches descending
sort(targetMatches_.begin(), targetMatches_.end(), compMatchDotScore);
sort(decoyMatches_.begin(), decoyMatches_.end(), compMatchDotScore);
setRank();
if( printAll_ ){
cout << "spec " << s.getScanNumber() << endl;
}
if( compute_pvalues_ ){
weibullEstimator_.estimateParams(allScores);
// print params to file
if( weibullParamFile_.is_open() ){
weibullParamFile_ << s.getScanNumber() << "\t"
<< weibullEstimator_.getEta() << "\t"
<< weibullEstimator_.getBeta() << "\t"
<< weibullEstimator_.getShift() << "\t"
<< weibullEstimator_.getCorrelation() << "\t"
<< weibullEstimator_.getNumPointsFit()
//(int)(allScores.size() * fraction_to_fit_)
<< endl;
}
setMatchesPvalues(allScores.size());
}
}
/**
* Generate more scores by creating decoy spectra and comparing them
* to query. Create one decoy for each target until there are the
* minimum number of scores. Do not save decoy spectrum or its
* Match. Makes no changes to cachedSpectra_.
*/
void SearchLibrary::addNullScores(Spectrum s, vector<double>& allScores){
int shiftAmount = 5;
while((int)allScores.size() < minWeibullScores_) {
int specAdded = 0; // make sure spectra were successfully added
//loop through all candidate refs, create shifted spectrum, compare
for(size_t i=0; i < targetMatches_.size(); i++) {
const RefSpectrum* targetSpec = targetMatches_.at(i).getRefSpec();
RefSpectrum* decoySpec = targetSpec->newDecoy(shiftAmount,
shiftRawSpectra_);
if( decoySpec == NULL ){
continue;
}
specAdded++;
Match thisMatch(&s, decoySpec); // spectrum deleted with match?
thisMatch.setMatchLibID(0);
Verbosity::comment(V_ALL,
"Comparing query spec %d and shifted library spec %d",
s.getScanNumber(), decoySpec->getLibSpecID() );
DotProduct::compare(thisMatch);
allScores.push_back(thisMatch.getScore(DOTP));
} // next ref spectrum
shiftAmount += 5;
if( specAdded == 0 ){
break; // avoid infinite loop if number of scores not increasing
}
} // next pass through all ref spectra
}
/**
* Compare the given query spectrum to all library spectra. Create a
* match for each and add to matches.
*/
void SearchLibrary::scoreMatches(Spectrum& s, deque<RefSpectrum*>& spectra,
vector<Match>& matches ){
Verbosity::debug("Scoring %d matches.", spectra.size());
// get the charge states we will search
const vector<int>& charges = s.getPossibleCharges();
// compare all ref spec to query, create match for each
for(size_t i=0; i< spectra.size(); i++) {
// is there a better place to check this?
if(spectra.at(i)->getNumProcessedPeaks() == 0 ){
Verbosity::debug("Skipping library spectrum %d. No peaks.",
spectra.at(i)->getLibSpecID());
continue;
}
if( ! checkCharge(charges, spectra.at(i)->getCharge()) ){
continue;
}
Match thisMatch(&s, spectra.at(i));
thisMatch.setMatchLibID(spectra.at(i)->getLibID());
Verbosity::comment(V_ALL, "Comparing query spec %d and library spec %d",
s.getScanNumber(), spectra.at(i)->getLibSpecID());
DotProduct::compare(thisMatch); //static method
// save match for reporting
matches.push_back(thisMatch);
}
}
bool CPeptideDatabase::readMascot(char* mascotFile, char* dataFile){
MSReader r;
MascotParser p;
Spectrum s;
MascotLite m;
int i;
int percent=0;
p.clear();
p.readText(mascotFile);
if(p.isDistiller()){
cout << mascotFile << "\t" << p.size() << " peptides from Distiller, no need for RT lookup." << endl;
} else if(p.hasRTime()){
cout << mascotFile << "\t" << p.size() << " peptides have RT already, no need for lookup." << endl;
} else {
cout << mascotFile << "\tLooking up retention times for " << p.size() << " Mascot peptides..." << endl;
cout << percent;
r.readFile(dataFile,s);
}
for(i=0;i<p.size();i++){
if(p.isDistiller()||p.hasRTime()){
m.rTime=p[i].rTime;
} else {
if( (int)i*100/p.size() > percent){
percent=(int)i*100/p.size();
cout << "\b\b\b" << percent;
}
r.readFile(NULL,s,p[i].scanNum);
if(s.getScanNumber()==0){
cout << "Scan not found. Please check that correct data file is used. Stopping Mascot parsing." << endl;
return false;
}
m.rTime=s.getRTime();
}
m.charge=p[i].charge;
m.fileID=fileID;
m.monoMass=p[i].zeroMass;
m.scanNum=p[i].scanNum;
strcpy(m.peptide,p[i].sequence_Long);
strcpy(m.protein,p[i].gene);
mascot.push_back(m);
}
cout << " Done!" << endl;
fileID++;
return true;
}
bool CNoiseReduction::DeNoiseD(Spectrum& sp){
Spectrum tmpSpec;
vector<int> v;
sp.clear();
if(pos==0){
if(cs.scan.iLower>0) {
if(!NewScanAverage(sp,cs.inFile,(int)(cs.boxcar/2),0.1f,cs.scan.iLower)) return false;
} else {
NewScanAverage(sp,cs.inFile,(int)(cs.boxcar/2),0.1f);
}
pos=1;
} else {
NewScanAverage(sp,NULL,(int)(cs.boxcar/2),0.1f);
}
if(sp.getScanNumber()==0) return false;
FirstDerivativePeaks(sp,1);
return true;
}
bool CNoiseReduction::NewScanAverage(Spectrum& sp, char* file, int width, float cutoff, int scanNum){
Spectrum ts;
vector<int> vPos;
int i;
int j;
int k;
int m;
int widthCount=0;
int numScans=1;
double dif;
double prec;
double dt;
double c;
bool bLeft=true;
int posLeft;
int posRight;
int index;
char cFilter1[256];
//char cFilter2[256];
//double slope;
//double intercept;
sp.clear();
Spectrum* specs;
specs = new Spectrum[width*2+1];
//if file is not null, create new buffer
if(file!=NULL){
strcpy(lastFile,file);
bs.clear();
if(scanNum>0) r->readFile(file,ts,scanNum);
else r->readFile(file,ts);
if(ts.getScanNumber()==0) {
delete [] specs;
return false;
}
bs.push_back(ts);
specs[0]=bs[0];
c=CParam(specs[0],3);
posA=0;
} else {
posA++;
if(posA>=(int)bs.size()) { //end of buffer, no more data
delete [] specs;
return false;
}
specs[0]=bs[posA];
c=CParam(specs[0],3);
}
specs[0].getRawFilter(cFilter1,256);
posLeft=posA;
posRight=posA;
while(widthCount<(width*2)){
index=-1;
//Alternate looking left and right
if(bLeft){
bLeft=false;
widthCount++;
while(true){
posLeft--;
if(posLeft<0) { //buffer is too short on left, add spectra
i=bs[0].getScanNumber();
while(true){
i--;
if(i==0) break;
r->readFile(lastFile,ts,i);
if(ts.getScanNumber()==0) continue;
else break;
}
if(i==0) break;
bs.push_front(ts);
posA++;
posRight++;
posLeft=0;
//ts.getRawFilter(cFilter2,256);
if(ts.getMsLevel()==cs.msLevel) {
index=posLeft;
break;
}
} else {
//bs[posLeft].getRawFilter(cFilter2,256);
if(bs[posLeft].getMsLevel()==cs.msLevel) {
index=posLeft;
break;
}
}
}
} else {
bLeft=true;
widthCount++;
while(true){
posRight++;
if(posRight>=(int)bs.size()) { //buffer is too short on right, add spectra
r->readFile(lastFile,ts,bs[bs.size()-1].getScanNumber());
r->readFile(NULL,ts);
if(ts.getScanNumber()==0) {
posRight--;
break;
}
bs.push_back(ts);
//ts.getRawFilter(cFilter2,256);
if(ts.getMsLevel()==cs.msLevel) {
index=posRight;
break;
}
} else {
//bs[posRight].getRawFilter(cFilter2,256);
if(bs[posRight].getMsLevel()==cs.msLevel) {
index=posRight;
break;
}
}
}
}
if(index==-1) continue;
specs[numScans++]=bs[index];
}
double tmz;
int mzcount=0;
/* Ledford equation correction
double freq;
double conA=0.0;
double conB=0.0;
for(m=0;m<numScans;m++){
conA+=specs[m].getConversionA();
conB+=specs[m].getConversionB();
}
conA/=numScans;
conB/=numScans;
printf("%.10lf, %.10lf\n",conA,conB);
for(k=0;k<numScans;k++){
for(j=0;j<specs[k].size();j++){
freq = sqrt(specs[k].getConversionA()*specs[k].getConversionA() - (4*specs[k].at(j).mz*(-specs[k].getConversionB())));
freq += specs[k].getConversionA();
freq /= (2*specs[k].at(j).mz);
specs[k].at(j).mz = conA/freq + conB/(freq*freq);
}
}
*/
//Match peaks between pivot scan (0) and neighbors (the rest)
for(m=0;m<numScans;m++){
vPos.clear();
for(i=0;i<numScans;i++) vPos.push_back(0);
for(i=0;i<specs[m].size();i++){ //iterate all points
if(specs[m].at(i).intensity<0.1) continue;
tmz=specs[m].at(i).mz;
mzcount=1;
prec = c * tmz * tmz / 2;
for(k=m+1;k<numScans;k++){ //iterate all neighbors
dif=100000.0;
for(j=vPos[k];j<specs[k].size();j++){ //check if point is a match
if(specs[k].at(j).intensity<0.1) continue; //skip meaningless datapoints to speed along
dt=fabs(tmz-specs[k].at(j).mz);
if(dt<=dif) {
if(dt<prec) {
//linear interpolate
//if(specs[k].at(j).mz<tmz && j<specs[k].size()-1){
// slope=(specs[k].at(j+1).intensity-specs[k].at(j).intensity)/(specs[k].at(j+1).mz-specs[k].at(j).mz);
// intercept=specs[k].at(j).intensity-specs[k].at(j).mz*slope;
// specs[m].at(i).intensity+=(tmz*slope+intercept);
//} else if(specs[k].at(j).mz>tmz && j>0) {
// slope=(specs[k].at(j).intensity-specs[k].at(j-1).intensity)/(specs[k].at(j).mz-specs[k].at(j-1).mz);
// intercept=specs[k].at(j).intensity-specs[k].at(j).mz*slope;
// specs[m].at(i).intensity+=(tmz*slope+intercept);
//} else {
specs[m].at(i).intensity += specs[k].at(j).intensity;
//}
//Averaging the mz values appears equivalent to realigning all spectra against
//an average Ledford correction.
specs[m].at(i).mz += specs[k].at(j).mz;
vPos[k]=j+1;
specs[k].at(j).intensity=-1.0;
mzcount++;
break;
}
dif=dt;
} else {
vPos[k]=j-1;
break;
}
}
}//for k
sp.add(specs[m].at(i).mz/mzcount,specs[m].at(i).intensity/numScans);
} //next i
} //next m
if(sp.size()>0) sp.sortMZ();
sp.setScanNumber(specs[0].getScanNumber());
sp.setScanNumber(specs[0].getScanNumber(true),true);
sp.setRTime(specs[0].getRTime());
sp.setRawFilter(cFilter1);
if(posLeft>0){
while(posLeft>0){
bs.pop_front();
posLeft--;
posA--;
}
}
delete [] specs;
return true;
}
bool CNoiseReduction::ScanAverage(Spectrum& sp, char* file, int width, float cutoff){
Spectrum ts;
Spectrum ps=sp;
MSReader r;
int i;
int j;
int k;
int widthCount=0;
int numScans=1;
double dif;
double prec;
double dt;
double c=CParam(ps,3);
bool bLeft=true;
int posLeft=ps.getScanNumber()-1;
int posRight=ps.getScanNumber()+1;
char cFilter1[256];
char cFilter2[256];
ps.getRawFilter(cFilter1,256);
while(widthCount<(width*2)){
//Alternate looking left and right
if(bLeft){
bLeft=false;
widthCount++;
while(posLeft>0){
r.readFile(file,ts,posLeft);
if(ts.getScanNumber()==0) break;
ts.getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) break;
posLeft--;
}
} else {
bLeft=true;
widthCount++;
while(true){
r.readFile(file,ts,posRight);
if(ts.getScanNumber()==0) break;
ts.getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) break;
posRight++;
}
}
if(ts.getScanNumber()==0) continue;
numScans++;
//Match peaks between pivot scan and temp scan
k=0;
for(i=0;i<ps.size();i++){
dif=100000.0;
prec = c * ps.at(i).mz * ps.at(i).mz / 2;
for(j=k;j<ts.size();j++){
dt=fabs(ps.at(i).mz-ts.at(j).mz);
if(dt<=dif) {
if(dt<prec) {
ps.at(i).intensity+=ts.at(j).intensity;
ts.at(j).mz=-1.0;
k=j+1;
break;
}
dif=dt;
} else {
k=j-1;
break;
}
}
}
//Add unmatched peaks from temp scan
for(i=0;i<ts.size();i++){
if(ts.at(i).mz>-1.0) ps.add(ts.at(i));
}
//Sort pivot scan peaks for fast traversal against next temp scan
ps.sortMZ();
}
//Average points and apply cutoff
sp.clear();
for(i=0;i<ps.size();i++) {
ps.at(i).intensity/=numScans;
if(ps.at(i).intensity>=cutoff) sp.add(ps.at(i));
}
sp.setScanNumber(ps.getScanNumber());
sp.setScanNumber(ps.getScanNumber(true),true);
sp.setRTime(ps.getRTime());
return true;
}
bool CNoiseReduction::DeNoise(Spectrum& sp){
double ppm;
int i,j,k;
int index;
int matchCount;
char cFilter1[256];
char cFilter2[256];
Spectrum tmpSpec;
vector<int> v;
sp.clear();
if(pos==0){
if((cs.scan.iLower>0)) {
k=cs.scan.iLower;
r->readFile(&cs.inFile[0],tmpSpec,cs.scan.iLower);
} else {
r->readFile(&cs.inFile[0],tmpSpec);
k=tmpSpec.getScanNumber();
}
if(tmpSpec.getScanNumber()==0) return false;
tmpSpec.getRawFilter(cFilter1,256);
//Gather left side of scan
i=1;
j=0;
while( k-i > 0){
r->readFile(&cs.inFile[0],tmpSpec,cs.scan.iLower-i);
if(tmpSpec.getScanNumber()==0) {
i++;
continue;
}
tmpSpec.getRawFilter(cFilter2,256);
//Assume High resolution data at all times
if(!cs.centroid) {
FirstDerivativePeaks(tmpSpec,1);
tmpSpec.setRawFilter(cFilter2);
}
s.push_front(tmpSpec);
if(strcmp(cFilter1,cFilter2)==0){
j++;
if(j==(int)(cs.boxcar/2)) break;
}
i++;
}
//cout << "Done left " << s.size() << " " << cs.rawAvgWidth << endl;
//Get our position again
r->readFile(&cs.inFile[0],tmpSpec,k);
//Assume High resolution data at all times
if(!cs.centroid) {
FirstDerivativePeaks(tmpSpec,1);
tmpSpec.setRawFilter(cFilter1);
}
//Add first target scan
s.push_back(tmpSpec);
pos=s.size()-1;
//Add right side of scan
i=1;
j=0;
while(true){
r->readFile(NULL,tmpSpec);
if(tmpSpec.getScanNumber()==0) break;
tmpSpec.getRawFilter(cFilter2,256);
//Assume High resolution data at all times
if(!cs.centroid) {
FirstDerivativePeaks(tmpSpec,1);
tmpSpec.setRawFilter(cFilter2);
}
s.push_back(tmpSpec);
//cout << tmpSpec.getScanNumber() << " " << cFilter1 << " xx " << cFilter2 << endl;
if(strcmp(cFilter1,cFilter2)==0){
j++;
if(j==(int)(cs.boxcar/2)) break;
}
i++;
}
}
//return false if we reached the end
if(pos>=(int)s.size()) return false;
//Because Deque may have mixed spectra, create list of Deque indexes to compare
//Shrink Deque on the left if some spectra are not needed
//Expand Deque to the right if needed and possible
//cout << "Checking scan " << s[pos].getScanNumber() << endl;
v.clear();
s[pos].getRawFilter(cFilter1,256);
//look left
for(i=pos-1;i>=0;i--){
s[i].getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
v.push_back(i);
if(v.size()==(int)(cs.boxcar/2)) break;
}
}
//erase unneeded left items
while(i>0){
s.pop_front();
i--;
for(j=0;j<(int)v.size();j++) v[j]--;
pos--;
}
//look right
j=0;
for(i=pos+1;i<(int)s.size();i++){
s[i].getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
v.push_back(i);
j++;
if(j==(int)(cs.boxcar/2)) break;
}
}
//extend right side if needed
while(j<(int)(cs.boxcar/2)){
r->readFile(NULL,tmpSpec);
if(tmpSpec.getScanNumber()==0) break;
tmpSpec.getRawFilter(cFilter2,256);
//Assume High resolution data at all times
if(!cs.centroid) {
FirstDerivativePeaks(tmpSpec,1);
tmpSpec.setRawFilter(cFilter2);
}
s.push_back(tmpSpec);
if(strcmp(cFilter1,cFilter2)==0) {
v.push_back(s.size()-1);
j++;
}
}
//if there is nothing to compare to, exit now
if(v.size()==0) return false;
//compare peaks
//cout << "Checking " << s[pos].size() << " peaks." << endl;
for(i=0;i<s[pos].size();i++){
matchCount=1;
for(j=0;j<(int)v.size();j++){
if(s[v[j]].size()<1) continue;
index = NearestPeak(s[v[j]],s[pos].at(i).mz);
ppm=fabs( (s[v[j]].at(index).mz-s[pos].at(i).mz)/s[pos].at(i).mz * 1000000);
if(ppm<cs.ppm) matchCount++;
}
if(matchCount>=cs.boxcarFilter || matchCount==v.size()) sp.add(s[pos].at(i));
}
sp.setScanNumber(s[pos].getScanNumber());
sp.setScanNumber(s[pos].getScanNumber(true),true);
sp.setRTime(s[pos].getRTime());
pos++;
//cout << sp.getScanNumber() << endl;
//cout << "Done! " << sp.size() << " " << pos << endl;
return true;
}
//First derivative method taken from CSpecAnalyze, returns base peak intensity of the set
void CNoiseReduction::FirstDerivativePeaks(Spectrum& sp, int winSize){
int i,j;
float maxIntensity;
int bestPeak;
bool bLastPos;
Spectrum gp;
int nextBest;
double FWHM;
Peak_T centroid;
bLastPos=false;
for(i=0;i<sp.size()-winSize;i++){
if(sp.at(i).intensity<sp.at(i+winSize).intensity) {
bLastPos=true;
continue;
} else {
if(bLastPos){
bLastPos=false;
//find max and add peak
maxIntensity=0;
for(j=i;j<i+winSize;j++){
if (sp.at(j).intensity>maxIntensity){
maxIntensity=sp.at(j).intensity;
bestPeak = j;
}
}
//Best estimate of Gaussian centroid
//Get 2nd highest point of peak
if(bestPeak==sp.size()-1){
nextBest=bestPeak-1;
} else if(sp.at(bestPeak-1).intensity > sp.at(bestPeak+1).intensity){
nextBest=bestPeak-1;
} else {
nextBest=bestPeak+1;
}
//Get FWHM
FWHM = calcFWHM(sp.at(bestPeak).mz);
//Calc centroid MZ (in three lines for easy reading)
centroid.mz = pow(FWHM,2)*log(sp.at(bestPeak).intensity/sp.at(nextBest).intensity);
centroid.mz /= GC*(sp.at(bestPeak).mz-sp.at(nextBest).mz);
centroid.mz += (sp.at(bestPeak).mz+sp.at(nextBest).mz)/2;
//Calc centroid intensity
centroid.intensity=(float)(sp.at(bestPeak).intensity/exp(-pow((sp.at(bestPeak).mz-centroid.mz)/FWHM,2)*GC));
//some peaks are funny shaped and have bad gaussian fit.
//if error is more than 10%, keep existing intensity
if( fabs((sp.at(bestPeak).intensity - centroid.intensity) / centroid.intensity * 100) > 10 ||
//not a good check for infinity
centroid.intensity>999999999999.9 ||
centroid.intensity < 0 ) {
centroid.intensity=sp.at(bestPeak).intensity;
}
//Hack until I put in mass ranges
if(centroid.mz<0 || centroid.mz>2000) {
//do nothing if invalid mz
} else {
gp.add(centroid);
}
i+=winSize-1;
}
}
}
int scanNumber=sp.getScanNumber();
int scanNumber2=sp.getScanNumber(false);
float rTime=sp.getRTime();
sp = gp;
sp.setRTime(rTime);
sp.setScanNumber(scanNumber);
sp.setScanNumber(scanNumber2,true);
}
bool CNoiseReduction::NewScanAveragePlusDeNoise(Spectrum& sp, char* file, int width, float cutoff, int scanNum){
Spectrum ts;
vector<int> vPos;
int i;
int j;
int k;
int m;
int widthCount=0;
int numScans=1;
int match;
double dif;
double prec;
double dt;
double c;
bool bLeft=true;
int posLeft;
int posRight;
int index;
char cFilter1[256];
char cFilter2[256];
sp.clear();
Spectrum* specs;
specs = new Spectrum[width*2+1];
//if file is not null, create new buffer
if(file!=NULL){
strcpy(lastFile,file);
bs.clear();
if(scanNum>0) r->readFile(file,ts,scanNum);
else r->readFile(file,ts);
if(ts.getScanNumber()==0) {
delete [] specs;
return false;
}
bs.push_back(ts);
specs[0]=bs[0];
c=CParam(specs[0],3);
posA=0;
} else {
posA++;
if(posA>=(int)bs.size()) { //end of buffer, no more data
delete [] specs;
return false;
}
specs[0]=bs[posA];
c=CParam(specs[0],3);
}
//set our pivot spectrum
specs[0].getRawFilter(cFilter1,256);
posLeft=posA;
posRight=posA;
while(widthCount<(width*2)){
index=-1;
//Alternate looking left and right
if(bLeft){
bLeft=false;
widthCount++;
while(true){
posLeft--;
if(posLeft<0) { //buffer is too short on left, add spectra
i=bs[0].getScanNumber();
while(true){
i--;
if(i==0) break;
r->readFile(lastFile,ts,i);
if(ts.getScanNumber()==0) continue;
else break;
}
if(i==0) break;
bs.push_front(ts);
posA++;
posRight++;
posLeft=0;
ts.getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
index=posLeft;
break;
}
} else {
bs[posLeft].getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
index=posLeft;
break;
}
}
}
} else {
bLeft=true;
widthCount++;
while(true){
posRight++;
if(posRight>=(int)bs.size()) { //buffer is too short on right, add spectra
r->readFile(lastFile,ts,bs[bs.size()-1].getScanNumber());
r->readFile(NULL,ts);
if(ts.getScanNumber()==0) {
posRight--;
break;
}
bs.push_back(ts);
ts.getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
index=posRight;
break;
}
} else {
bs[posRight].getRawFilter(cFilter2,256);
if(strcmp(cFilter1,cFilter2)==0) {
index=posRight;
break;
}
}
}
}
if(index==-1) continue;
//ts=bs[index];
specs[numScans++]=bs[index];
//cout << "NumScans: " << numScans << endl;
}
//Match peaks between pivot scan (0) and neighbors (the rest)
//for(m=0;m<cs.ppMatch && m<numScans;m++){
for(m=0;m<1;m++){
//cout << "m " << m << " = " << specs[m].getScanNumber() << endl;
vPos.clear();
for(i=0;i<numScans;i++) vPos.push_back(0);
//cout << "Checking " << m << " of " << numScans << " points remaining: " << specs[m].size() << endl;
for(i=0;i<specs[m].size();i++){ //iterate all points
if(specs[m].at(i).intensity<0.1) continue;
prec = c * specs[m].at(i).mz * specs[m].at(i).mz / 2;
match=1;
for(k=m+1;k<numScans;k++){ //iterate all neighbors
dif=100000.0;
for(j=vPos[k];j<specs[k].size();j++){ //check if point is a match
//cout << "Checking " << j << " of " << specs[k].size() << endl;
if(specs[k].at(j).intensity<0.1) continue; //skip meaningless datapoints to speed along
dt=fabs(specs[m].at(i).mz-specs[k].at(j).mz);
//dt=specs[m].at(i).mz-specs[k].at(j).mz;
//if(dt<0.0)dt=-dt;
if(dt<=dif) {
if(dt<prec) {
specs[m].at(i).intensity+=specs[k].at(j).intensity;
vPos[k]=j+1;
specs[k].at(j).intensity=-1.0;
match++;
break;
}
dif=dt;
} else {
vPos[k]=j-1;
break;
}
}
}//for k
//if data point was not visible across enough scans, set it to 0
if(match<cs.boxcarFilter && match<numScans) {
//ignore point
//cout << "BAH! " << specs[m].at(i).mz << " has " << match << " matches." << endl;
} else {
//add to temp spectrum
//cout << specs[m].at(i).mz << " has " << match << " matches." << endl;
//sp.add(specs[m].at(i).mz,specs[m].at(i).intensity/numScans);
sp.add(specs[m].at(i).mz,specs[m].at(i).intensity/match);
}
} //next i
} //next m
//sort
//cout << "Done " << sp.size() << endl;
if(sp.size()>0) sp.sortMZ();
sp.setScanNumber(specs[0].getScanNumber());
sp.setScanNumber(specs[0].getScanNumber(true),true);
sp.setRTime(specs[0].getRTime());
sp.setRawFilter(cFilter1);
//clear unused buffer
if(posLeft>0){
while(posLeft>0){
bs.pop_front();
posLeft--;
posA--;
}
}
delete [] specs;
return true;
}
bool CNoiseReduction::ScanAveragePlusDeNoise(Spectrum& sp, char* file, int width, float cutoff, int scanNum){
Spectrum ts;
Spectrum ps;
//MSReader r;
vector<int> v;
vector<int> vPos;
int i;
int j;
int k;
int widthCount=0;
int numScans=1;
int match;
double dif;
double prec;
double dt;
double c;
bool bLeft=true;
int posLeft=ps.getScanNumber()-1;
int posRight=ps.getScanNumber()+1;
int index;
char cFilter1[256];
//char cFilter2[256];
sp.clear();
//if file is not null, create new buffer
if(file!=NULL){
strcpy(lastFile,file);
bs.clear();
if(scanNum>0) r->readFile(file,ts,scanNum);
else r->readFile(file,ts);
if(ts.getScanNumber()==0) return false;
bs.push_back(ts);
ps=bs[0];
c=CParam(ps,3);
posA=0;
} else {
posA++;
//cout << "ER: " << posA << " " << bs.size() << endl;
if(posA>=(int)bs.size()) return false; //end of buffer, no more data
ps=bs[posA];
c=CParam(ps,3);
}
//set our pivot spectrum
//ps=bs[posA];
ps.getRawFilter(cFilter1,256);
//cout << "Averaging: " << ps.getScanNumber() << endl;
posLeft=posA;
posRight=posA;
while(widthCount<(width*2)){
index=-1;
//Alternate looking left and right
if(bLeft){
bLeft=false;
widthCount++;
while(true){
posLeft--;
//cout << posLeft << endl;
if(posLeft<0) { //buffer is too short on left, add spectra
i=bs[0].getScanNumber();
while(true){
i--;
//cout << "I: " << i << endl;
if(i==0) break;
r->readFile(lastFile,ts,i);
if(ts.getScanNumber()==0) continue;
else break;
}
if(i==0) break;
bs.push_front(ts);
for(i=0;i<(int)v.size();i++)v[i]++;
posA++;
posRight++;
posLeft=0;
//ts.getRawFilter(cFilter2,256);
if(ts.getMsLevel()==cs.msLevel) {
index=posLeft;
break;
}
} else {
//bs[posLeft].getRawFilter(cFilter2,256);
if(bs[posLeft].getMsLevel()==cs.msLevel) {
index=posLeft;
break;
}
}
}
} else {
bLeft=true;
widthCount++;
while(true){
posRight++;
if(posRight>=(int)bs.size()) { //buffer is too short on right, add spectra
r->readFile(lastFile,ts,bs[bs.size()-1].getScanNumber());
r->readFile(NULL,ts);
if(ts.getScanNumber()==0) {
posRight--;
break;
}
bs.push_back(ts);
//ts.getRawFilter(cFilter2,256);
if(ts.getMsLevel()==cs.msLevel) {
index=posRight;
break;
}
} else {
//bs[posRight].getRawFilter(cFilter2,256);
if(bs[posRight].getMsLevel()==cs.msLevel) {
index=posRight;
break;
}
}
}
}
if(index==-1) continue;
//ts=bs[index];
v.push_back(index);
numScans++;
}
//cout << "Still Averaging: " << ps.getScanNumber() << endl;
//cout << " with: ";
//for(i=0;i<v.size();i++) cout << bs[v[i]].getScanNumber() << " ";
//cout << endl;
//cout << numScans << " " << v.size() << endl;
//Match peaks between pivot scan and neighbors
for(i=0;i<(int)v.size();i++) vPos.push_back(0);
for(i=0;i<(int)ps.size();i++){ //iterate all points
prec = c * ps.at(i).mz * ps.at(i).mz / 2;
match=1;
for(k=0;k<(int)v.size();k++){ //iterate all neighbors
dif=100000.0;
//cout << "Checking " << bs[v[k]].getScanNumber() << " pos " << vPos[k] << endl;
for(j=vPos[k];j<bs[v[k]].size();j++){ //check if point is a match
dt=fabs(ps.at(i).mz-bs[v[k]].at(j).mz);
if(dt<=dif) {
if(dt<prec) {
ps.at(i).intensity+=bs[v[k]].at(j).intensity;
vPos[k]=j+1;
match++;
break;
}
dif=dt;
} else {
vPos[k]=j-1;
break;
}
}
}
//if data point was not visible across enough scans, set it to 0
if(match<cs.boxcarFilter && match<(int)v.size()) ps.at(i).intensity=0.0;
}
//Average points and apply cutoff
for(i=0;i<ps.size();i++) {
ps.at(i).intensity/=numScans;
sp.add(ps.at(i));
//if(ps.at(i).intensity>=cutoff) sp.add(ps.at(i));
}
sp.setScanNumber(ps.getScanNumber());
sp.setScanNumber(ps.getScanNumber(true),true);
sp.setRTime(ps.getRTime());
sp.setRawFilter(cFilter1);
//clear unused buffer
if(posLeft>0){
while(posLeft>0){
bs.pop_front();
posLeft--;
posA--;
}
}
//cout << "Done averaging" << endl;
return true;
}
// Function reads in the spectra from the data stored for each PSMClass object
void PepXMLClass::readInSpectra() {
string curSpectrumFilePath;
deque<PSMClass>::iterator curPSM;
SpecStruct *spec = NULL;
bool status;
string spectrumFileName;
int scanNum;
int ctr = 0;
int N = (signed) PSMvec->size();
if(g_ext == "mgf") {
parseMGF();
return;
}
// mstoolkit to read MS2 spectra
MSReader *reader = new MSReader();
reader->setFilter(MS2);
// mstoolkit spectrum object
Spectrum *S = NULL;
// Extract from each PSM, it's parent spectrum file and scan number
for(curPSM = PSMvec->begin(); curPSM != PSMvec->end(); curPSM++) {
spectrumFileName = curPSM->getSpectrumFileName();
scanNum = curPSM->getScanNumber();
// code to get the full path to the input spectrum file
filesystem::path curFile( spectrumFileName.c_str() );
filesystem::path spectral_dir( g_srcDir.c_str() );
filesystem::path curFilePath( spectral_dir/curFile );
curSpectrumFilePath = curFilePath.file_string();
// if the source file is not located, drop the spectrum
if( !boost::filesystem::exists(curSpectrumFilePath) ) {
curPSM = PSMvec->erase(curPSM);
continue;
}
// Read the spectrum for 'scanNum' into 'S'
S = new Spectrum();
status = reader->readFile(curSpectrumFilePath.c_str(), *S, scanNum);
if(S->getScanNumber() == 0) {
cerr << "Failed to get " << curSpectrumFilePath.c_str() << " scan#: " << scanNum << endl;
exit(0);
}
spec = new SpecStruct();
for(int j = 0; j < S->size(); j++) {
spec->mz.push_back( S->at(j).mz );
spec->intensity.push_back( S->at(j).intensity );
}
delete(S); S = NULL;
// Assign the data in 'S' to curPSM
curPSM->recordSpectrum(*spec);
delete(spec); spec = NULL;
ctr++;
printProgress("Reading in spectra (please be patient)...", ctr, N);
}
delete(reader); reader = NULL;
cerr << endl; // prettier stderr
}