X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=pintail.cpp;h=b9f2434d19d4cc842bfac1dce4f6de5a0d1b742d;hp=39ac13f786ad3c6e94d4b1a8252f126a23200522;hb=a8e2df1b96a57f5f29576b08361b86a96a8eff4f;hpb=fe346922fe0af5b1a025beacb211078d37598fd4 diff --git a/pintail.cpp b/pintail.cpp index 39ac13f..b9f2434 100644 --- a/pintail.cpp +++ b/pintail.cpp @@ -9,106 +9,72 @@ #include "pintail.h" #include "ignoregaps.h" +#include "eachgapdist.h" +//******************************************************************************************************************** +//sorts lowest to highest +inline bool compareQuanMembers(quanMember left, quanMember right){ + return (left.score < right.score); +} //*************************************************************************************************************** -Pintail::Pintail(string filename, string temp) { fastafile = filename; templateFile = temp; } -//*************************************************************************************************************** - -Pintail::~Pintail() { +Pintail::Pintail(string filename, string temp, bool f, int p, string mask, string cons, string q, int win, int inc, string o) : Chimera() { try { - for (int i = 0; i < querySeqs.size(); i++) { delete querySeqs[i]; } - for (int i = 0; i < templateSeqs.size(); i++) { delete templateSeqs[i]; } + + fastafile = filename; + templateFileName = temp; templateSeqs = readSeqs(temp); + filter = f; + processors = p; + setMask(mask); + consfile = cons; + quanfile = q; + window = win; + increment = inc; + outputDir = o; + + distcalculator = new eachGapDist(); + decalc = new DeCalculator(); - if (processors != 1) { for (int i = 0; i < bestfit.size(); i++) { delete bestfit[i]; } } + doPrep(); } catch(exception& e) { - errorOut(e, "Pintail", "~Pintail"); + m->errorOut(e, "Pintail", "Pintail"); exit(1); } -} + +} //*************************************************************************************************************** -void Pintail::print(ostream& out) { + +Pintail::~Pintail() { try { - for (int i = 0; i < querySeqs.size(); i++) { - - int index = ceil(deviation[i]); -float quan = 2.64 * log10(deviation[i]) + 1.46; -cout << "dist = " << index << endl; -cout << "de = " << DE[i] << endl; -cout << "mallard quantile = " << quan << endl; -cout << "my quantile = " << quantiles[index][4] << endl; - - //is your DE value higher than the 95% - string chimera; - if (DE[i] > quantiles[index][4]) { chimera = "Yes"; } - else { chimera = "No"; } - - out << querySeqs[i]->getName() << '\t' << "div: " << deviation[i] << "\tstDev: " << DE[i] << "\tchimera flag: " << chimera << endl; - if (chimera == "Yes") { - mothurOut(querySeqs[i]->getName() + "\tdiv: " + toString(deviation[i]) + "\tstDev: " + toString(DE[i]) + "\tchimera flag: " + chimera); mothurOutEndLine(); - } - out << "Observed\t"; - - for (int j = 0; j < obsDistance[i].size(); j++) { out << obsDistance[i][j] << '\t'; } - out << endl; - - out << "Expected\t"; - - for (int m = 0; m < expectedDistance[i].size(); m++) { out << expectedDistance[i][m] << '\t'; } - out << endl; - - } + delete distcalculator; + delete decalc; } catch(exception& e) { - errorOut(e, "Pintail", "print"); + m->errorOut(e, "Pintail", "~Pintail"); exit(1); } } - //*************************************************************************************************************** -void Pintail::getChimeras() { +int Pintail::doPrep() { try { - //read in query sequences and subject sequences - mothurOut("Reading sequences and template file... "); cout.flush(); - querySeqs = readSeqs(fastafile); - templateSeqs = readSeqs(templateFile); - mothurOut("Done."); mothurOutEndLine(); - - int numSeqs = querySeqs.size(); - - obsDistance.resize(numSeqs); - expectedDistance.resize(numSeqs); - seqCoef.resize(numSeqs); - DE.resize(numSeqs); - Qav.resize(numSeqs); - bestfit.resize(numSeqs); - deviation.resize(numSeqs); - trimmed.resize(numSeqs); - windowSizes.resize(numSeqs, window); + mergedFilterString = ""; windowSizesTemplate.resize(templateSeqs.size(), window); - windowsForeachQuery.resize(numSeqs); - h.resize(numSeqs); quantiles.resize(100); //one for every percent mismatch + quantilesMembers.resize(100); //one for every percent mismatch - //break up file if needed - int linesPerProcess = numSeqs / processors ; + //if the user does not enter a mask then you want to keep all the spots in the alignment + if (seqMask.length() == 0) { decalc->setAlignmentLength(templateSeqs[0]->getAligned().length()); } + else { decalc->setAlignmentLength(seqMask.length()); } - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - //find breakup of sequences for all times we will Parallelize - if (processors == 1) { lines.push_back(new linePair(0, numSeqs)); } - else { - //fill line pairs - for (int i = 0; i < (processors-1); i++) { - lines.push_back(new linePair((i*linesPerProcess), ((i*linesPerProcess) + linesPerProcess))); - } - //this is necessary to get remainder of processors / numSeqs so you don't miss any lines at the end - int i = processors - 1; - lines.push_back(new linePair((i*linesPerProcess), numSeqs)); - } - + decalc->setMask(seqMask); + + #ifdef USE_MPI + //do nothing + #else + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) //find breakup of templatefile for quantiles if (processors == 1) { templateLines.push_back(new linePair(0, templateSeqs.size())); } else { @@ -119,577 +85,454 @@ void Pintail::getChimeras() { } } #else - lines.push_back(new linePair(0, numSeqs)); templateLines.push_back(new linePair(0, templateSeqs.size())); #endif + #endif - distcalculator = new ignoreGaps(); - decalc = new DeCalculator(); - - //if the user does enter a mask then you want to keep all the spots in the alignment - if (seqMask.length() == 0) { decalc->setAlignmentLength(querySeqs[0]->getAligned().length()); } - else { decalc->setAlignmentLength(seqMask.length()); } - - decalc->setMask(seqMask); - - //find pairs - if (processors == 1) { - mothurOut("Finding closest sequence in template to each sequence... "); cout.flush(); - bestfit = findPairs(lines[0]->start, lines[0]->end); - mothurOut("Done."); mothurOutEndLine(); - }else { createProcessesPairs(); } - - - for (int j = 0; j < bestfit.size(); j++) { - //chops off beginning and end of sequences so they both start and end with a base - ofstream out; - string s = querySeqs[j]->getName(); - - openOutputFile(s, out); - out << ">" << querySeqs[j]->getName() << endl; - out << querySeqs[j]->getAligned() << endl; - out.close(); - - string t =querySeqs[j]->getName() + ".ref"; - openOutputFile(t, out); - out << ">" << bestfit[j]->getName() << endl; - out << bestfit[j]->getAligned() << endl; - out.close(); - } - - - //find P - mothurOut("Getting conservation... "); cout.flush(); + m->mothurOut("Getting conservation... "); cout.flush(); if (consfile == "") { - mothurOut("Calculating probability of conservation for your template sequences. This can take a while... I will output the frequency of the highest base in each position to a .freq file so that you can input them using the conservation parameter next time you run this command. Providing the .freq file will improve speed. "); cout.flush(); - probabilityProfile = decalc->calcFreq(templateSeqs, templateFile); - mothurOut("Done."); mothurOutEndLine(); - }else { probabilityProfile = readFreq(); } - - //make P into Q - for (int i = 0; i < probabilityProfile.size(); i++) { probabilityProfile[i] = 1 - probabilityProfile[i]; } //cout << i << '\t' << probabilityProfile[i] << endl; - mothurOut("Done."); mothurOutEndLine(); - - //mask querys - for (int i = 0; i < querySeqs.size(); i++) { - //cout << querySeqs[i]->getName() << " before mask = " << querySeqs[i]->getAligned() << endl << endl; - decalc->runMask(querySeqs[i]); - //cout << querySeqs[i]->getName() << " after mask = " << querySeqs[i]->getAligned() << endl << endl; - } - - //mask templates - for (int i = 0; i < templateSeqs.size(); i++) { - decalc->runMask(templateSeqs[i]); - } + m->mothurOut("Calculating probability of conservation for your template sequences. This can take a while... I will output the frequency of the highest base in each position to a .freq file so that you can input them using the conservation parameter next time you run this command. Providing the .freq file will improve speed. "); cout.flush(); + probabilityProfile = decalc->calcFreq(templateSeqs, templateFileName); + if (m->control_pressed) { return 0; } + m->mothurOut("Done."); m->mothurOutEndLine(); + }else { probabilityProfile = readFreq(); m->mothurOut("Done."); } + m->mothurOutEndLine(); -//for (int i = 0; i < lines.size(); i++) { cout << "line pair " << i << " = " << lines[i]->start << '\t' << lines[i]->end << endl; } - - if (processors == 1) { - - for (int j = 0; j < bestfit.size(); j++) { - //cout << querySeqs[j]->getName() << " after mask = " << querySeqs[j]->getAligned() << endl << endl; - ///cout << bestfit[j]->getName() << " after mask = " << bestfit[j]->getAligned() << endl << endl; - decalc->trimSeqs(querySeqs[j], bestfit[j], trimmed[j]); - } - - mothurOut("Finding window breaks... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - it = trimmed[i].begin(); -//cout << querySeqs[i]->getName() << '\t' << "trimmed = " << it->first << '\t' << it->second << endl; - vector win = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment); - windowsForeachQuery[i] = win; - } - mothurOut("Done."); mothurOutEndLine(); - - }else { createProcessesSpots(); } - + //make P into Q + for (int i = 0; i < probabilityProfile.size(); i++) { probabilityProfile[i] = 1 - probabilityProfile[i]; } // - if (processors == 1) { + bool reRead = false; + //create filter if needed for later + if (filter) { - mothurOut("Calculating observed distance... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - //cout << querySeqs[i]->getName() << '\t' << bestfit[i]->getName() << " windows = " << windowsForeachQuery[i].size() << " size = " << windowSizes[i] << endl; - vector obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windowsForeachQuery[i], windowSizes[i]); + //read in all query seqs + vector tempQuerySeqs = readSeqs(fastafile); + + vector temp; + //merge query seqs and template seqs + temp = templateSeqs; + for (int i = 0; i < tempQuerySeqs.size(); i++) { temp.push_back(tempQuerySeqs[i]); } - for (int j = 0; j < obsi.size(); j++) { - //cout << obsi[j] << '\t'; - } - //cout << endl; - obsDistance[i] = obsi; + if (seqMask != "") { + reRead = true; + //mask templates + for (int i = 0; i < temp.size(); i++) { + if (m->control_pressed) { + for (int i = 0; i < tempQuerySeqs.size(); i++) { delete tempQuerySeqs[i]; } + return 0; + } + decalc->runMask(temp[i]); + } } - mothurOut("Done."); mothurOutEndLine(); - - - mothurOut("Finding variability... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - vector q = decalc->findQav(windowsForeachQuery[i], windowSizes[i], probabilityProfile); - - Qav[i] = q; -//cout << querySeqs[i]->getName() << endl; -for (int j = 0; j < Qav[i].size(); j++) { - //cout << Qav[i][j] << '\t'; -} -//cout << endl << endl; - } - mothurOut("Done."); mothurOutEndLine(); - + mergedFilterString = createFilter(temp, 0.5); - mothurOut("Calculating alpha... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - float alpha = decalc->getCoef(obsDistance[i], Qav[i]); -//cout << querySeqs[i]->getName() << "\tcoef = " << alpha << endl; - seqCoef[i] = alpha; - } - mothurOut("Done."); mothurOutEndLine(); - - - mothurOut("Calculating expected distance... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - vector exp = decalc->calcExpected(Qav[i], seqCoef[i]); - expectedDistance[i] = exp; - } - mothurOut("Done."); mothurOutEndLine(); - - - mothurOut("Finding deviation... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - float de = decalc->calcDE(obsDistance[i], expectedDistance[i]); - DE[i] = de; -//cout << querySeqs[i]->getName() << '\t' << "de value = " << de << endl; - it = trimmed[i].begin(); - float dist = decalc->calcDist(querySeqs[i], bestfit[i], it->first, it->second); -//cout << querySeqs[i]->getName() << '\t' << "dist value = " << dist << endl; - deviation[i] = dist; + if (m->control_pressed) { + for (int i = 0; i < tempQuerySeqs.size(); i++) { delete tempQuerySeqs[i]; } + return 0; } - mothurOut("Done."); mothurOutEndLine(); - } - else { createProcesses(); } + //reread template seqs + for (int i = 0; i < tempQuerySeqs.size(); i++) { delete tempQuerySeqs[i]; } + } //quantiles are used to determine whether the de values found indicate a chimera //if you have to calculate them, its time intensive because you are finding the de and deviation values for each //combination of sequences in the template - if (quanfile != "") { quantiles = readQuantiles(); } - else { + if (quanfile != "") { + quantiles = readQuantiles(); + }else { + if ((!filter) && (seqMask != "")) { //if you didn't filter but you want to mask. if you filtered then you did mask first above. + reRead = true; + //mask templates + for (int i = 0; i < templateSeqs.size(); i++) { + if (m->control_pressed) { return 0; } + decalc->runMask(templateSeqs[i]); + } + } + + if (filter) { + reRead = true; + for (int i = 0; i < templateSeqs.size(); i++) { + if (m->control_pressed) { return 0; } + runFilter(templateSeqs[i]); + } + } - mothurOut("Calculating quantiles for your template. This can take a while... I will output the quantiles to a .quan file that you can input them using the quantiles parameter next time you run this command. Providing the .quan file will dramatically improve speed. "); cout.flush(); + m->mothurOut("Calculating quantiles for your template. This can take a while... I will output the quantiles to a .quan file that you can input them using the quantiles parameter next time you run this command. Providing the .quan file will dramatically improve speed. "); cout.flush(); if (processors == 1) { - quantiles = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size()); + quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size()); }else { createProcessesQuan(); } + + if (m->control_pressed) { return 0; } + string noOutliers, outliers; - decalc->removeObviousOutliers(quantiles); - - ofstream out4; - string o = getRootName(templateFile) + "quan"; + if ((!filter) && (seqMask == "")) { + noOutliers = m->getRootName(m->getSimpleName(templateFileName)) + "pintail.quan"; + }else if ((!filter) && (seqMask != "")) { + noOutliers =m->getRootName(m->getSimpleName(templateFileName)) + "pintail.masked.quan"; + }else if ((filter) && (seqMask != "")) { + noOutliers = m->getRootName(m->getSimpleName(templateFileName)) + "pintail.filtered." + m->getSimpleName(m->getRootName(fastafile)) + "masked.quan"; + }else if ((filter) && (seqMask == "")) { + noOutliers = m->getRootName(m->getSimpleName(templateFileName)) + "pintail.filtered." + m->getSimpleName(m->getRootName(fastafile)) + "quan"; + } + + decalc->removeObviousOutliers(quantilesMembers, templateSeqs.size()); - openOutputFile(o, out4); + if (m->control_pressed) { return 0; } + + string outputString = "#" + m->getVersion() + "\n"; //adjust quantiles - for (int i = 0; i < quantiles.size(); i++) { - if (quantiles[i].size() == 0) { + for (int i = 0; i < quantilesMembers.size(); i++) { + vector temp; + + if (quantilesMembers[i].size() == 0) { //in case this is not a distance found in your template files for (int g = 0; g < 6; g++) { - quantiles[i].push_back(0.0); + temp.push_back(0.0); } }else{ - sort(quantiles[i].begin(), quantiles[i].end()); + sort(quantilesMembers[i].begin(), quantilesMembers[i].end()); - vector temp; //save 10% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.10)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.10)]); //save 25% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.25)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.25)]); //save 50% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.5)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.5)]); //save 75% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.75)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.75)]); //save 95% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.95)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.95)]); //save 99% - temp.push_back(quantiles[i][int(quantiles[i].size() * 0.99)]); + temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.99)]); - quantiles[i] = temp; } //output quan value - out4 << i+1 << '\t'; - for (int u = 0; u < quantiles[i].size(); u++) { out4 << quantiles[i][u] << '\t'; } - out4 << endl; - + outputString += toString(i+1) + "\t"; + for (int u = 0; u < temp.size(); u++) { outputString += toString(temp[u]) + "\t"; } + outputString += "\n"; + + quantiles[i] = temp; + } - mothurOut("Done."); mothurOutEndLine(); + printQuanFile(noOutliers, outputString); + + //free memory + quantilesMembers.clear(); + + m->mothurOut("Done."); m->mothurOutEndLine(); } - + + if (reRead) { + for (int i = 0; i < templateSeqs.size(); i++) { delete templateSeqs[i]; } + templateSeqs.clear(); + templateSeqs = readSeqs(templateFileName); + } + + //free memory - for (int i = 0; i < lines.size(); i++) { delete lines[i]; } - for (int i = 0; i < templateLines.size(); i++) { delete templateLines[i]; } - - delete distcalculator; - delete decalc; + for (int i = 0; i < templateLines.size(); i++) { delete templateLines[i]; } + + return 0; + } catch(exception& e) { - errorOut(e, "Pintail", "getChimeras"); + m->errorOut(e, "Pintail", "doPrep"); exit(1); } } - //*************************************************************************************************************** - -vector Pintail::readFreq() { +Sequence Pintail::print(ostream& out, ostream& outAcc) { try { - - ifstream in; - openInputFile(consfile, in); - - vector prob; - set h = decalc->getPos(); //positions of bases in masking sequence - //read in probabilities and store in vector - int pos; float num; + int index = ceil(deviation); - while(!in.eof()){ - - in >> pos >> num; - - if (h.count(pos) > 0) { - float Pi; - Pi = (num - 0.25) / 0.75; - - //cannot have probability less than 0. - if (Pi < 0) { Pi = 0.0; } - - //do you want this spot - prob.push_back(Pi); + //is your DE value higher than the 95% + string chimera; + if (index != 0) { //if index is 0 then its an exact match to a template seq + if (quantiles[index][4] == 0.0) { + chimera = "Your template does not include sequences that provide quantile values at distance " + toString(index); + }else { + if (DE > quantiles[index][4]) { chimera = "Yes"; } + else { chimera = "No"; } } - - gobble(in); + }else{ chimera = "No"; } + + out << querySeq->getName() << '\t' << "div: " << deviation << "\tstDev: " << DE << "\tchimera flag: " << chimera << endl; + if (chimera == "Yes") { + m->mothurOut(querySeq->getName() + "\tdiv: " + toString(deviation) + "\tstDev: " + toString(DE) + "\tchimera flag: " + chimera); m->mothurOutEndLine(); + outAcc << querySeq->getName() << endl; } + out << "Observed\t"; - in.close(); - return prob; + for (int j = 0; j < obsDistance.size(); j++) { out << obsDistance[j] << '\t'; } + out << endl; + + out << "Expected\t"; + + for (int m = 0; m < expectedDistance.size(); m++) { out << expectedDistance[m] << '\t'; } + out << endl; + + return *querySeq; } catch(exception& e) { - errorOut(e, "Pintail", "readFreq"); + m->errorOut(e, "Pintail", "print"); exit(1); } } - +#ifdef USE_MPI //*************************************************************************************************************** - -vector< vector > Pintail::readQuantiles() { +Sequence Pintail::print(MPI_File& out, MPI_File& outAcc) { try { - - ifstream in; - openInputFile(quanfile, in); - vector< vector > quan; - - int num; float ten, twentyfive, fifty, seventyfive, ninetyfive, ninetynine; - - while(!in.eof()){ - - in >> num >> ten >> twentyfive >> fifty >> seventyfive >> ninetyfive >> ninetynine; - - vector temp; - - temp.push_back(ten); - temp.push_back(twentyfive); - temp.push_back(fifty); - temp.push_back(seventyfive); - temp.push_back(ninetyfive); - temp.push_back(ninetynine); - - quan.push_back(temp); + string outputString = ""; + int index = ceil(deviation); + + //is your DE value higher than the 95% + string chimera; + if (index != 0) { //if index is 0 then its an exact match to a template seq + if (quantiles[index][4] == 0.0) { + chimera = "Your template does not include sequences that provide quantile values at distance " + toString(index); + }else { + if (DE > quantiles[index][4]) { chimera = "Yes"; } + else { chimera = "No"; } + } + }else{ chimera = "No"; } + + outputString += querySeq->getName() + "\tdiv: " + toString(deviation) + "\tstDev: " + toString(DE) + "\tchimera flag: " + chimera + "\n"; + if (chimera == "Yes") { + cout << querySeq->getName() << "\tdiv: " << toString(deviation) << "\tstDev: " << toString(DE) << "\tchimera flag: " << chimera << endl; + string outAccString = querySeq->getName() + "\n"; - gobble(in); + MPI_Status statusAcc; + int length = outAccString.length(); + char* buf = new char[length]; + memcpy(buf, outAccString.c_str(), length); + + MPI_File_write_shared(outAcc, buf, length, MPI_CHAR, &statusAcc); + delete buf; + + return *querySeq; } + outputString += "Observed\t"; - in.close(); - return quan; + for (int j = 0; j < obsDistance.size(); j++) { outputString += toString(obsDistance[j]) + "\t"; } + outputString += "\n"; + + outputString += "Expected\t"; + + for (int m = 0; m < expectedDistance.size(); m++) { outputString += toString(expectedDistance[m]) + "\t"; } + outputString += "\n"; + MPI_Status status; + int length = outputString.length(); + char* buf2 = new char[length]; + memcpy(buf2, outputString.c_str(), length); + + MPI_File_write_shared(out, buf2, length, MPI_CHAR, &status); + delete buf2; + + return *querySeq; } catch(exception& e) { - errorOut(e, "Pintail", "readQuantiles"); + m->errorOut(e, "Pintail", "print"); exit(1); } } +#endif //*************************************************************************************************************** -//calculate the distances from each query sequence to all sequences in the template to find the closest sequence -vector Pintail::findPairs(int start, int end) { +int Pintail::getChimeras(Sequence* query) { try { + querySeq = query; + trimmed.clear(); + windowSizes = window; + + //find pairs has to be done before a mask + bestfit = findPairs(query); + + if (m->control_pressed) { return 0; } + + //if they mask + if (seqMask != "") { + decalc->runMask(query); + decalc->runMask(bestfit); + } + + if (filter) { //must be done after a mask + runFilter(query); + runFilter(bestfit); + } - vector seqsMatches; + + //trim seq + decalc->trimSeqs(query, bestfit, trimmed); - for(int i = start; i < end; i++){ + //find windows + it = trimmed.begin(); + windowsForeachQuery = decalc->findWindows(query, it->first, it->second, windowSizes, increment); + + //find observed distance + obsDistance = decalc->calcObserved(query, bestfit, windowsForeachQuery, windowSizes); - float smallest = 10000.0; - Sequence query = *(querySeqs[i]); - Sequence* match; - - for(int j = 0; j < templateSeqs.size(); j++){ - - Sequence temp = *(templateSeqs[j]); + if (m->control_pressed) { return 0; } - distcalculator->calcDist(query, temp); - float dist = distcalculator->getDist(); - - if (dist < smallest) { - match = templateSeqs[j]; - smallest = dist; - } - } - - seqsMatches.push_back(match); - } + Qav = decalc->findQav(windowsForeachQuery, windowSizes, probabilityProfile); - return seqsMatches; - + if (m->control_pressed) { return 0; } + + //find alpha + seqCoef = decalc->getCoef(obsDistance, Qav); + + //calculating expected distance + expectedDistance = decalc->calcExpected(Qav, seqCoef); + + if (m->control_pressed) { return 0; } + + //finding de + DE = decalc->calcDE(obsDistance, expectedDistance); + + if (m->control_pressed) { return 0; } + + //find distance between query and closest match + it = trimmed.begin(); + deviation = decalc->calcDist(query, bestfit, it->first, it->second); + + delete bestfit; + + return 0; } catch(exception& e) { - errorOut(e, "Pintail", "findPairs"); + m->errorOut(e, "Pintail", "getChimeras"); exit(1); } } -/**************************************************************************************************/ +//*************************************************************************************************************** -void Pintail::createProcessesSpots() { +vector Pintail::readFreq() { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - vector processIDS; + //read in probabilities and store in vector + int pos; float num; + + vector prob; + set h = decalc->getPos(); //positions of bases in masking sequence - //loop through and create all the processes you want - while (process != processors) { - int pid = fork(); - - if (pid > 0) { - processIDS.push_back(pid); - process++; - }else if (pid == 0){ - - for (int j = lines[process]->start; j < lines[process]->end; j++) { - - //chops off beginning and end of sequences so they both start and end with a base - map trim; - decalc->trimSeqs(querySeqs[j], bestfit[j], trim); - trimmed[j] = trim; - - } + #ifdef USE_MPI + + MPI_File inMPI; + MPI_Offset size; + MPI_Status status; - mothurOut("Finding window breaks for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); - for (int i = lines[process]->start; i < lines[process]->end; i++) { - it = trimmed[i].begin(); - windowsForeachQuery[i] = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment); - } - mothurOut("Done finding window breaks for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); - - //write out data to file so parent can read it - ofstream out; - string s = toString(getpid()) + ".temp"; - openOutputFile(s, out); - - //output windowsForeachQuery - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << windowsForeachQuery[i].size() << '\t'; - for (int j = 0; j < windowsForeachQuery[i].size(); j++) { - out << windowsForeachQuery[i][j] << '\t'; - } - out << endl; - } - - //output windowSizes - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << windowSizes[i] << '\t'; - } - out << endl; - - //output trimmed values - for (int i = lines[process]->start; i < lines[process]->end; i++) { - it = trimmed[i].begin(); - - out << it->first << '\t' << it->second << endl; - } - out.close(); - - exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } - } + //char* inFileName = new char[consfile.length()]; + //memcpy(inFileName, consfile.c_str(), consfile.length()); - //force parent to wait until all the processes are done - for (int i=0;i size) { tempBuf = tempBuf.substr(0, size); } + istringstream iss (tempBuf,istringstream::in); - //get data created by processes - for (int i=0;iend - lines[i]->start; - - int count = lines[i]->start; - for (int m = 0; m < size; m++) { - int num; - in >> num; - - vector win; int w; - for (int j = 0; j < num; j++) { - in >> w; - win.push_back(w); - } + //read version + string line = m->getline(iss); m->gobble(iss); + + while(!iss.eof()) { + iss >> pos >> num; + + if (h.count(pos) > 0) { + float Pi; + Pi = (num - 0.25) / 0.75; - windowsForeachQuery[count] = win; - count++; - gobble(in); + //cannot have probability less than 0. + if (Pi < 0) { Pi = 0.0; } + + //do you want this spot + prob.push_back(Pi); } + + m->gobble(iss); + } + + MPI_File_close(&inMPI); - gobble(in); - count = lines[i]->start; - for (int m = 0; m < size; m++) { - int num; - in >> num; + #else + + ifstream in; + m->openInputFile(consfile, in); + + //read version + string line = m->getline(in); m->gobble(in); - windowSizes[count] = num; - count++; - } + while(!in.eof()){ - gobble(in); + in >> pos >> num; - count = lines[i]->start; - for (int m = 0; m < size; m++) { - int front, back; - in >> front >> back; - - map t; - - t[front] = back; - - trimmed[count] = t; - count++; - - gobble(in); - } + if (h.count(pos) > 0) { + float Pi; + Pi = (num - 0.25) / 0.75; + + //cannot have probability less than 0. + if (Pi < 0) { Pi = 0.0; } + //do you want this spot + prob.push_back(Pi); + } - in.close(); - remove(s.c_str()); + m->gobble(in); } - + in.close(); + + #endif -#else - for (int j = 0; j < bestfit.size(); j++) { - //chops off beginning and end of sequences so they both start and end with a base - decalc->trimSeqs(querySeqs[j], bestfit[j], trimmed[j]); - } - - for (int i = lines[0]->start; i < lines[0]->end; i++) { - it = trimmed[i].begin(); - map win = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment); - windows[i] = win; - } - -#endif + return prob; + } catch(exception& e) { - errorOut(e, "Pintail", "createProcessesSpots"); + m->errorOut(e, "Pintail", "readFreq"); exit(1); } } -/**************************************************************************************************/ -void Pintail::createProcessesPairs() { +//*************************************************************************************************************** +//calculate the distances from each query sequence to all sequences in the template to find the closest sequence +Sequence* Pintail::findPairs(Sequence* q) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - vector processIDS; - //loop through and create all the processes you want - while (process != processors) { - int pid = fork(); - - if (pid > 0) { - processIDS.push_back(pid); - process++; - }else if (pid == 0){ - - mothurOut("Finding pairs for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); - bestfit = findPairs(lines[process]->start, lines[process]->end); - mothurOut("Done finding pairs for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); - - //write out data to file so parent can read it - ofstream out; - string s = toString(getpid()) + ".temp"; - openOutputFile(s, out); - - //output range and size - out << bestfit.size() << endl; - - //output pairs - for (int i = 0; i < bestfit.size(); i++) { - out << ">" << bestfit[i]->getName() << endl << bestfit[i]->getAligned() << endl; - } - out.close(); - - exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } - } + Sequence* seqsMatches; - //force parent to wait until all the processes are done - for (int i=0;i> size; gobble(in); - - //get pairs - int count = lines[i]->start; - for (int m = 0; m < size; m++) { - Sequence* temp = new Sequence(in); - bestfit[count] = temp; - - count++; - gobble(in); - } - - in.close(); - remove(s.c_str()); - } - + seqsMatches = decalc->findClosest(q, templateSeqs); + return seqsMatches; -#else - bestfit = findPairs(lines[0]->start, lines[0]->end); -#endif } catch(exception& e) { - errorOut(e, "Pintail", "createProcessesPairs"); + m->errorOut(e, "Pintail", "findPairs"); exit(1); } } -/**************************************************************************************************/ - -void Pintail::createProcesses() { +//************************************************************************************************** +void Pintail::createProcessesQuan() { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + int process = 1; vector processIDS; - + //loop through and create all the processes you want while (process != processors) { int pid = fork(); @@ -699,304 +542,234 @@ void Pintail::createProcesses() { process++; }else if (pid == 0){ - mothurOut("Calculating observed, expected and de values for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); - for (int i = lines[process]->start; i < lines[process]->end; i++) { - - vector obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windowsForeachQuery[i], windowSizes[i]); - obsDistance[i] = obsi; - - //calc Qav - vector q = decalc->findQav(windowsForeachQuery[i], windowSizes[i], probabilityProfile); - - //get alpha - float alpha = decalc->getCoef(obsDistance[i], q); - - //find expected - vector exp = decalc->calcExpected(q, alpha); - expectedDistance[i] = exp; - - //get de and deviation - float dei = decalc->calcDE(obsi, exp); - DE[i] = dei; - - it = trimmed[i].begin(); - float dist = decalc->calcDist(querySeqs[i], bestfit[i], it->first, it->second); - deviation[i] = dist; - } - mothurOut("Done calculating observed, expected and de values for sequences " + toString(lines[process]->start) + " to " + toString(lines[process]->end)); mothurOutEndLine(); + quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, templateLines[process]->start, templateLines[process]->end); //write out data to file so parent can read it ofstream out; string s = toString(getpid()) + ".temp"; - openOutputFile(s, out); - - int size = lines[process]->end - lines[process]->start; - out << size << endl; + m->openOutputFile(s, out); //output observed distances - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << obsDistance[i].size() << '\t'; - for (int j = 0; j < obsDistance[i].size(); j++) { - out << obsDistance[i][j] << '\t'; + for (int i = 0; i < quantilesMembers.size(); i++) { + out << quantilesMembers[i].size() << '\t'; + for (int j = 0; j < quantilesMembers[i].size(); j++) { + out << quantilesMembers[i][j] << '\t'; } out << endl; } - - //output expected distances - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << expectedDistance[i].size() << '\t'; - for (int j = 0; j < expectedDistance[i].size(); j++) { - out << expectedDistance[i][j] << '\t'; - } - out << endl; - } - - - //output de values - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << DE[i] << '\t'; - } - out << endl; - - //output de values - for (int i = lines[process]->start; i < lines[process]->end; i++) { - out << deviation[i] << '\t'; - } - out << endl; - out.close(); - + + exit(0); + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } + } } + //parent does its part + quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, templateLines[0]->start, templateLines[0]->end); + //force parent to wait until all the processes are done - for (int i=0;i> size; gobble(in); - - //get observed distances - int count = lines[i]->start; - for (int m = 0; m < size; m++) { - int num; - in >> num; - - vector obs; float w; - for (int j = 0; j < num; j++) { - in >> w; - obs.push_back(w); - } - - obsDistance[count] = obs; - count++; - gobble(in); - } + m->openInputFile(s, in); - gobble(in); + vector< vector > quan; + quan.resize(100); - //get expected distances - count = lines[i]->start; - for (int m = 0; m < size; m++) { + //get quantiles + for (int h = 0; h < quan.size(); h++) { int num; - in >> num; - - vector exp; float w; + in >> num; + + m->gobble(in); + + vector q; float w; for (int j = 0; j < num; j++) { in >> w; - exp.push_back(w); + q.push_back(w); } - - expectedDistance[count] = exp; - count++; - gobble(in); - } - gobble(in); - - count = lines[i]->start; - for (int m = 0; m < size; m++) { - float num; - in >> num; - - DE[count] = num; - count++; + quan[h] = q; + m->gobble(in); } - gobble(in); - - count = lines[i]->start; - for (int m = 0; m < size; m++) { - float num; - in >> num; - - deviation[count] = num; - count++; + + //save quan in quantiles + for (int j = 0; j < quan.size(); j++) { + //put all values of q[i] into quan[i] + for (int l = 0; l < quan[j].size(); l++) { quantilesMembers[j].push_back(quan[j][l]); } + //quantilesMembers[j].insert(quantilesMembers[j].begin(), quan[j].begin(), quan[j].end()); } - + in.close(); - remove(s.c_str()); + m->mothurRemove(s); } - #else - mothurOut("Calculating observed distance... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - vector obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windows[i], windowSizes[i]); - obsDistance[i] = obsi; - } - mothurOut("Done."); mothurOutEndLine(); - - + quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size()); +#endif + } + catch(exception& e) { + m->errorOut(e, "Pintail", "createProcessesQuan"); + exit(1); + } +} +//*************************************************************************************************************** +vector< vector > Pintail::readQuantiles() { + try { + int num; + float ten, twentyfive, fifty, seventyfive, ninetyfive, ninetynine; + + vector< vector > quan; + vector temp; temp.resize(6, 0); + + //to fill 0 + quan.push_back(temp); + + #ifdef USE_MPI + + MPI_File inMPI; + MPI_Offset size; + MPI_Status status; + + //char* inFileName = new char[quanfile.length()]; + //memcpy(inFileName, quanfile.c_str(), quanfile.length()); + + char inFileName[1024]; + strcpy(inFileName, quanfile.c_str()); + + MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); + MPI_File_get_size(inMPI, &size); + //delete inFileName; + + + char* buffer = new char[size]; + MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status); + + string tempBuf = buffer; + if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); } + istringstream iss (tempBuf,istringstream::in); + delete buffer; + + //read version + string line = m->getline(iss); m->gobble(iss); + + while(!iss.eof()) { + iss >> num >> ten >> twentyfive >> fifty >> seventyfive >> ninetyfive >> ninetynine; - mothurOut("Finding variability... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - vector q = decalc->findQav(windows[i], windowSizes[i], probabilityProfile, h[i]); - Qav[i] = q; - } - mothurOut("Done."); mothurOutEndLine(); + temp.clear(); + temp.push_back(ten); + temp.push_back(twentyfive); + temp.push_back(fifty); + temp.push_back(seventyfive); + temp.push_back(ninetyfive); + temp.push_back(ninetynine); + quan.push_back(temp); - mothurOut("Calculating alpha... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - float alpha = decalc->getCoef(obsDistance[i], Qav[i]); - seqCoef.push_back(alpha); - } - mothurOut("Done."); mothurOutEndLine(); - + m->gobble(iss); + } + + MPI_File_close(&inMPI); + #else + + ifstream in; + m->openInputFile(quanfile, in); - mothurOut("Calculating expected distance... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - vector exp = decalc->calcExpected(Qav[i], seqCoef[i]); - expectedDistance[i] = exp; - } - mothurOut("Done."); mothurOutEndLine(); + //read version + string line = m->getline(in); m->gobble(in); + while(!in.eof()){ + in >> num >> ten >> twentyfive >> fifty >> seventyfive >> ninetyfive >> ninetynine; - mothurOut("Finding deviation... "); cout.flush(); - for (int i = lines[0]->start; i < lines[0]->end; i++) { - float de = decalc->calcDE(obsDistance[i], expectedDistance[i]); - DE[i] = de; - - it = trimmed[i].begin(); - float dist = decalc->calcDist(querySeqs[i], bestfit[i], it->first, it->second); - deviation[i] = dist; - } - mothurOut("Done."); mothurOutEndLine(); - -#endif + temp.clear(); + + temp.push_back(ten); + temp.push_back(twentyfive); + temp.push_back(fifty); + temp.push_back(seventyfive); + temp.push_back(ninetyfive); + temp.push_back(ninetynine); + + quan.push_back(temp); + + m->gobble(in); + } + in.close(); + #endif + + return quan; + } catch(exception& e) { - errorOut(e, "Pintail", "createProcesses"); + m->errorOut(e, "Pintail", "readQuantiles"); exit(1); } } +//***************************************************************************************************************/ - -/**************************************************************************************************/ - -void Pintail::createProcessesQuan() { +void Pintail::printQuanFile(string file, string outputString) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - vector processIDS; - - //loop through and create all the processes you want - while (process != processors) { - int pid = fork(); - - if (pid > 0) { - processIDS.push_back(pid); - process++; - }else if (pid == 0){ - - quantiles = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, templateLines[process]->start, templateLines[process]->end); - - //write out data to file so parent can read it - ofstream out; - string s = toString(getpid()) + ".temp"; - openOutputFile(s, out); - - - //output observed distances - for (int i = 0; i < quantiles.size(); i++) { - out << quantiles[i].size() << '\t'; - for (int j = 0; j < quantiles[i].size(); j++) { - out << quantiles[i][j] << '\t'; - } - out << endl; - } - - out.close(); - - exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } - } - - //force parent to wait until all the processes are done - for (int i=0;i > quan; quan.resize(100); - - //get quantiles - for (int m = 0; m < quan.size(); m++) { - int num; - in >> num; + int pid; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; + + //char* FileName = new char[file.length()]; + //memcpy(FileName, file.c_str(), file.length()); - vector q; float w; - for (int j = 0; j < num; j++) { - in >> w; - q.push_back(w); - } + char FileName[1024]; + strcpy(FileName, file.c_str()); - quan[m] = q; - gobble(in); + if (pid == 0) { + MPI_File_open(MPI_COMM_SELF, FileName, outMode, MPI_INFO_NULL, &outQuan); //comm, filename, mode, info, filepointer + + int length = outputString.length(); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write(outQuan, buf, length, MPI_CHAR, &status); + delete buf; + + MPI_File_close(&outQuan); } + + //delete FileName; + #else + ofstream outQuan; + m->openOutputFile(file, outQuan); + outQuan << outputString; - //save quan in quantiles - for (int i = 0; i < quan.size(); i++) { - //put all values of q[i] into quan[i] - quantiles[i].insert(quantiles[i].begin(), quan[i].begin(), quan[i].end()); - } - - in.close(); - remove(s.c_str()); - } - -#else - quantiles = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size()); -#endif + outQuan.close(); + #endif } catch(exception& e) { - errorOut(e, "Pintail", "createProcessesQuan"); + m->errorOut(e, "Pintail", "printQuanFile"); exit(1); } } +//***************************************************************************************************************/ -//***************************************************************************************************************