#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(string filename, string temp, string o) { fastafile = filename; templateFile = temp; outputDir = o; }
//***************************************************************************************************************
Pintail::~Pintail() {
try {
for (int i = 0; i < querySeqs.size(); i++) { delete querySeqs[i]; }
for (int i = 0; i < templateSeqs.size(); i++) { delete templateSeqs[i]; }
+ for (int i = 0; i < bestfit.size(); i++) { delete bestfit[i]; }
}
catch(exception& e) {
errorOut(e, "Pintail", "~Pintail");
void Pintail::print(ostream& out) {
try {
+ mothurOutEndLine();
+
for (int i = 0; i < querySeqs.size(); i++) {
int index = ceil(deviation[i]);
-
+
//is your DE value higher than the 95%
string chimera;
- if (DE[i] > quantiles[index][4]) { chimera = "Yes"; }
- else { chimera = "No"; }
+ if (quantiles[index][4] == 0.0) {
+ chimera = "Your template does not include sequences that provide quantile values at distance " + toString(index);
+ }else {
+ 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") {
}
//***************************************************************************************************************
-void Pintail::getChimeras() {
+int Pintail::getChimeras() {
try {
//read in query sequences and subject sequences
int numSeqs = querySeqs.size();
+ if (unaligned) { mothurOut("Your sequences need to be aligned when you use the pintail method."); mothurOutEndLine(); return 1; }
+
obsDistance.resize(numSeqs);
expectedDistance.resize(numSeqs);
seqCoef.resize(numSeqs);
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 ;
templateLines.push_back(new linePair(0, templateSeqs.size()));
#endif
- distcalculator = new ignoreGaps();
+ distcalculator = new eachGapDist();
decalc = new DeCalculator();
- decalc->setMask(seqMask);
-
- //mask querys
- for (int i = 0; i < querySeqs.size(); i++) {
- decalc->runMask(querySeqs[i]);
- }
+ //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()); }
- //mask templates
- for (int i = 0; i < templateSeqs.size(); i++) {
- decalc->runMask(templateSeqs[i]);
- }
+ decalc->setMask(seqMask);
-for (int i = 0; i < lines.size(); i++) { cout << "line pair " << i << " = " << lines[i]->start << '\t' << lines[i]->end << endl; }
-
+ //find pairs
if (processors == 1) {
mothurOut("Finding closest sequence in template to each sequence... "); cout.flush();
bestfit = findPairs(lines[0]->start, lines[0]->end);
-
- //ex.align matches from wigeon
-for (int m = 0; m < templateSeqs.size(); m++) {
- if (templateSeqs[m]->getName() == "159481") { bestfit[17] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "100137") { bestfit[16] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "112956") { bestfit[15] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "102326") { bestfit[14] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "66229") { bestfit[13] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "206276") { bestfit[12] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "63607") { bestfit[11] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "7056") { bestfit[10] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "7088") { bestfit[9] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "17553") { bestfit[8] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "131723") { bestfit[7] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "69013") { bestfit[6] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "24543") { bestfit[5] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "27824") { bestfit[4] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "1456") { bestfit[3] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "1456") { bestfit[2] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "141312") { bestfit[1] = *(templateSeqs[m]); }
- if (templateSeqs[m]->getName() == "141312") { bestfit[0] = *(templateSeqs[m]); }
+ mothurOut("Done."); mothurOutEndLine();
+ }else { createProcessesPairs(); }
+
+//string o = "closestmatch.eachgap.fasta";
+//ofstream out7;
+//openOutputFile(o, out7);
+//for (int i = 0; i < bestfit.size(); i++) {
+ //out7 << ">" << querySeqs[i]->getName() << "-"<< bestfit[i]->getName() << endl;
+ //out7 << bestfit[i]->getAligned() << endl;
+//}
+//out7.close();
+ //find P
+ 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, outputDir + getSimpleName(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 sequences if the user wants to
+ if (seqMask != "") {
+ //mask querys
+ for (int i = 0; i < querySeqs.size(); i++) {
+ decalc->runMask(querySeqs[i]);
+ }
+
+ //mask templates
+ for (int i = 0; i < templateSeqs.size(); i++) {
+ decalc->runMask(templateSeqs[i]);
+ }
+
+ for (int i = 0; i < bestfit.size(); i++) {
+ decalc->runMask(bestfit[i]);
+ }
+
+ }
+
+ if (filter) {
+ vector<Sequence*> temp = templateSeqs;
+ for (int i = 0; i < querySeqs.size(); i++) { temp.push_back(querySeqs[i]); }
+
+ createFilter(temp);
+ runFilter(querySeqs);
+ runFilter(templateSeqs);
+ runFilter(bestfit);
+ }
+
+
+ if (processors == 1) {
+
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]);
}
- mothurOut("Done."); mothurOutEndLine();
mothurOut("Finding window breaks... "); cout.flush();
for (int i = lines[0]->start; i < lines[0]->end; i++) {
it = trimmed[i].begin();
-//cout << "trimmed = " << it->first << '\t' << it->second << endl;
vector<int> win = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment);
windowsForeachQuery[i] = win;
}
mothurOut("Done."); mothurOutEndLine();
}else { createProcessesSpots(); }
-
- //find P
- 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 quantiles to a .prob file so that you can input them using the conservation parameter next time you run this command. Providing the .prob file will dramatically 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]; }
- mothurOut("Done."); mothurOutEndLine();
if (processors == 1) {
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<float> obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windowsForeachQuery[i], windowSizes[i]);
+
obsDistance[i] = obsi;
}
mothurOut("Done."); mothurOutEndLine();
vector<float> q = decalc->findQav(windowsForeachQuery[i], windowSizes[i], probabilityProfile);
Qav[i] = q;
-//cout << i+1 << endl;
-//for (int j = 0; j < Qav[i].size(); j++) {
- //cout << Qav[i][j] << '\t';
-//}
-//cout << endl << endl;
-
}
mothurOut("Done."); mothurOutEndLine();
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 << i+1 << "\tcoef = " << alpha << endl;
seqCoef[i] = alpha;
}
mothurOut("Done."); mothurOutEndLine();
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;
}
else { createProcesses(); }
-
//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(); }
+ if (quanfile != "") { quantiles = readQuantiles(); }
else {
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(); }
+
+
+ ofstream out4, out5;
+ string noOutliers, outliers;
- ofstream out4;
- string o = getRootName(templateFile) + "quan";
+ if ((!filter) && (seqMask == "")) {
+ noOutliers = outputDir + getRootName(getSimpleName(templateFile)) + "pintail.quan";
+ }else if ((filter) && (seqMask == "")) {
+ noOutliers = outputDir + getRootName(getSimpleName(templateFile)) + "pintail.filtered.quan";
+ }else if ((!filter) && (seqMask != "")) {
+ noOutliers = outputDir + getRootName(getSimpleName(templateFile)) + "pintail.masked.quan";
+ }else if ((filter) && (seqMask != "")) {
+ noOutliers = outputDir + getRootName(getSimpleName(templateFile)) + "pintail.filtered.masked.quan";
+ }
+
+ //outliers = getRootName(templateFile) + "pintail.quanYESOUTLIERS";
- openOutputFile(o, out4);
+ /*openOutputFile(outliers, out4);
//adjust quantiles
- for (int i = 0; i < quantiles.size(); i++) {
- if (quantiles[i].size() == 0) {
+ for (int i = 0; i < quantilesMembers.size(); i++) {
+ vector<float> 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(), compareQuanMembers);
- vector<float> 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)].score);
//save 25%
- temp.push_back(quantiles[i][int(quantiles[i].size() * 0.25)]);
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.25)].score);
//save 50%
- temp.push_back(quantiles[i][int(quantiles[i].size() * 0.5)]);
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.5)].score);
//save 75%
- temp.push_back(quantiles[i][int(quantiles[i].size() * 0.75)]);
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.75)].score);
//save 95%
- temp.push_back(quantiles[i][int(quantiles[i].size() * 0.95)]);
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.95)].score);
//save 99%
- temp.push_back(quantiles[i][int(quantiles[i].size() * 0.99)]);
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.99)].score);
- quantiles[i] = temp;
}
//output quan value
out4 << i+1 << '\t';
- for (int u = 0; u < quantiles[i].size(); u++) { out4 << quantiles[i][u] << '\t'; }
+ for (int u = 0; u < temp.size(); u++) { out4 << temp[u] << '\t'; }
out4 << endl;
-
+
+ quantiles[i] = temp;
+
}
+ out4.close();*/
+
+ decalc->removeObviousOutliers(quantilesMembers, templateSeqs.size());
+
+ openOutputFile(noOutliers, out5);
+
+ //adjust quantiles
+ for (int i = 0; i < quantilesMembers.size(); i++) {
+ vector<float> 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++) {
+ temp.push_back(0.0);
+ }
+ }else{
+
+ sort(quantilesMembers[i].begin(), quantilesMembers[i].end(), compareQuanMembers);
+
+ //save 10%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.10)].score);
+ //save 25%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.25)].score);
+ //save 50%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.5)].score);
+ //save 75%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.75)].score);
+ //save 95%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.95)].score);
+ //save 99%
+ temp.push_back(quantilesMembers[i][int(quantilesMembers[i].size() * 0.99)].score);
+
+ }
+
+ //output quan value
+ out5 << i+1 << '\t';
+ for (int u = 0; u < temp.size(); u++) { out5 << temp[u] << '\t'; }
+ out5 << endl;
+
+ quantiles[i] = temp;
+
+ }
+
mothurOut("Done."); mothurOutEndLine();
+
}
-
+
//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;
+
+ return 0;
}
catch(exception& e) {
errorOut(e, "Pintail", "getChimeras");
openInputFile(consfile, in);
vector<float> prob;
+ set<int> h = decalc->getPos(); //positions of bases in masking sequence
//read in probabilities and store in vector
int pos; float num;
in >> pos >> num;
- //do you want this spot
- prob.push_back(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);
+ }
gobble(in);
}
}
}
-//***************************************************************************************************************
-
-vector< vector<float> > Pintail::readQuantiles() {
- try {
-
- ifstream in;
- openInputFile(quanfile, in);
-
- vector< vector<float> > quan;
-
- int num; float ten, twentyfive, fifty, seventyfive, ninetyfive, ninetynine;
-
- while(!in.eof()){
-
- in >> num >> ten >> twentyfive >> fifty >> seventyfive >> ninetyfive >> ninetynine;
-
- vector <float> 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);
-
- gobble(in);
- }
-
- in.close();
- return quan;
-
- }
- catch(exception& e) {
- errorOut(e, "Pintail", "readQuantiles");
- exit(1);
- }
-}
//***************************************************************************************************************
//calculate the distances from each query sequence to all sequences in the template to find the closest sequence
-vector<Sequence> Pintail::findPairs(int start, int end) {
+vector<Sequence*> Pintail::findPairs(int start, int end) {
try {
- vector<Sequence> seqsMatches;
+ vector<Sequence*> seqsMatches;
for(int i = start; i < end; i++){
-
- float smallest = 10000.0;
- Sequence query = *(querySeqs[i]);
- Sequence match;
- for(int j = 0; j < templateSeqs.size(); j++){
-
- Sequence temp = *(templateSeqs[j]);
-
- distcalculator->calcDist(query, temp);
- float dist = distcalculator->getDist();
-
- if (dist < smallest) {
- match = *(templateSeqs[j]);
- smallest = dist;
- }
- }
-
- seqsMatches.push_back(match);
+ vector<Sequence*> copy = decalc->findClosest(querySeqs[i], templateSeqs, 1);
+ seqsMatches.push_back(copy[0]);
}
return seqsMatches;
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();
-
- int count = lines[process]->start;
- for (int j = 0; j < bestfit.size(); j++) {
+ 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<int, int> trim;
- decalc->trimSeqs(querySeqs[count], bestfit[j], trim);
- trimmed[count] = trim;
+
+ decalc->trimSeqs(querySeqs[j], bestfit[j], trim);
+ trimmed[j] = trim;
- count++;
}
mothurOut("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(pid) + ".temp";
+ 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;
- }
-
//output windowsForeachQuery
- for (int i = 0; i < windowsForeachQuery.size(); i++) {
+ 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 = 0; i < windowSizes.size(); i++) {
+ for (int i = lines[process]->start; i < lines[process]->end; i++) {
out << windowSizes[i] << '\t';
}
- out << endl;
+ 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);
string s = toString(processIDS[i]) + ".temp";
openInputFile(s, in);
- int size;
- in >> size; gobble(in);
-
- //get pairs
+ int size = lines[i]->end - lines[i]->start;
+
int count = lines[i]->start;
- for (int m = 0; m < size; m++) {
- Sequence temp(in);
- bestfit[count] = temp;
-
- count++;
- gobble(in);
- }
-
- gobble(in);
-
- count = lines[i]->start;
for (int m = 0; m < size; m++) {
int num;
in >> num;
-
+
vector<int> win; int w;
for (int j = 0; j < num; j++) {
in >> w;
win.push_back(w);
}
-
+
windowsForeachQuery[count] = win;
count++;
gobble(in);
}
-
+
gobble(in);
count = lines[i]->start;
- for (int i = 0; i < size; i++) {
+ for (int m = 0; m < size; m++) {
int num;
in >> num;
count++;
}
+ gobble(in);
+
+ count = lines[i]->start;
+ for (int m = 0; m < size; m++) {
+ int front, back;
+ in >> front >> back;
+
+ map<int, int> t;
+
+ t[front] = back;
+
+ trimmed[count] = t;
+ count++;
+
+ gobble(in);
+ }
+
+
in.close();
+ remove(s.c_str());
}
-
+
#else
- bestfit = findPairs(lines[0]->start, lines[0]->end);
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]);
+ //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<int, int> win = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment);
- windows[i] = win;
+ vector<int> win = decalc->findWindows(querySeqs[i], it->first, it->second, windowSizes[i], increment);
+ windowsForeachQuery[i] = win;
}
#endif
exit(1);
}
}
-
-
/**************************************************************************************************/
-void Pintail::createProcesses() {
+void Pintail::createProcessesPairs() {
try {
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
int process = 0;
vector<int> processIDS;
- vector< vector<float> > exp; exp.resize(querySeqs.size());
- vector<float> de; de.resize(querySeqs.size());
- vector< vector<float> > obs; obs.resize(querySeqs.size());
- vector<float> dev; dev.resize(querySeqs.size());
+ //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); }
+ }
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processors;i++) {
+ int temp = processIDS[i];
+ wait(&temp);
+ }
+ //get data created by processes
+ for (int i=0;i<processors;i++) {
+ ifstream in;
+ string s = toString(processIDS[i]) + ".temp";
+ openInputFile(s, in);
+
+ int size;
+ in >> 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());
+ }
+
+
+#else
+ bestfit = findPairs(lines[0]->start, lines[0]->end);
+#endif
+ }
+ catch(exception& e) {
+ errorOut(e, "Pintail", "createProcessesPairs");
+ exit(1);
+ }
+}
+/**************************************************************************************************/
+
+void Pintail::createProcesses() {
+ try {
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ int process = 0;
+ vector<int> processIDS;
//loop through and create all the processes you want
while (process != processors) {
for (int i = lines[process]->start; i < lines[process]->end; i++) {
vector<float> obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windowsForeachQuery[i], windowSizes[i]);
- obs[i] = obsi;
+ obsDistance[i] = obsi;
//calc Qav
vector<float> q = decalc->findQav(windowsForeachQuery[i], windowSizes[i], probabilityProfile);
//get de and deviation
float dei = decalc->calcDE(obsi, exp);
- de[i] = dei;
+ DE[i] = dei;
it = trimmed[i].begin();
float dist = decalc->calcDist(querySeqs[i], bestfit[i], it->first, it->second);
- dev[i] = dist;
+ deviation[i] = dist;
}
mothurOut("Done calculating observed, expected and de values 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);
+
+ int size = lines[process]->end - lines[process]->start;
+ out << size << endl;
+
+ //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';
+ }
+ 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 { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); }
}
wait(&temp);
}
- obsDistance = obs;
- expectedDistance = exp;
- DE = de;
- deviation = dev;
-
+ //get data created by processes
+ for (int i=0;i<processors;i++) {
+ ifstream in;
+ string s = toString(processIDS[i]) + ".temp";
+ openInputFile(s, in);
+
+ int size;
+ in >> size; gobble(in);
+
+ //get observed distances
+ int count = lines[i]->start;
+ for (int m = 0; m < size; m++) {
+ int num;
+ in >> num;
+
+ vector<float> obs; float w;
+ for (int j = 0; j < num; j++) {
+ in >> w;
+ obs.push_back(w);
+ }
+
+ obsDistance[count] = obs;
+ count++;
+ gobble(in);
+ }
+
+ gobble(in);
+
+ //get expected distances
+ count = lines[i]->start;
+ for (int m = 0; m < size; m++) {
+ int num;
+ in >> num;
+
+ vector<float> exp; float w;
+ for (int j = 0; j < num; j++) {
+ in >> w;
+ exp.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++;
+ }
+
+ gobble(in);
+
+ count = lines[i]->start;
+ for (int m = 0; m < size; m++) {
+ float num;
+ in >> num;
+
+ deviation[count] = num;
+ count++;
+ }
+
+ in.close();
+ remove(s.c_str());
+ }
+
+
#else
mothurOut("Calculating observed distance... "); cout.flush();
for (int i = lines[0]->start; i < lines[0]->end; i++) {
- vector<float> obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windows[i], windowSizes[i]);
+ vector<float> obsi = decalc->calcObserved(querySeqs[i], bestfit[i], windowsForeachQuery[i], windowSizes[i]);
obsDistance[i] = obsi;
}
mothurOut("Done."); mothurOutEndLine();
mothurOut("Finding variability... "); cout.flush();
for (int i = lines[0]->start; i < lines[0]->end; i++) {
- vector<float> q = decalc->findQav(windows[i], windowSizes[i], probabilityProfile, h[i]);
+ vector<float> q = decalc->findQav(windowsForeachQuery[i], windowSizes[i], probabilityProfile);
Qav[i] = q;
}
mothurOut("Done."); mothurOutEndLine();
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
int process = 0;
vector<int> processIDS;
- vector< vector<float> > quan; quan.resize(100);
//loop through and create all the processes you want
while (process != processors) {
process++;
}else if (pid == 0){
- vector< vector<float> > q = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, templateLines[process]->start, templateLines[process]->end);
+ quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, templateLines[process]->start, templateLines[process]->end);
- for (int i = 0; i < q.size(); i++) {
- //put all values of q[i] into quan[i]
- quan[i].insert(quan[i].begin(), q[i].begin(), q[i].end());
- }
+ //write out data to file so parent can read it
+ ofstream out;
+ string s = toString(getpid()) + ".temp";
+ openOutputFile(s, out);
- for (int i = 0; i < quan.size(); i++) {
- cout << i+1 << '\t';
- for (int j = 0; j < quan[i].size(); j++) { cout << quan[i][j] << '\t'; }
- cout << endl;
+
+ //output observed distances
+ 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].score << '\t' << quantilesMembers[i][j].member1 << '\t' << quantilesMembers[i][j].member2 << '\t';
+ }
+ out << endl;
}
-
+
+ out.close();
+
exit(0);
}else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); }
}
int temp = processIDS[i];
wait(&temp);
}
+
+ //get data created by processes
+ for (int i=0;i<processors;i++) {
+ ifstream in;
+ string s = toString(processIDS[i]) + ".temp";
+ openInputFile(s, in);
+
+ vector< vector<quanMember> > quan;
+ quan.resize(100);
+
+ //get quantiles
+ for (int m = 0; m < quan.size(); m++) {
+ int num;
+ in >> num;
+
+ gobble(in);
+
+ vector<quanMember> q; float w; int b, n;
+ for (int j = 0; j < num; j++) {
+ in >> w >> b >> n;
+ //cout << w << '\t' << b << '\t' n << endl;
+ quanMember newMember(w, b, n);
+ q.push_back(newMember);
+ }
+//cout << "here" << endl;
+ quan[m] = q;
+//cout << "now here" << endl;
+ gobble(in);
+ }
+
+
+ //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());
+ }
- quantiles = quan;
#else
- quantiles = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size());
+ quantilesMembers = decalc->getQuantiles(templateSeqs, windowSizesTemplate, window, probabilityProfile, increment, 0, templateSeqs.size());
#endif
}
catch(exception& e) {