unifracweightedcommand.cpp

   1 /*
   2  *  unifracweightedcommand.cpp
   3  *  Mothur
   4  *
   5  *  Created by Sarah Westcott on 2/9/09.
   6  *  Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
   7  *
   8  */
   9
  10 #include "unifracweightedcommand.h"
  11
  12 //**********************************************************************************************************************
  13 vector<string> UnifracWeightedCommand::getValidParameters(){
  14         try {
  15                 string Array[] =  {"groups","iters","distance","random","processors","outputdir","inputdir"};
  16                 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  17                 return myArray;
  18         }
  19         catch(exception& e) {
  20                 m->errorOut(e, "UnifracWeightedCommand", "getValidParameters");
  21                 exit(1);
  22         }
  23 }
  24 //**********************************************************************************************************************
  25 UnifracWeightedCommand::UnifracWeightedCommand(){
  26         try {
  27                 //initialize outputTypes
  28                 vector<string> tempOutNames;
  29                 outputTypes["weighted"] = tempOutNames;
  30                 outputTypes["wsummary"] = tempOutNames;
  31                 outputTypes["phylip"] = tempOutNames;
  32         }
  33         catch(exception& e) {
  34                 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
  35                 exit(1);
  36         }
  37 }
  38 //**********************************************************************************************************************
  39 vector<string> UnifracWeightedCommand::getRequiredParameters(){
  40         try {
  41                 vector<string> myArray;
  42                 return myArray;
  43         }
  44         catch(exception& e) {
  45                 m->errorOut(e, "UnifracWeightedCommand", "getRequiredParameters");
  46                 exit(1);
  47         }
  48 }
  49 //**********************************************************************************************************************
  50 vector<string> UnifracWeightedCommand::getRequiredFiles(){
  51         try {
  52                 string Array[] =  {"tree","group"};
  53                 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  54
  55                 return myArray;
  56         }
  57         catch(exception& e) {
  58                 m->errorOut(e, "UnifracWeightedCommand", "getRequiredFiles");
  59                 exit(1);
  60         }
  61 }
  62 /***********************************************************/
  63 UnifracWeightedCommand::UnifracWeightedCommand(string option) {
  64         try {
  65                 globaldata = GlobalData::getInstance();
  66                 abort = false;
  67                 Groups.clear();
  68
  69                 //allow user to run help
  70                 if(option == "help") { help(); abort = true; }
  71
  72                 else {
  73                         //valid paramters for this command
  74                         string Array[] =  {"groups","iters","distance","random","processors","outputdir","inputdir"};
  75                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  76
  77                         OptionParser parser(option);
  78                         map<string,string> parameters=parser.getParameters();
  79
  80                         ValidParameters validParameter;
  81
  82                         //check to make sure all parameters are valid for command
  83                         for (map<string,string>::iterator it = parameters.begin(); it != parameters.end(); it++) {
  84                                 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) {  abort = true;  }
  85                         }
  86
  87                         //initialize outputTypes
  88                         vector<string> tempOutNames;
  89                         outputTypes["weighted"] = tempOutNames;
  90                         outputTypes["wsummary"] = tempOutNames;
  91                         outputTypes["phylip"] = tempOutNames;
  92
  93                         if (globaldata->gTree.size() == 0) {//no trees were read
  94                                 m->mothurOut("You must execute the read.tree command, before you may execute the unifrac.weighted command."); m->mothurOutEndLine(); abort = true;  }
  95
  96                         //if the user changes the output directory command factory will send this info to us in the output parameter
  97                         outputDir = validParameter.validFile(parameters, "outputdir", false);           if (outputDir == "not found"){
  98                                 outputDir = "";
  99                                 outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it
 100                         }
 101
 102                         //check for optional parameter and set defaults
 103                         // ...at some point should added some additional type checking...
 104                         groups = validParameter.validFile(parameters, "groups", false);
 105                         if (groups == "not found") { groups = ""; }
 106                         else {
 107                                 m->splitAtDash(groups, Groups);
 108                                 globaldata->Groups = Groups;
 109                         }
 110
 111                         itersString = validParameter.validFile(parameters, "iters", false);                     if (itersString == "not found") { itersString = "1000"; }
 112                         convert(itersString, iters);
 113
 114                         string temp = validParameter.validFile(parameters, "distance", false);                  if (temp == "not found") { temp = "false"; }
 115                         phylip = m->isTrue(temp);
 116
 117                         temp = validParameter.validFile(parameters, "random", false);                                   if (temp == "not found") { temp = "F"; }
 118                         random = m->isTrue(temp);
 119
 120                         temp = validParameter.validFile(parameters, "processors", false);       if (temp == "not found"){       temp = "1";                             }
 121                         convert(temp, processors);
 122
 123                         if (!random) {  iters = 0;  } //turn off random calcs
 124
 125
 126                         if (abort == false) {
 127                                 T = globaldata->gTree;
 128                                 tmap = globaldata->gTreemap;
 129                                 sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary";
 130                                 m->openOutputFile(sumFile, outSum);
 131                                 outputNames.push_back(sumFile);  outputTypes["wsummary"].push_back(sumFile);
 132
 133                                 util = new SharedUtil();
 134                                 string s; //to make work with setgroups
 135                                 util->setGroups(globaldata->Groups, tmap->namesOfGroups, s, numGroups, "weighted");     //sets the groups the user wants to analyze
 136                                 util->getCombos(groupComb, globaldata->Groups, numComp);
 137
 138                                 weighted = new Weighted(tmap);
 139
 140                         }
 141                 }
 142
 143
 144         }
 145         catch(exception& e) {
 146                 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
 147                 exit(1);
 148         }
 149 }
 150 //**********************************************************************************************************************
 151
 152 void UnifracWeightedCommand::help(){
 153         try {
 154                 m->mothurOut("The unifrac.weighted command can only be executed after a successful read.tree command.\n");
 155                 m->mothurOut("The unifrac.weighted command parameters are groups, iters, distance, processors and random.  No parameters are required.\n");
 156                 m->mothurOut("The groups parameter allows you to specify which of the groups in your groupfile you would like analyzed.  You must enter at least 2 valid groups.\n");
 157                 m->mothurOut("The group names are separated by dashes.  The iters parameter allows you to specify how many random trees you would like compared to your tree.\n");
 158                 m->mothurOut("The distance parameter allows you to create a distance file from the results. The default is false.\n");
 159                 m->mothurOut("The random parameter allows you to shut off the comparison to random trees. The default is false, meaning don't compare your trees with randomly generated trees.\n");
 160                 m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n");
 161                 m->mothurOut("The unifrac.weighted command should be in the following format: unifrac.weighted(groups=yourGroups, iters=yourIters).\n");
 162                 m->mothurOut("Example unifrac.weighted(groups=A-B-C, iters=500).\n");
 163                 m->mothurOut("The default value for groups is all the groups in your groupfile, and iters is 1000.\n");
 164                 m->mothurOut("The unifrac.weighted command output two files: .weighted and .wsummary their descriptions are in the manual.\n");
 165                 m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n");
 166         }
 167         catch(exception& e) {
 168                 m->errorOut(e, "UnifracWeightedCommand", "help");
 169                 exit(1);
 170         }
 171 }
 172
 173 /***********************************************************/
 174 int UnifracWeightedCommand::execute() {
 175         try {
 176
 177                 if (abort == true) { return 0; }
 178
 179                 int start = time(NULL);
 180
 181                 //get weighted for users tree
 182                 userData.resize(numComp,0);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 183                 randomData.resize(numComp,0); //data[0] = weightedscore AB, data[1] = weightedscore AC...
 184
 185                 //get weighted scores for users trees
 186                 for (int i = 0; i < T.size(); i++) {
 187
 188                         if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0; }
 189
 190                         counter = 0;
 191                         rScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 192                         uScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 193
 194                         if (random) {
 195                                 output = new ColumnFile(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted", itersString);
 196                                 outputNames.push_back(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted");
 197                                 outputTypes["weighted"].push_back(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted");
 198                         }
 199
 200                         userData = weighted->getValues(T[i], processors, outputDir);  //userData[0] = weightedscore
 201
 202                         if (m->control_pressed) { if (random) { delete output; } outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str());  } return 0; }
 203
 204                         //save users score
 205                         for (int s=0; s<numComp; s++) {
 206                                 //add users score to vector of user scores
 207                                 uScores[s].push_back(userData[s]);
 208
 209                                 //save users tree score for summary file
 210                                 utreeScores.push_back(userData[s]);
 211                         }
 212
 213                         if (random) {
 214
 215                                 //calculate number of comparisons i.e. with groups A,B,C = AB, AC, BC = 3;
 216                                 vector< vector<string> > namesOfGroupCombos;
 217                                 for (int a=0; a<numGroups; a++) {
 218                                         for (int l = 0; l < a; l++) {
 219                                                 vector<string> groups; groups.push_back(globaldata->Groups[a]); groups.push_back(globaldata->Groups[l]);
 220                                                 namesOfGroupCombos.push_back(groups);
 221                                         }
 222                                 }
 223
 224                                 lines.clear();
 225
 226                                 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 227                                         if(processors != 1){
 228                                                 int numPairs = namesOfGroupCombos.size();
 229                                                 int numPairsPerProcessor = numPairs / processors;
 230
 231                                                 for (int i = 0; i < processors; i++) {
 232                                                         int startPos = i * numPairsPerProcessor;
 233                                                         if(i == processors - 1){
 234                                                                 numPairsPerProcessor = numPairs - i * numPairsPerProcessor;
 235                                                         }
 236                                                         lines.push_back(linePair(startPos, numPairsPerProcessor));
 237                                                 }
 238                                         }
 239                                 #endif
 240
 241
 242                                 //get scores for random trees
 243                                 for (int j = 0; j < iters; j++) {
 244
 245                                         #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 246                                                 if(processors == 1){
 247                                                         driver(T[i],  namesOfGroupCombos, 0, namesOfGroupCombos.size(),  rScores);
 248                                                 }else{
 249                                                         createProcesses(T[i],  namesOfGroupCombos, rScores);
 250                                                 }
 251                                         #else
 252                                                 driver(T[i], namesOfGroupCombos, 0, namesOfGroupCombos.size(), rScores);
 253                                         #endif
 254
 255                                         if (m->control_pressed) { delete output; outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str());  } return 0; }
 256
 257                                         //report progress
 258                                         m->mothurOut("Iter: " + toString(j+1)); m->mothurOutEndLine();
 259                                 }
 260                                 lines.clear();
 261
 262                                 //find the signifigance of the score for summary file
 263                                 for (int f = 0; f < numComp; f++) {
 264                                         //sort random scores
 265                                         sort(rScores[f].begin(), rScores[f].end());
 266
 267                                         //the index of the score higher than yours is returned
 268                                         //so if you have 1000 random trees the index returned is 100
 269                                         //then there are 900 trees with a score greater then you.
 270                                         //giving you a signifigance of 0.900
 271                                         int index = findIndex(userData[f], f);    if (index == -1) { m->mothurOut("error in UnifracWeightedCommand"); m->mothurOutEndLine(); exit(1); } //error code
 272
 273                                         //the signifigance is the number of trees with the users score or higher
 274                                         WScoreSig.push_back((iters-index)/(float)iters);
 275                                 }
 276
 277                                 //out << "Tree# " << i << endl;
 278                                 calculateFreqsCumuls();
 279                                 printWeightedFile();
 280
 281                                 delete output;
 282
 283                         }
 284
 285                         //clear data
 286                         rScores.clear();
 287                         uScores.clear();
 288                         validScores.clear();
 289                 }
 290
 291
 292                 if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0;  }
 293
 294                 printWSummaryFile();
 295
 296                 if (phylip) {   createPhylipFile();             }
 297
 298                 //clear out users groups
 299                 globaldata->Groups.clear();
 300
 301
 302                 if (m->control_pressed) {
 303                         for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 304                         return 0;
 305                 }
 306
 307                 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run unifrac.weighted."); m->mothurOutEndLine();
 308
 309                 m->mothurOutEndLine();
 310                 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
 311                 for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
 312                 m->mothurOutEndLine();
 313
 314                 return 0;
 315
 316         }
 317         catch(exception& e) {
 318                 m->errorOut(e, "UnifracWeightedCommand", "execute");
 319                 exit(1);
 320         }
 321 }
 322 /**************************************************************************************************/
 323
 324 int UnifracWeightedCommand::createProcesses(Tree* t, vector< vector<string> > namesOfGroupCombos, vector< vector<double> >& scores) {
 325         try {
 326 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 327                 int process = 1;
 328                 int num = 0;
 329                 vector<int> processIDS;
 330
 331                 EstOutput results;
 332
 333                 //loop through and create all the processes you want
 334                 while (process != processors) {
 335                         int pid = fork();
 336
 337                         if (pid > 0) {
 338                                 processIDS.push_back(pid);  //create map from line number to pid so you can append files in correct order later
 339                                 process++;
 340                         }else if (pid == 0){
 341                                 driver(t, namesOfGroupCombos, lines[process].start, lines[process].num, scores);
 342
 343                                 //pass numSeqs to parent
 344                                 ofstream out;
 345                                 string tempFile = outputDir + toString(getpid()) + ".weightedcommand.results.temp";
 346                                 m->openOutputFile(tempFile, out);
 347                                 for (int i = lines[process].start; i < (lines[process].start + lines[process].num); i++) { out << scores[i][(scores[i].size()-1)] << '\t';  } out << endl;
 348                                 out.close();
 349
 350                                 exit(0);
 351                         }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
 352                 }
 353
 354                 driver(t, namesOfGroupCombos, lines[0].start, lines[0].num, scores);
 355
 356                 //force parent to wait until all the processes are done
 357                 for (int i=0;i<(processors-1);i++) {
 358                         int temp = processIDS[i];
 359                         wait(&temp);
 360                 }
 361
 362                 //get data created by processes
 363                 for (int i=0;i<(processors-1);i++) {
 364
 365                         ifstream in;
 366                         string s = outputDir + toString(processIDS[i]) + ".weightedcommand.results.temp";
 367                         m->openInputFile(s, in);
 368
 369                         double tempScore;
 370                         for (int j = lines[(i+1)].start; j < (lines[(i+1)].start + lines[(i+1)].num); j++) { in >> tempScore; scores[j].push_back(tempScore); }
 371                         in.close();
 372                         remove(s.c_str());
 373                 }
 374
 375                 return 0;
 376 #endif
 377         }
 378         catch(exception& e) {
 379                 m->errorOut(e, "UnifracWeightedCommand", "createProcesses");
 380                 exit(1);
 381         }
 382 }
 383
 384 /**************************************************************************************************/
 385 int UnifracWeightedCommand::driver(Tree* t, vector< vector<string> > namesOfGroupCombos, int start, int num, vector< vector<double> >& scores) {
 386  try {
 387                 Tree* randT = new Tree();
 388
 389                 for (int h = start; h < (start+num); h++) {
 390
 391                         if (m->control_pressed) { return 0; }
 392
 393                         //initialize weighted score
 394                         string groupA = namesOfGroupCombos[h][0];
 395                         string groupB = namesOfGroupCombos[h][1];
 396
 397                         //copy T[i]'s info.
 398                         randT->getCopy(t);
 399
 400                         //create a random tree with same topology as T[i], but different labels
 401                         randT->assembleRandomUnifracTree(groupA, groupB);
 402
 403                         if (m->control_pressed) { delete randT;  return 0;  }
 404
 405                         //get wscore of random tree
 406                         EstOutput randomData = weighted->getValues(randT, groupA, groupB);
 407
 408                         if (m->control_pressed) { delete randT;  return 0;  }
 409
 410                         //save scores
 411                         scores[h].push_back(randomData[0]);
 412                 }
 413
 414                 delete randT;
 415
 416                 return 0;
 417
 418         }
 419         catch(exception& e) {
 420                 m->errorOut(e, "UnifracWeightedCommand", "driver");
 421                 exit(1);
 422         }
 423 }
 424 /***********************************************************/
 425 void UnifracWeightedCommand::printWeightedFile() {
 426         try {
 427                 vector<double> data;
 428                 vector<string> tags;
 429                 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
 430
 431                 for(int a = 0; a < numComp; a++) {
 432                         output->initFile(groupComb[a], tags);
 433                         //print each line
 434                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 435                                 data.push_back(it->first);  data.push_back(rScoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
 436                                 output->output(data);
 437                                 data.clear();
 438                         }
 439                         output->resetFile();
 440                 }
 441         }
 442         catch(exception& e) {
 443                 m->errorOut(e, "UnifracWeightedCommand", "printWeightedFile");
 444                 exit(1);
 445         }
 446 }
 447
 448
 449 /***********************************************************/
 450 void UnifracWeightedCommand::printWSummaryFile() {
 451         try {
 452                 //column headers
 453                 outSum << "Tree#" << '\t' << "Groups" << '\t' << "WScore" << '\t' << "WSig" <<  endl;
 454                 m->mothurOut("Tree#\tGroups\tWScore\tWSig"); m->mothurOutEndLine();
 455
 456                 //format output
 457                 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
 458
 459                 //print each line
 460                 int count = 0;
 461                 for (int i = 0; i < T.size(); i++) {
 462                         for (int j = 0; j < numComp; j++) {
 463                                 if (random) {
 464                                         if (WScoreSig[count] > (1/(float)iters)) {
 465                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 466                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 467                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" +  toString(WScoreSig[count]) + "\n");
 468                                         }else{
 469                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 470                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 471                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" +  toString((1/float(iters))) + "\n");
 472                                         }
 473                                 }else{
 474                                         outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 475                                         cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 476                                         m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00\n");
 477                                 }
 478                                 count++;
 479                         }
 480                 }
 481                 outSum.close();
 482         }
 483         catch(exception& e) {
 484                 m->errorOut(e, "UnifracWeightedCommand", "printWSummaryFile");
 485                 exit(1);
 486         }
 487 }
 488 /***********************************************************/
 489 void UnifracWeightedCommand::createPhylipFile() {
 490         try {
 491                 int count = 0;
 492                 //for each tree
 493                 for (int i = 0; i < T.size(); i++) {
 494
 495                         string phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted.dist";
 496                         outputNames.push_back(phylipFileName);
 497                         outputTypes["phylip"].push_back(phylipFileName);
 498                         ofstream out;
 499                         m->openOutputFile(phylipFileName, out);
 500
 501                         //output numSeqs
 502                         out << globaldata->Groups.size() << endl;
 503
 504                         //make matrix with scores in it
 505                         vector< vector<float> > dists;  dists.resize(globaldata->Groups.size());
 506                         for (int i = 0; i < globaldata->Groups.size(); i++) {
 507                                 dists[i].resize(globaldata->Groups.size(), 0.0);
 508                         }
 509
 510                         //flip it so you can print it
 511                         for (int r=0; r<globaldata->Groups.size(); r++) {
 512                                 for (int l = r+1; l < globaldata->Groups.size(); l++) {
 513                                         dists[r][l] = utreeScores[count];
 514                                         dists[l][r] = utreeScores[count];
 515                                         count++;
 516                                 }
 517                         }
 518
 519                         //output to file
 520                         for (int r=0; r<globaldata->Groups.size(); r++) {
 521                                 //output name
 522                                 string name = globaldata->Groups[r];
 523                                 if (name.length() < 10) { //pad with spaces to make compatible
 524                                         while (name.length() < 10) {  name += " ";  }
 525                                 }
 526                                 out << name << '\t';
 527
 528                                 //output distances
 529                                 for (int l = 0; l < r; l++) {   out  << dists[r][l] << '\t';  }
 530                                 out << endl;
 531                         }
 532                         out.close();
 533                 }
 534         }
 535         catch(exception& e) {
 536                 m->errorOut(e, "UnifracWeightedCommand", "createPhylipFile");
 537                 exit(1);
 538         }
 539 }
 540 /***********************************************************/
 541 int UnifracWeightedCommand::findIndex(float score, int index) {
 542         try{
 543                 for (int i = 0; i < rScores[index].size(); i++) {
 544                         if (rScores[index][i] >= score) {       return i;       }
 545                 }
 546                 return rScores[index].size();
 547         }
 548         catch(exception& e) {
 549                 m->errorOut(e, "UnifracWeightedCommand", "findIndex");
 550                 exit(1);
 551         }
 552 }
 553
 554 /***********************************************************/
 555
 556 void UnifracWeightedCommand::calculateFreqsCumuls() {
 557         try {
 558                 //clear out old tree values
 559                 rScoreFreq.clear();
 560                 rScoreFreq.resize(numComp);
 561                 rCumul.clear();
 562                 rCumul.resize(numComp);
 563                 validScores.clear();
 564
 565                 //calculate frequency
 566                 for (int f = 0; f < numComp; f++) {
 567                         for (int i = 0; i < rScores[f].size(); i++) { //looks like 0,0,1,1,1,2,4,7...  you want to make a map that say rScoreFreq[0] = 2, rScoreFreq[1] = 3...
 568                                 validScores[rScores[f][i]] = rScores[f][i];
 569                                 map<float,float>::iterator it = rScoreFreq[f].find(rScores[f][i]);
 570                                 if (it != rScoreFreq[f].end()) {
 571                                         rScoreFreq[f][rScores[f][i]]++;
 572                                 }else{
 573                                         rScoreFreq[f][rScores[f][i]] = 1;
 574                                 }
 575                         }
 576                 }
 577
 578                 //calculate rcumul
 579                 for(int a = 0; a < numComp; a++) {
 580                         float rcumul = 1.0000;
 581                         //this loop fills the cumulative maps and put 0.0000 in the score freq map to make it easier to print.
 582                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 583                                 //make rscoreFreq map and rCumul
 584                                 map<float,float>::iterator it2 = rScoreFreq[a].find(it->first);
 585                                 rCumul[a][it->first] = rcumul;
 586                                 //get percentage of random trees with that info
 587                                 if (it2 != rScoreFreq[a].end()) {  rScoreFreq[a][it->first] /= iters; rcumul-= it2->second;  }
 588                                 else { rScoreFreq[a][it->first] = 0.0000; } //no random trees with that score
 589                         }
 590                 }
 591
 592         }
 593         catch(exception& e) {
 594                 m->errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums");
 595                 exit(1);
 596         }
 597 }
 598
 599 /***********************************************************/
 600
 601
 602
 603
 604