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 UnifracWeightedCommand::UnifracWeightedCommand(string option) {
  14         try {
  15                 globaldata = GlobalData::getInstance();
  16                 abort = false;
  17                 Groups.clear();
  18
  19                 //allow user to run help
  20                 if(option == "help") { help(); abort = true; }
  21
  22                 else {
  23                         //valid paramters for this command
  24                         string Array[] =  {"groups","iters","distance","random","outputdir","inputdir"};
  25                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  26
  27                         OptionParser parser(option);
  28                         map<string,string> parameters=parser.getParameters();
  29
  30                         ValidParameters validParameter;
  31
  32                         //check to make sure all parameters are valid for command
  33                         for (map<string,string>::iterator it = parameters.begin(); it != parameters.end(); it++) {
  34                                 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) {  abort = true;  }
  35                         }
  36
  37                         if (globaldata->gTree.size() == 0) {//no trees were read
  38                                 m->mothurOut("You must execute the read.tree command, before you may execute the unifrac.weighted command."); m->mothurOutEndLine(); abort = true;  }
  39
  40                         //if the user changes the output directory command factory will send this info to us in the output parameter
  41                         outputDir = validParameter.validFile(parameters, "outputdir", false);           if (outputDir == "not found"){
  42                                 outputDir = "";
  43                                 outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it
  44                         }
  45
  46                         //check for optional parameter and set defaults
  47                         // ...at some point should added some additional type checking...
  48                         groups = validParameter.validFile(parameters, "groups", false);
  49                         if (groups == "not found") { groups = ""; }
  50                         else {
  51                                 m->splitAtDash(groups, Groups);
  52                                 globaldata->Groups = Groups;
  53                         }
  54
  55                         itersString = validParameter.validFile(parameters, "iters", false);                     if (itersString == "not found") { itersString = "1000"; }
  56                         convert(itersString, iters);
  57
  58                         string temp = validParameter.validFile(parameters, "distance", false);                  if (temp == "not found") { temp = "false"; }
  59                         phylip = m->isTrue(temp);
  60
  61                         temp = validParameter.validFile(parameters, "random", false);                                   if (temp == "not found") { temp = "F"; }
  62                         random = m->isTrue(temp);
  63
  64                         if (!random) {  iters = 0;  } //turn off random calcs
  65
  66
  67                         if (abort == false) {
  68                                 T = globaldata->gTree;
  69                                 tmap = globaldata->gTreemap;
  70                                 sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary";
  71                                 m->openOutputFile(sumFile, outSum);
  72                                 outputNames.push_back(sumFile);
  73
  74                                 util = new SharedUtil();
  75                                 string s; //to make work with setgroups
  76                                 util->setGroups(globaldata->Groups, tmap->namesOfGroups, s, numGroups, "weighted");     //sets the groups the user wants to analyze
  77                                 util->getCombos(groupComb, globaldata->Groups, numComp);
  78
  79                                 weighted = new Weighted(tmap);
  80
  81                         }
  82                 }
  83
  84
  85         }
  86         catch(exception& e) {
  87                 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
  88                 exit(1);
  89         }
  90 }
  91 //**********************************************************************************************************************
  92
  93 void UnifracWeightedCommand::help(){
  94         try {
  95                 m->mothurOut("The unifrac.weighted command can only be executed after a successful read.tree command.\n");
  96                 m->mothurOut("The unifrac.weighted command parameters are groups, iters, distance and random.  No parameters are required.\n");
  97                 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");
  98                 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");
  99                 m->mothurOut("The distance parameter allows you to create a distance file from the results. The default is false.\n");
 100                 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");
 101                 m->mothurOut("The unifrac.weighted command should be in the following format: unifrac.weighted(groups=yourGroups, iters=yourIters).\n");
 102                 m->mothurOut("Example unifrac.weighted(groups=A-B-C, iters=500).\n");
 103                 m->mothurOut("The default value for groups is all the groups in your groupfile, and iters is 1000.\n");
 104                 m->mothurOut("The unifrac.weighted command output two files: .weighted and .wsummary their descriptions are in the manual.\n");
 105                 m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n");
 106         }
 107         catch(exception& e) {
 108                 m->errorOut(e, "UnifracWeightedCommand", "help");
 109                 exit(1);
 110         }
 111 }
 112
 113 /***********************************************************/
 114 int UnifracWeightedCommand::execute() {
 115         try {
 116
 117                 if (abort == true) { return 0; }
 118
 119                 int start = time(NULL);
 120
 121                 Progress* reading;
 122                 if (random) {   reading = new Progress("Comparing to random:", iters);  }
 123
 124                 //get weighted for users tree
 125                 userData.resize(numComp,0);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 126                 randomData.resize(numComp,0); //data[0] = weightedscore AB, data[1] = weightedscore AC...
 127
 128                 //create new tree with same num nodes and leaves as users
 129                 randT = new Tree();
 130
 131                 //get weighted scores for users trees
 132                 for (int i = 0; i < T.size(); i++) {
 133
 134                         if (m->control_pressed) {
 135                                 delete randT;
 136                                 if (random) { delete reading; }
 137                                 outSum.close();
 138                                 for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 139                                 return 0;
 140                         }
 141
 142                         counter = 0;
 143                         rScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 144                         uScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 145
 146                         if (random) {
 147                                 output = new ColumnFile(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted", itersString);
 148                                 outputNames.push_back(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted");
 149                         }
 150
 151                         userData = weighted->getValues(T[i]);  //userData[0] = weightedscore
 152
 153                         if (m->control_pressed) {
 154                                 delete randT;
 155                                 if (random) { delete reading; delete output; }
 156                                 outSum.close();
 157                                 for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 158                                 return 0;
 159                         }
 160
 161
 162                         //save users score
 163                         for (int s=0; s<numComp; s++) {
 164                                 //add users score to vector of user scores
 165                                 uScores[s].push_back(userData[s]);
 166
 167                                 //save users tree score for summary file
 168                                 utreeScores.push_back(userData[s]);
 169                         }
 170
 171                         //get scores for random trees
 172                         for (int j = 0; j < iters; j++) {
 173                                 int count = 0;
 174                                 for (int r=0; r<numGroups; r++) {
 175                                         for (int l = r+1; l < numGroups; l++) {
 176                                                 //copy T[i]'s info.
 177                                                 randT->getCopy(T[i]);
 178
 179                                                 //create a random tree with same topology as T[i], but different labels
 180                                                 randT->assembleRandomUnifracTree(globaldata->Groups[r], globaldata->Groups[l]);
 181
 182                                                 if (m->control_pressed) {
 183                                                         delete randT;
 184                                                         if (random) { delete reading; delete output; }
 185                                                         outSum.close();
 186                                                         for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 187                                                         return 0;
 188                                                 }
 189
 190
 191                                                 //get wscore of random tree
 192                                                 randomData = weighted->getValues(randT, globaldata->Groups[r], globaldata->Groups[l]);
 193
 194                                                 if (m->control_pressed) {
 195                                                         delete randT;
 196                                                         if (random) { delete reading; delete output; }
 197                                                         outSum.close();
 198                                                         for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 199                                                         return 0;
 200                                                 }
 201
 202                                                 //save scores
 203                                                 rScores[count].push_back(randomData[0]);
 204                                                 count++;
 205                                         }
 206                                 }
 207
 208                                 //update progress bar
 209                                 reading->update(j);
 210
 211                         }
 212
 213                         //removeValidScoresDuplicates();
 214                         //find the signifigance of the score for summary file
 215                         if (random) {
 216                                 for (int f = 0; f < numComp; f++) {
 217                                         //sort random scores
 218                                         sort(rScores[f].begin(), rScores[f].end());
 219
 220                                         //the index of the score higher than yours is returned
 221                                         //so if you have 1000 random trees the index returned is 100
 222                                         //then there are 900 trees with a score greater then you.
 223                                         //giving you a signifigance of 0.900
 224                                         int index = findIndex(userData[f], f);    if (index == -1) { m->mothurOut("error in UnifracWeightedCommand"); m->mothurOutEndLine(); exit(1); } //error code
 225
 226                                         //the signifigance is the number of trees with the users score or higher
 227                                         WScoreSig.push_back((iters-index)/(float)iters);
 228                                 }
 229
 230                                 //out << "Tree# " << i << endl;
 231                                 calculateFreqsCumuls();
 232                                 printWeightedFile();
 233
 234                                 delete output;
 235                         }
 236
 237                         //clear data
 238                         rScores.clear();
 239                         uScores.clear();
 240                         validScores.clear();
 241                 }
 242
 243
 244                 if (m->control_pressed) {
 245                                 delete randT;
 246                                 if (random) { delete reading;  }
 247                                 outSum.close();
 248                                 for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 249                                 return 0;
 250                 }
 251
 252                 //finish progress bar
 253                 if (random) {   reading->finish();      delete reading;         }
 254
 255                 printWSummaryFile();
 256
 257                 if (phylip) {   createPhylipFile();             }
 258
 259                 //clear out users groups
 260                 globaldata->Groups.clear();
 261
 262                 delete randT;
 263
 264                 if (m->control_pressed) {
 265                         for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 266                         return 0;
 267                 }
 268
 269                 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run unifrac.weighted."); m->mothurOutEndLine();
 270
 271                 m->mothurOutEndLine();
 272                 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
 273                 for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
 274                 m->mothurOutEndLine();
 275
 276                 return 0;
 277
 278         }
 279         catch(exception& e) {
 280                 m->errorOut(e, "UnifracWeightedCommand", "execute");
 281                 exit(1);
 282         }
 283 }
 284 /***********************************************************/
 285 void UnifracWeightedCommand::printWeightedFile() {
 286         try {
 287                 vector<double> data;
 288                 vector<string> tags;
 289                 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
 290
 291                 for(int a = 0; a < numComp; a++) {
 292                         output->initFile(groupComb[a], tags);
 293                         //print each line
 294                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 295                                 data.push_back(it->first);  data.push_back(rScoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
 296                                 output->output(data);
 297                                 data.clear();
 298                         }
 299                         output->resetFile();
 300                 }
 301         }
 302         catch(exception& e) {
 303                 m->errorOut(e, "UnifracWeightedCommand", "printWeightedFile");
 304                 exit(1);
 305         }
 306 }
 307
 308
 309 /***********************************************************/
 310 void UnifracWeightedCommand::printWSummaryFile() {
 311         try {
 312                 //column headers
 313                 outSum << "Tree#" << '\t' << "Groups" << '\t' << "WScore" << '\t' << "WSig" <<  endl;
 314                 m->mothurOut("Tree#\tGroups\tWScore\tWSig"); m->mothurOutEndLine();
 315
 316                 //format output
 317                 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
 318
 319                 //print each line
 320                 int count = 0;
 321                 for (int i = 0; i < T.size(); i++) {
 322                         for (int j = 0; j < numComp; j++) {
 323                                 if (random) {
 324                                         if (WScoreSig[count] > (1/(float)iters)) {
 325                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 326                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 327                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" +  toString(WScoreSig[count])); m->mothurOutEndLine();
 328                                         }else{
 329                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 330                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 331                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" +  toString((1/float(iters)))); m->mothurOutEndLine();
 332                                         }
 333                                 }else{
 334                                         outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 335                                         cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 336                                         m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00"); m->mothurOutEndLine();
 337                                 }
 338                                 count++;
 339                         }
 340                 }
 341                 outSum.close();
 342         }
 343         catch(exception& e) {
 344                 m->errorOut(e, "UnifracWeightedCommand", "printWSummaryFile");
 345                 exit(1);
 346         }
 347 }
 348 /***********************************************************/
 349 void UnifracWeightedCommand::createPhylipFile() {
 350         try {
 351                 int count = 0;
 352                 //for each tree
 353                 for (int i = 0; i < T.size(); i++) {
 354
 355                         string phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted.dist";
 356                         outputNames.push_back(phylipFileName);
 357                         ofstream out;
 358                         m->openOutputFile(phylipFileName, out);
 359
 360                         //output numSeqs
 361                         out << globaldata->Groups.size() << endl;
 362
 363                         //make matrix with scores in it
 364                         vector< vector<float> > dists;  dists.resize(globaldata->Groups.size());
 365                         for (int i = 0; i < globaldata->Groups.size(); i++) {
 366                                 dists[i].resize(globaldata->Groups.size(), 0.0);
 367                         }
 368
 369                         //flip it so you can print it
 370                         for (int r=0; r<globaldata->Groups.size(); r++) {
 371                                 for (int l = r+1; l < globaldata->Groups.size(); l++) {
 372                                         dists[r][l] = utreeScores[count];
 373                                         dists[l][r] = utreeScores[count];
 374                                         count++;
 375                                 }
 376                         }
 377
 378                         //output to file
 379                         for (int r=0; r<globaldata->Groups.size(); r++) {
 380                                 //output name
 381                                 string name = globaldata->Groups[r];
 382                                 if (name.length() < 10) { //pad with spaces to make compatible
 383                                         while (name.length() < 10) {  name += " ";  }
 384                                 }
 385                                 out << name << '\t';
 386
 387                                 //output distances
 388                                 for (int l = 0; l < r; l++) {   out  << dists[r][l] << '\t';  }
 389                                 out << endl;
 390                         }
 391                         out.close();
 392                 }
 393         }
 394         catch(exception& e) {
 395                 m->errorOut(e, "UnifracWeightedCommand", "createPhylipFile");
 396                 exit(1);
 397         }
 398 }
 399 /***********************************************************/
 400 int UnifracWeightedCommand::findIndex(float score, int index) {
 401         try{
 402                 for (int i = 0; i < rScores[index].size(); i++) {
 403                         if (rScores[index][i] >= score) {       return i;       }
 404                 }
 405                 return rScores[index].size();
 406         }
 407         catch(exception& e) {
 408                 m->errorOut(e, "UnifracWeightedCommand", "findIndex");
 409                 exit(1);
 410         }
 411 }
 412
 413 /***********************************************************/
 414
 415 void UnifracWeightedCommand::calculateFreqsCumuls() {
 416         try {
 417                 //clear out old tree values
 418                 rScoreFreq.clear();
 419                 rScoreFreq.resize(numComp);
 420                 rCumul.clear();
 421                 rCumul.resize(numComp);
 422                 validScores.clear();
 423
 424                 //calculate frequency
 425                 for (int f = 0; f < numComp; f++) {
 426                         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...
 427                                 validScores[rScores[f][i]] = rScores[f][i];
 428                                 map<float,float>::iterator it = rScoreFreq[f].find(rScores[f][i]);
 429                                 if (it != rScoreFreq[f].end()) {
 430                                         rScoreFreq[f][rScores[f][i]]++;
 431                                 }else{
 432                                         rScoreFreq[f][rScores[f][i]] = 1;
 433                                 }
 434                         }
 435                 }
 436
 437                 //calculate rcumul
 438                 for(int a = 0; a < numComp; a++) {
 439                         float rcumul = 1.0000;
 440                         //this loop fills the cumulative maps and put 0.0000 in the score freq map to make it easier to print.
 441                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 442                                 //make rscoreFreq map and rCumul
 443                                 map<float,float>::iterator it2 = rScoreFreq[a].find(it->first);
 444                                 rCumul[a][it->first] = rcumul;
 445                                 //get percentage of random trees with that info
 446                                 if (it2 != rScoreFreq[a].end()) {  rScoreFreq[a][it->first] /= iters; rcumul-= it2->second;  }
 447                                 else { rScoreFreq[a][it->first] = 0.0000; } //no random trees with that score
 448                         }
 449                 }
 450
 451         }
 452         catch(exception& e) {
 453                 m->errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums");
 454                 exit(1);
 455         }
 456 }
 457
 458 /***********************************************************/
 459
 460
 461
 462
 463