bayesian.cpp

   1 /*
   2  *  bayesian.cpp
   3  *  Mothur
   4  *
   5  *  Created by westcott on 11/3/09.
   6  *  Copyright 2009 Schloss Lab. All rights reserved.
   7  *
   8  */
   9
  10 #include "bayesian.h"
  11 #include "kmer.hpp"
  12 #include "phylosummary.h"
  13 #include "referencedb.h"
  14 /**************************************************************************************************/
  15 Bayesian::Bayesian(string tfile, string tempFile, string method, int ksize, int cutoff, int i, int tid) :
  16 Classify(), kmerSize(ksize), confidenceThreshold(cutoff), iters(i)  {
  17         try {
  18                 ReferenceDB* rdb = ReferenceDB::getInstance();
  19
  20                 threadID = tid;
  21                 string baseName = tempFile;
  22
  23                 if (baseName == "saved") { baseName = rdb->getSavedReference(); }
  24
  25                 string baseTName = tfile;
  26                 if (baseTName == "saved") { baseTName = rdb->getSavedTaxonomy(); }
  27
  28                 /************calculate the probablity that each word will be in a specific taxonomy*************/
  29                 string tfileroot = baseTName.substr(0,baseTName.find_last_of(".")+1);
  30                 string tempfileroot = m->getRootName(m->getSimpleName(baseName));
  31                 string phyloTreeName = tfileroot + "tree.train";
  32                 string phyloTreeSumName = tfileroot + "tree.sum";
  33                 string probFileName = tfileroot + tempfileroot + char('0'+ kmerSize) + "mer.prob";
  34                 string probFileName2 = tfileroot + tempfileroot + char('0'+ kmerSize) + "mer.numNonZero";
  35
  36                 ofstream out;
  37                 ofstream out2;
  38
  39                 ifstream phyloTreeTest(phyloTreeName.c_str());
  40                 ifstream probFileTest2(probFileName2.c_str());
  41                 ifstream probFileTest(probFileName.c_str());
  42                 ifstream probFileTest3(phyloTreeSumName.c_str());
  43
  44                 int start = time(NULL);
  45
  46                 //if they are there make sure they were created after this release date
  47                 bool FilesGood = false;
  48                 if(probFileTest && probFileTest2 && phyloTreeTest && probFileTest3){
  49                         FilesGood = checkReleaseDate(probFileTest, probFileTest2, phyloTreeTest, probFileTest3);
  50                 }
  51
  52                 //if you want to save, but you dont need to calculate then just read
  53                 if (rdb->save && probFileTest && probFileTest2 && phyloTreeTest && probFileTest3 && FilesGood && (tempFile != "saved")) {
  54                         ifstream saveIn;
  55                         m->openInputFile(tempFile, saveIn);
  56
  57                         while (!saveIn.eof()) {
  58                                 Sequence temp(saveIn);
  59                                 m->gobble(saveIn);
  60
  61                                 rdb->referenceSeqs.push_back(temp);
  62                         }
  63                         saveIn.close();
  64                 }
  65
  66                 if(probFileTest && probFileTest2 && phyloTreeTest && probFileTest3 && FilesGood){
  67                         if (tempFile == "saved") { m->mothurOutEndLine();  m->mothurOut("Using sequences from " + rdb->getSavedReference() + " that are saved in memory.");     m->mothurOutEndLine(); }
  68
  69                         m->mothurOut("Reading template taxonomy...     "); cout.flush();
  70
  71                         phyloTree = new PhyloTree(phyloTreeTest, phyloTreeName);
  72
  73                         m->mothurOut("DONE."); m->mothurOutEndLine();
  74
  75                         genusNodes = phyloTree->getGenusNodes();
  76                         genusTotals = phyloTree->getGenusTotals();
  77
  78                         if (tfile == "saved") {
  79                                 m->mothurOutEndLine();  m->mothurOut("Using probabilties from " + rdb->getSavedTaxonomy() + " that are saved in memory...    ");        cout.flush();;
  80                                 wordGenusProb = rdb->wordGenusProb;
  81                         }else {
  82                                 m->mothurOut("Reading template probabilities...     "); cout.flush();
  83                                 readProbFile(probFileTest, probFileTest2, probFileName, probFileName2);
  84                         }
  85
  86                         //save probabilities
  87                         if (rdb->save) { rdb->wordGenusProb = wordGenusProb; }
  88                 }else{
  89
  90                         //create search database and names vector
  91                         generateDatabaseAndNames(tfile, tempFile, method, ksize, 0.0, 0.0, 0.0, 0.0);
  92
  93                         //prevents errors caused by creating shortcut files if you had an error in the sanity check.
  94                         if (m->control_pressed) {  m->mothurRemove(phyloTreeName);  m->mothurRemove(probFileName); m->mothurRemove(probFileName2); }
  95                         else{
  96                                 genusNodes = phyloTree->getGenusNodes();
  97                                 genusTotals = phyloTree->getGenusTotals();
  98
  99                                 m->mothurOut("Calculating template taxonomy tree...     "); cout.flush();
 100
 101                                 phyloTree->printTreeNodes(phyloTreeName);
 102
 103                                 m->mothurOut("DONE."); m->mothurOutEndLine();
 104
 105                                 m->mothurOut("Calculating template probabilities...     "); cout.flush();
 106
 107                                 numKmers = database->getMaxKmer() + 1;
 108
 109                                 //initialze probabilities
 110                                 wordGenusProb.resize(numKmers);
 111                         //cout << numKmers << '\t' << genusNodes.size() << endl;
 112                                 for (int j = 0; j < wordGenusProb.size(); j++) {        wordGenusProb[j].resize(genusNodes.size());             }
 113                         //cout << numKmers << '\t' << genusNodes.size() << endl;
 114                                 //ofstream out;
 115                                 //ofstream out2;
 116
 117                                 #ifdef USE_MPI
 118                                         int pid;
 119                                         MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 120
 121                                         if (pid == 0) {
 122                                 #endif
 123
 124
 125                                 //m->openOutputFile(probFileName, out);
 126
 127                                 //output mothur version
 128                                 //out << "#" << m->getVersion() << endl;
 129
 130                                 //out << numKmers << endl;
 131
 132                                 //m->openOutputFile(probFileName2, out2);
 133
 134                                 //output mothur version
 135                                 //out2 << "#" << m->getVersion() << endl;
 136
 137                                 #ifdef USE_MPI
 138                                         }
 139                                 #endif
 140
 141
 142                                 //for each word
 143                                 for (int i = 0; i < numKmers; i++) {
 144                                         if (m->control_pressed) {  break; }
 145
 146                                         #ifdef USE_MPI
 147                                                 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 148
 149                                                 if (pid == 0) {
 150                                         #endif
 151
 152                                         //out << i << '\t';
 153
 154                                         #ifdef USE_MPI
 155                                                 }
 156                                         #endif
 157
 158                                         vector<int> seqsWithWordi = database->getSequencesWithKmer(i);
 159
 160                                         map<int, int> count;
 161                                         for (int k = 0; k < genusNodes.size(); k++) {  count[genusNodes[k]] = 0;  }
 162
 163                                         //for each sequence with that word
 164                                         for (int j = 0; j < seqsWithWordi.size(); j++) {
 165                                                 int temp = phyloTree->getIndex(names[seqsWithWordi[j]]);
 166                                                 count[temp]++;  //increment count of seq in this genus who have this word
 167                                         }
 168
 169                                         //probabilityInTemplate = (# of seqs with that word in template + 0.50) / (total number of seqs in template + 1);
 170                                         float probabilityInTemplate = (seqsWithWordi.size() + 0.50) / (float) (names.size() + 1);
 171
 172                                         int numNotZero = 0;
 173                                         for (int k = 0; k < genusNodes.size(); k++) {
 174                                                 //probabilityInThisTaxonomy = (# of seqs with that word in this taxonomy + probabilityInTemplate) / (total number of seqs in this taxonomy + 1);
 175
 176
 177                                                 wordGenusProb[i][k] = log((count[genusNodes[k]] + probabilityInTemplate) / (float) (genusTotals[k] + 1));
 178
 179                                                 if (count[genusNodes[k]] != 0) {
 180                                                         #ifdef USE_MPI
 181                                                                 int pid;
 182                                                                 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 183
 184                                                                 if (pid == 0) {
 185                                                         #endif
 186
 187                                                         //out << k << '\t' << wordGenusProb[i][k] << '\t';
 188
 189                                                         #ifdef USE_MPI
 190                                                                 }
 191                                                         #endif
 192
 193                                                         numNotZero++;
 194                                                 }
 195                                         }
 196
 197                                         #ifdef USE_MPI
 198                                                 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 199
 200                                                 if (pid == 0) {
 201                                         #endif
 202
 203                                         //out << endl;
 204                                         //out2 << probabilityInTemplate << '\t' << numNotZero << endl;
 205
 206                                         #ifdef USE_MPI
 207                                                 }
 208                                         #endif
 209                                 }
 210
 211                                 #ifdef USE_MPI
 212                                         MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 213
 214                                         if (pid == 0) {
 215                                 #endif
 216
 217                                 //out.close();
 218                                 //out2.close();
 219
 220                                 #ifdef USE_MPI
 221                                         }
 222                                 #endif
 223
 224                                 //read in new phylotree with less info. - its faster
 225                                 ifstream phyloTreeTest(phyloTreeName.c_str());
 226                                 delete phyloTree;
 227
 228                                 phyloTree = new PhyloTree(phyloTreeTest, phyloTreeName);
 229
 230                                 //save probabilities
 231                                 if (rdb->save) { rdb->wordGenusProb = wordGenusProb; }
 232                         }
 233                 }
 234
 235                 m->mothurOut("DONE."); m->mothurOutEndLine();
 236                 m->mothurOut("It took " + toString(time(NULL) - start) + " seconds get probabilities. "); m->mothurOutEndLine();
 237         }
 238         catch(exception& e) {
 239                 m->errorOut(e, "Bayesian", "Bayesian");
 240                 exit(1);
 241         }
 242 }
 243 /**************************************************************************************************/
 244 Bayesian::~Bayesian() {
 245         try {
 246
 247                  delete phyloTree;
 248                  if (database != NULL) {  delete database; }
 249         }
 250         catch(exception& e) {
 251                 m->errorOut(e, "Bayesian", "~Bayesian");
 252                 exit(1);
 253         }
 254 }
 255
 256 /**************************************************************************************************/
 257 string Bayesian::getTaxonomy(Sequence* seq) {
 258         try {
 259                 string tax = "";
 260                 Kmer kmer(kmerSize);
 261
 262                 //get words contained in query
 263                 //getKmerString returns a string where the index in the string is hte kmer number
 264                 //and the character at that index can be converted to be the number of times that kmer was seen
 265
 266                 string queryKmerString = kmer.getKmerString(seq->getUnaligned());
 267
 268                 vector<int> queryKmers;
 269                 for (int i = 0; i < queryKmerString.length()-1; i++) {  // the -1 is to ignore any kmer with an N in it
 270                         if (queryKmerString[i] != '!') { //this kmer is in the query
 271                                 queryKmers.push_back(i);
 272                         }
 273                 }
 274
 275                 if (queryKmers.size() == 0) {  m->mothurOut(seq->getName() + "is bad."); m->mothurOutEndLine(); return "bad seq"; }
 276
 277
 278                 int index = getMostProbableTaxonomy(queryKmers);
 279
 280                 if (m->control_pressed) { return tax; }
 281
 282                 //bootstrap - to set confidenceScore
 283                 int numToSelect = queryKmers.size() / 8;
 284
 285                 tax = bootstrapResults(queryKmers, index, numToSelect);
 286
 287                 return tax;
 288         }
 289         catch(exception& e) {
 290                 m->errorOut(e, "Bayesian", "getTaxonomy");
 291                 exit(1);
 292         }
 293 }
 294 /**************************************************************************************************/
 295 string Bayesian::bootstrapResults(vector<int> kmers, int tax, int numToSelect) {
 296         try {
 297
 298                 map<int, int> confidenceScores;
 299
 300                 //initialize confidences to 0
 301                 int seqIndex = tax;
 302                 TaxNode seq = phyloTree->get(tax);
 303                 confidenceScores[tax] = 0;
 304
 305                 while (seq.level != 0) { //while you are not at the root
 306                         seqIndex = seq.parent;
 307                         confidenceScores[seqIndex] = 0;
 308                         seq = phyloTree->get(seq.parent);
 309                 }
 310
 311                 map<int, int>::iterator itBoot;
 312                 map<int, int>::iterator itBoot2;
 313                 map<int, int>::iterator itConvert;
 314
 315                 for (int i = 0; i < iters; i++) {
 316                         if (m->control_pressed) { return "control"; }
 317
 318                         vector<int> temp;
 319                         for (int j = 0; j < numToSelect; j++) {
 320                                 int index = int(rand() % kmers.size());
 321
 322                                 //add word to temp
 323                                 temp.push_back(kmers[index]);
 324                         }
 325
 326                         //get taxonomy
 327                         int newTax = getMostProbableTaxonomy(temp);
 328                         //int newTax = 1;
 329                         TaxNode taxonomyTemp = phyloTree->get(newTax);
 330
 331                         //add to confidence results
 332                         while (taxonomyTemp.level != 0) { //while you are not at the root
 333                                 itBoot2 = confidenceScores.find(newTax); //is this a classification we already have a count on
 334
 335                                 if (itBoot2 != confidenceScores.end()) { //this is a classification we need a confidence for
 336                                         (itBoot2->second)++;
 337                                 }
 338
 339                                 newTax = taxonomyTemp.parent;
 340                                 taxonomyTemp = phyloTree->get(newTax);
 341                         }
 342
 343                 }
 344
 345                 string confidenceTax = "";
 346                 simpleTax = "";
 347
 348                 int seqTaxIndex = tax;
 349                 TaxNode seqTax = phyloTree->get(tax);
 350
 351                 while (seqTax.level != 0) { //while you are not at the root
 352
 353                                 itBoot2 = confidenceScores.find(seqTaxIndex); //is this a classification we already have a count on
 354
 355                                 int confidence = 0;
 356                                 if (itBoot2 != confidenceScores.end()) { //already in confidence scores
 357                                         confidence = itBoot2->second;
 358                                 }
 359
 360                                 if (((confidence/(float)iters) * 100) >= confidenceThreshold) {
 361                                         confidenceTax = seqTax.name + "(" + toString(((confidence/(float)iters) * 100)) + ");" + confidenceTax;
 362                                         simpleTax = seqTax.name + ";" + simpleTax;
 363                                 }
 364
 365                                 seqTaxIndex = seqTax.parent;
 366                                 seqTax = phyloTree->get(seqTax.parent);
 367                 }
 368
 369                 if (confidenceTax == "") { confidenceTax = "unclassified;"; simpleTax = "unclassified;"; }
 370                 return confidenceTax;
 371
 372         }
 373         catch(exception& e) {
 374                 m->errorOut(e, "Bayesian", "bootstrapResults");
 375                 exit(1);
 376         }
 377 }
 378 /**************************************************************************************************/
 379 int Bayesian::getMostProbableTaxonomy(vector<int> queryKmer) {
 380         try {
 381                 int indexofGenus = 0;
 382
 383                 double maxProbability = -1000000.0;
 384                 //find taxonomy with highest probability that this sequence is from it
 385
 386
 387 //              cout << genusNodes.size() << endl;
 388
 389
 390                 for (int k = 0; k < genusNodes.size(); k++) {
 391                         //for each taxonomy calc its probability
 392
 393                         double prob = 0.0000;
 394                         for (int i = 0; i < queryKmer.size(); i++) {
 395                                 prob += wordGenusProb[queryKmer[i]][k];
 396                         }
 397
 398 //                      cout << phyloTree->get(genusNodes[k]).name << '\t' << prob << endl;
 399
 400                         //is this the taxonomy with the greatest probability?
 401                         if (prob > maxProbability) {
 402                                 indexofGenus = genusNodes[k];
 403                                 maxProbability = prob;
 404                         }
 405                 }
 406
 407
 408                 return indexofGenus;
 409         }
 410         catch(exception& e) {
 411                 m->errorOut(e, "Bayesian", "getMostProbableTaxonomy");
 412                 exit(1);
 413         }
 414 }
 415 /*************************************************************************************************
 416 map<string, int> Bayesian::parseTaxMap(string newTax) {
 417         try{
 418
 419                 map<string, int> parsed;
 420
 421                 newTax = newTax.substr(0, newTax.length()-1);  //get rid of last ';'
 422
 423                 //parse taxonomy
 424                 string individual;
 425                 while (newTax.find_first_of(';') != -1) {
 426                         individual = newTax.substr(0,newTax.find_first_of(';'));
 427                         newTax = newTax.substr(newTax.find_first_of(';')+1, newTax.length());
 428                         parsed[individual] = 1;
 429                 }
 430
 431                 //get last one
 432                 parsed[newTax] = 1;
 433
 434                 return parsed;
 435
 436         }
 437         catch(exception& e) {
 438                 m->errorOut(e, "Bayesian", "parseTax");
 439                 exit(1);
 440         }
 441 }
 442 /**************************************************************************************************/
 443 void Bayesian::readProbFile(ifstream& in, ifstream& inNum, string inName, string inNumName) {
 444         try{
 445
 446                 #ifdef USE_MPI
 447
 448                         int pid, num, num2, processors;
 449                         vector<unsigned long long> positions;
 450                         vector<unsigned long long> positions2;
 451
 452                         MPI_Status status;
 453                         MPI_File inMPI;
 454                         MPI_File inMPI2;
 455                         MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 456                         MPI_Comm_size(MPI_COMM_WORLD, &processors);
 457                         int tag = 2001;
 458
 459                         char inFileName[1024];
 460                         strcpy(inFileName, inNumName.c_str());
 461
 462                         char inFileName2[1024];
 463                         strcpy(inFileName2, inName.c_str());
 464
 465                         MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);  //comm, filename, mode, info, filepointer
 466                         MPI_File_open(MPI_COMM_WORLD, inFileName2, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI2);  //comm, filename, mode, info, filepointer
 467
 468                         if (pid == 0) {
 469                                 positions = m->setFilePosEachLine(inNumName, num);
 470                                 positions2 = m->setFilePosEachLine(inName, num2);
 471
 472                                 for(int i = 1; i < processors; i++) {
 473                                         MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
 474                                         MPI_Send(&positions[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
 475
 476                                         MPI_Send(&num2, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
 477                                         MPI_Send(&positions2[0], (num2+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
 478                                 }
 479
 480                         }else{
 481                                 MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
 482                                 positions.resize(num+1);
 483                                 MPI_Recv(&positions[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
 484
 485                                 MPI_Recv(&num2, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
 486                                 positions2.resize(num2+1);
 487                                 MPI_Recv(&positions2[0], (num2+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
 488                         }
 489
 490                         //read version
 491                         int length = positions2[1] - positions2[0];
 492                         char* buf5 = new char[length];
 493
 494                         MPI_File_read_at(inMPI2, positions2[0], buf5, length, MPI_CHAR, &status);
 495                         delete buf5;
 496
 497                         //read numKmers
 498                         length = positions2[2] - positions2[1];
 499                         char* buf = new char[length];
 500
 501                         MPI_File_read_at(inMPI2, positions2[1], buf, length, MPI_CHAR, &status);
 502
 503                         string tempBuf = buf;
 504                         if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
 505                         delete buf;
 506
 507                         istringstream iss (tempBuf,istringstream::in);
 508                         iss >> numKmers;
 509
 510                         //initialze probabilities
 511                         wordGenusProb.resize(numKmers);
 512
 513                         for (int j = 0; j < wordGenusProb.size(); j++) {        wordGenusProb[j].resize(genusNodes.size());             }
 514
 515                         int kmer, name;
 516                         vector<int> numbers; numbers.resize(numKmers);
 517                         float prob;
 518                         vector<float> zeroCountProb; zeroCountProb.resize(numKmers);
 519
 520                         //read version
 521                         length = positions[1] - positions[0];
 522                         char* buf6 = new char[length];
 523
 524                         MPI_File_read_at(inMPI2, positions[0], buf6, length, MPI_CHAR, &status);
 525                         delete buf6;
 526
 527                         //read file
 528                         for(int i=1;i<num;i++){
 529                                 //read next sequence
 530                                 length = positions[i+1] - positions[i];
 531                                 char* buf4 = new char[length];
 532
 533                                 MPI_File_read_at(inMPI, positions[i], buf4, length, MPI_CHAR, &status);
 534
 535                                 tempBuf = buf4;
 536                                 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
 537                                 delete buf4;
 538
 539                                 istringstream iss (tempBuf,istringstream::in);
 540                                 iss >> zeroCountProb[i] >> numbers[i];
 541                         }
 542
 543                         MPI_File_close(&inMPI);
 544
 545                         for(int i=2;i<num2;i++){
 546                                 //read next sequence
 547                                 length = positions2[i+1] - positions2[i];
 548                                 char* buf4 = new char[length];
 549
 550                                 MPI_File_read_at(inMPI2, positions2[i], buf4, length, MPI_CHAR, &status);
 551
 552                                 tempBuf = buf4;
 553                                 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
 554                                 delete buf4;
 555
 556                                 istringstream iss (tempBuf,istringstream::in);
 557
 558                                 iss >> kmer;
 559
 560                                 //set them all to zero value
 561                                 for (int i = 0; i < genusNodes.size(); i++) {
 562                                         wordGenusProb[kmer][i] = log(zeroCountProb[kmer] / (float) (genusTotals[i]+1));
 563                                 }
 564
 565                                 //get probs for nonzero values
 566                                 for (int i = 0; i < numbers[kmer]; i++) {
 567                                         iss >> name >> prob;
 568                                         wordGenusProb[kmer][name] = prob;
 569                                 }
 570
 571                         }
 572                         MPI_File_close(&inMPI2);
 573                         MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
 574                 #else
 575                         //read version
 576                         string line = m->getline(in); m->gobble(in);
 577
 578                         in >> numKmers; m->gobble(in);
 579                         //cout << threadID << '\t' << line << '\t' << numKmers << &in << '\t' << &inNum << '\t' << genusNodes.size() << endl;
 580                         //initialze probabilities
 581                         wordGenusProb.resize(numKmers);
 582
 583                         for (int j = 0; j < wordGenusProb.size(); j++) {        wordGenusProb[j].resize(genusNodes.size());             }
 584
 585                         int kmer, name, count;  count = 0;
 586                         vector<int> num; num.resize(numKmers);
 587                         float prob;
 588                         vector<float> zeroCountProb; zeroCountProb.resize(numKmers);
 589
 590                         //read version
 591                         string line2 = m->getline(inNum); m->gobble(inNum);
 592                 //cout << threadID << '\t' << line2 << '\t' << this << endl;
 593                         while (inNum) {
 594                                 inNum >> zeroCountProb[count] >> num[count];
 595                                 count++;
 596                                 m->gobble(inNum);
 597                                 //cout << threadID << '\t' << count << endl;
 598                         }
 599                         inNum.close();
 600                 //cout << threadID << '\t' << "here1 " << &wordGenusProb << '\t' << &num << endl; //
 601                 //cout << threadID << '\t' << &genusTotals << '\t' << endl;
 602                 //cout << threadID << '\t' << genusNodes.size() << endl;
 603                         while(in) {
 604                                 in >> kmer;
 605                         //cout << threadID << '\t' << kmer << endl;
 606                                 //set them all to zero value
 607                                 for (int i = 0; i < genusNodes.size(); i++) {
 608                                         wordGenusProb[kmer][i] = log(zeroCountProb[kmer] / (float) (genusTotals[i]+1));
 609                                 }
 610                         //cout << threadID << '\t' << num[kmer] << "here" << endl;
 611                                 //get probs for nonzero values
 612                                 for (int i = 0; i < num[kmer]; i++) {
 613                                         in >> name >> prob;
 614                                         wordGenusProb[kmer][name] = prob;
 615                                 }
 616
 617                                 m->gobble(in);
 618                         }
 619                         in.close();
 620                 //cout << threadID << '\t' << "here" << endl;
 621                 #endif
 622         }
 623         catch(exception& e) {
 624                 m->errorOut(e, "Bayesian", "readProbFile");
 625                 exit(1);
 626         }
 627 }
 628 /**************************************************************************************************/
 629 bool Bayesian::checkReleaseDate(ifstream& file1, ifstream& file2, ifstream& file3, ifstream& file4) {
 630         try {
 631
 632                 bool good = true;
 633
 634                 vector<string> lines;
 635                 lines.push_back(m->getline(file1));
 636                 lines.push_back(m->getline(file2));
 637                 lines.push_back(m->getline(file3));
 638                 lines.push_back(m->getline(file4));
 639
 640                 //before we added this check
 641                 if ((lines[0][0] != '#') || (lines[1][0] != '#') || (lines[2][0] != '#') || (lines[3][0] != '#')) {  good = false;  }
 642                 else {
 643                         //rip off #
 644                         for (int i = 0; i < lines.size(); i++) { lines[i] = lines[i].substr(1);  }
 645
 646                         //get mothurs current version
 647                         string version = m->getVersion();
 648
 649                         vector<string> versionVector;
 650                         m->splitAtChar(version, versionVector, '.');
 651
 652                         //check each files version
 653                         for (int i = 0; i < lines.size(); i++) {
 654                                 vector<string> linesVector;
 655                                 m->splitAtChar(lines[i], linesVector, '.');
 656
 657                                 if (versionVector.size() != linesVector.size()) { good = false; break; }
 658                                 else {
 659                                         for (int j = 0; j < versionVector.size(); j++) {
 660                                                 int num1, num2;
 661                                                 convert(versionVector[j], num1);
 662                                                 convert(linesVector[j], num2);
 663
 664                                                 //if mothurs version is newer than this files version, then we want to remake it
 665                                                 if (num1 > num2) {  good = false; break;  }
 666                                         }
 667                                 }
 668
 669                                 if (!good) { break; }
 670                         }
 671                 }
 672
 673                 if (!good) {  file1.close(); file2.close(); file3.close(); file4.close();  }
 674                 else { file1.seekg(0);  file2.seekg(0);  file3.seekg(0);  file4.seekg(0);  }
 675
 676                 return good;
 677         }
 678         catch(exception& e) {
 679                 m->errorOut(e, "Bayesian", "checkReleaseDate");
 680                 exit(1);
 681         }
 682 }
 683 /**************************************************************************************************/
 684
 685
 686
 687
 688
 689