unifracweightedcommand.h

   1 #ifndef UNIFRACWEIGHTEDCOMMAND_H
   2 #define UNIFRACWEIGHTEDCOMMAND_H
   3
   4 /*
   5  *  unifracweightedcommand.h
   6  *  Mothur
   7  *
   8  *  Created by Sarah Westcott on 2/9/09.
   9  *  Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
  10  *
  11  */
  12
  13 #include "command.hpp"
  14 #include "weighted.h"
  15 #include "counttable.h"
  16 #include "progress.hpp"
  17 #include "sharedutilities.h"
  18 #include "fileoutput.h"
  19 #include "readtree.h"
  20
  21 class UnifracWeightedCommand : public Command {
  22
  23         public:
  24                 UnifracWeightedCommand(string);
  25                 UnifracWeightedCommand();
  26                 ~UnifracWeightedCommand() {}
  27
  28                 vector<string> setParameters();
  29                 string getCommandName()                 { return "unifrac.weighted";            }
  30                 string getCommandCategory()             { return "Hypothesis Testing";          }
  31
  32         string getHelpString();
  33     string getOutputPattern(string);
  34                 string getCitation() { return "Lozupone CA, Hamady M, Kelley ST, Knight R (2007). Quantitative and qualitative beta diversity measures lead to different insights into factors that structure microbial communities. Appl Environ Microbiol 73: 1576-85. \nhttp://www.mothur.org/wiki/Unifrac.weighted"; }
  35                 string getDescription()         { return "generic tests that describes whether two or more communities have the same structure"; }
  36
  37                 int execute();
  38                 void help() { m->mothurOut(getHelpString()); }
  39
  40         private:
  41                 struct linePair {
  42                         int start;
  43                         int num;
  44                         linePair(int i, int j) : start(i), num(j) {}
  45                 };
  46                 vector<linePair> lines;
  47         CountTable* ct;
  48                 FileOutput* output;
  49                 vector<Tree*> T;           //user trees
  50                 vector<double> utreeScores;  //user tree unweighted scores
  51                 vector<double> WScoreSig;  //tree weighted score signifigance when compared to random trees - percentage of random trees with that score or lower.
  52                 vector<string> groupComb; // AB. AC, BC...
  53                 string sumFile, outputDir;
  54                 int iters, numGroups, numComp, counter;
  55                 vector< vector<double> > rScores;  //vector<weighted scores for random trees.> each group comb has an entry
  56                 vector< vector<double> > uScores;  //vector<weighted scores for user trees.> each group comb has an entry
  57                 vector< map<double, double> > rScoreFreq;  //map <weighted score, number of random trees with that score.> -vector entry for each combination.
  58                 vector< map<double, double> > rCumul;  //map <weighted score, cumulative percentage of number of random trees with that score or higher.> -vector entry for each c
  59                 map<double, double>  validScores;  //map contains scores from random
  60
  61                 bool abort, phylip, random, includeRoot, subsample, consensus;
  62                 string groups, itersString, outputForm, treefile, groupfile, namefile, countfile;
  63                 vector<string> Groups, outputNames; //holds groups to be used
  64                 int processors, subsampleSize, subsampleIters;
  65                 ofstream outSum;
  66                 map<string, string> nameMap;
  67
  68                 void printWSummaryFile();
  69                 void printWeightedFile();
  70                 void createPhylipFile();
  71                 //void removeValidScoresDuplicates();
  72                 int findIndex(float, int);
  73                 void calculateFreqsCumuls();
  74                 int createProcesses(Tree*,  vector< vector<string> >,  vector< vector<double> >&);
  75                 int driver(Tree*, vector< vector<string> >, int, int,  vector< vector<double> >&);
  76         int runRandomCalcs(Tree*, vector<double>);
  77         vector<Tree*> buildTrees(vector< vector<double> >&, int, CountTable&);
  78         int getConsensusTrees(vector< vector<double> >&, int);
  79         int getAverageSTDMatrices(vector< vector<double> >&, int);
  80
  81 };
  82
  83 /***********************************************************************/
  84 struct weightedRandomData {
  85     int start;
  86         int num;
  87         MothurOut* m;
  88     vector< vector<double> > scores;
  89     vector< vector<string> > namesOfGroupCombos;
  90     Tree* t;
  91     CountTable* ct;
  92     bool includeRoot;
  93
  94         weightedRandomData(){}
  95         weightedRandomData(MothurOut* mout, int st, int en, vector< vector<string> > ngc, Tree* tree, CountTable* count, bool ir, vector< vector<double> > sc) {
  96         m = mout;
  97                 start = st;
  98                 num = en;
  99         namesOfGroupCombos = ngc;
 100         t = tree;
 101         ct = count;
 102         includeRoot = ir;
 103         scores = sc;
 104         }
 105 };
 106
 107 /**************************************************************************************************/
 108 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
 109 #else
 110 static DWORD WINAPI MyWeightedRandomThreadFunction(LPVOID lpParam){
 111         weightedRandomData* pDataArray;
 112         pDataArray = (weightedRandomData*)lpParam;
 113         try {
 114
 115         Tree* randT = new Tree(pDataArray->ct);
 116
 117         Weighted weighted(pDataArray->includeRoot);
 118
 119                 for (int h = pDataArray->start; h < (pDataArray->start+pDataArray->num); h++) {
 120
 121                         if (pDataArray->m->control_pressed) { return 0; }
 122
 123                         //initialize weighted score
 124                         string groupA = pDataArray->namesOfGroupCombos[h][0];
 125                         string groupB = pDataArray->namesOfGroupCombos[h][1];
 126
 127                         //copy T[i]'s info.
 128                         randT->getCopy(pDataArray->t);
 129
 130                         //create a random tree with same topology as T[i], but different labels
 131                         randT->assembleRandomUnifracTree(groupA, groupB);
 132
 133                         if (pDataArray->m->control_pressed) { delete randT;  return 0;  }
 134
 135                         //get wscore of random tree
 136                         EstOutput randomData = weighted.getValues(randT, groupA, groupB);
 137
 138                         if (pDataArray->m->control_pressed) { delete randT;  return 0;  }
 139
 140                         //save scores
 141                         pDataArray->scores[h].push_back(randomData[0]);
 142                 }
 143
 144                 delete randT;
 145
 146         return 0;
 147     }
 148         catch(exception& e) {
 149                 pDataArray->m->errorOut(e, "Weighted", "MyWeightedRandomThreadFunction");
 150                 exit(1);
 151         }
 152 }
 153 #endif
 154
 155
 156 #endif