2 * unifracweightedcommand.cpp
5 * Created by Sarah Westcott on 2/9/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "unifracweightedcommand.h"
12 /***********************************************************/
13 UnifracWeightedCommand::UnifracWeightedCommand(string option) {
15 globaldata = GlobalData::getInstance();
19 //allow user to run help
20 if(option == "help") { help(); abort = true; }
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)));
27 OptionParser parser(option);
28 map<string,string> parameters=parser.getParameters();
30 ValidParameters validParameter;
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; }
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; }
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"){
43 outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it
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 = ""; }
51 m->splitAtDash(groups, Groups);
52 globaldata->Groups = Groups;
55 itersString = validParameter.validFile(parameters, "iters", false); if (itersString == "not found") { itersString = "1000"; }
56 convert(itersString, iters);
58 string temp = validParameter.validFile(parameters, "distance", false); if (temp == "not found") { temp = "false"; }
59 phylip = m->isTrue(temp);
61 temp = validParameter.validFile(parameters, "random", false); if (temp == "not found") { temp = "F"; }
62 random = m->isTrue(temp);
64 if (!random) { iters = 0; } //turn off random calcs
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);
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);
79 weighted = new Weighted(tmap);
87 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
91 //**********************************************************************************************************************
93 void UnifracWeightedCommand::help(){
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");
107 catch(exception& e) {
108 m->errorOut(e, "UnifracWeightedCommand", "help");
113 /***********************************************************/
114 int UnifracWeightedCommand::execute() {
117 if (abort == true) { return 0; }
119 int start = time(NULL);
122 if (random) { reading = new Progress("Comparing to random:", iters); }
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...
128 //create new tree with same num nodes and leaves as users
131 //get weighted scores for users trees
132 for (int i = 0; i < T.size(); i++) {
134 if (m->control_pressed) {
136 if (random) { delete reading; }
138 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
143 rScores.resize(numComp); //data[0] = weightedscore AB, data[1] = weightedscore AC...
144 uScores.resize(numComp); //data[0] = weightedscore AB, data[1] = weightedscore AC...
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");
151 userData = weighted->getValues(T[i]); //userData[0] = weightedscore
153 if (m->control_pressed) {
155 if (random) { delete reading; delete output; }
157 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
163 for (int s=0; s<numComp; s++) {
164 //add users score to vector of user scores
165 uScores[s].push_back(userData[s]);
167 //save users tree score for summary file
168 utreeScores.push_back(userData[s]);
171 //get scores for random trees
172 for (int j = 0; j < iters; j++) {
174 for (int r=0; r<numGroups; r++) {
175 for (int l = r+1; l < numGroups; l++) {
177 randT->getCopy(T[i]);
179 //create a random tree with same topology as T[i], but different labels
180 randT->assembleRandomUnifracTree(globaldata->Groups[r], globaldata->Groups[l]);
182 if (m->control_pressed) {
184 if (random) { delete reading; delete output; }
186 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
191 //get wscore of random tree
192 randomData = weighted->getValues(randT, globaldata->Groups[r], globaldata->Groups[l]);
194 if (m->control_pressed) {
196 if (random) { delete reading; delete output; }
198 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
203 rScores[count].push_back(randomData[0]);
208 //update progress bar
213 //removeValidScoresDuplicates();
214 //find the signifigance of the score for summary file
216 for (int f = 0; f < numComp; f++) {
218 sort(rScores[f].begin(), rScores[f].end());
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
226 //the signifigance is the number of trees with the users score or higher
227 WScoreSig.push_back((iters-index)/(float)iters);
230 //out << "Tree# " << i << endl;
231 calculateFreqsCumuls();
244 if (m->control_pressed) {
246 if (random) { delete reading; }
248 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
252 //finish progress bar
253 if (random) { reading->finish(); delete reading; }
257 if (phylip) { createPhylipFile(); }
259 //clear out users groups
260 globaldata->Groups.clear();
264 if (m->control_pressed) {
265 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); }
269 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run unifrac.weighted."); m->mothurOutEndLine();
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();
279 catch(exception& e) {
280 m->errorOut(e, "UnifracWeightedCommand", "execute");
284 /***********************************************************/
285 void UnifracWeightedCommand::printWeightedFile() {
289 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
291 for(int a = 0; a < numComp; a++) {
292 output->initFile(groupComb[a], tags);
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);
302 catch(exception& e) {
303 m->errorOut(e, "UnifracWeightedCommand", "printWeightedFile");
309 /***********************************************************/
310 void UnifracWeightedCommand::printWSummaryFile() {
313 outSum << "Tree#" << '\t' << "Groups" << '\t' << "WScore" << '\t' << "WSig" << endl;
314 m->mothurOut("Tree#\tGroups\tWScore\tWSig"); m->mothurOutEndLine();
317 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
321 for (int i = 0; i < T.size(); i++) {
322 for (int j = 0; j < numComp; j++) {
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();
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();
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();
343 catch(exception& e) {
344 m->errorOut(e, "UnifracWeightedCommand", "printWSummaryFile");
348 /***********************************************************/
349 void UnifracWeightedCommand::createPhylipFile() {
353 for (int i = 0; i < T.size(); i++) {
355 string phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted.dist";
356 outputNames.push_back(phylipFileName);
358 m->openOutputFile(phylipFileName, out);
361 out << globaldata->Groups.size() << endl;
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);
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];
379 for (int r=0; r<globaldata->Groups.size(); r++) {
381 string name = globaldata->Groups[r];
382 if (name.length() < 10) { //pad with spaces to make compatible
383 while (name.length() < 10) { name += " "; }
388 for (int l = 0; l < r; l++) { out << dists[r][l] << '\t'; }
394 catch(exception& e) {
395 m->errorOut(e, "UnifracWeightedCommand", "createPhylipFile");
399 /***********************************************************/
400 int UnifracWeightedCommand::findIndex(float score, int index) {
402 for (int i = 0; i < rScores[index].size(); i++) {
403 if (rScores[index][i] >= score) { return i; }
405 return rScores[index].size();
407 catch(exception& e) {
408 m->errorOut(e, "UnifracWeightedCommand", "findIndex");
413 /***********************************************************/
415 void UnifracWeightedCommand::calculateFreqsCumuls() {
417 //clear out old tree values
419 rScoreFreq.resize(numComp);
421 rCumul.resize(numComp);
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]]++;
432 rScoreFreq[f][rScores[f][i]] = 1;
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
452 catch(exception& e) {
453 m->errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums");
458 /***********************************************************/