2 * treegroupscommand.cpp
5 * Created by Sarah Westcott on 4/8/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "treegroupscommand.h"
11 #include "sharedjabund.h"
12 #include "sharedsorabund.h"
13 #include "sharedjclass.h"
14 #include "sharedsorclass.h"
15 #include "sharedjest.h"
16 #include "sharedsorest.h"
17 #include "sharedthetayc.h"
18 #include "sharedthetan.h"
19 #include "sharedmorisitahorn.h"
20 #include "sharedbraycurtis.h"
23 //**********************************************************************************************************************
25 TreeGroupCommand::TreeGroupCommand(string option) {
27 globaldata = GlobalData::getInstance();
34 //allow user to run help
35 if(option == "help") { validCalculator = new ValidCalculators(); help(); abort = true; }
38 //valid paramters for this command
39 string Array[] = {"label","calc","groups", "phylip", "column", "name", "precision","cutoff","outputdir","inputdir"};
40 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
42 OptionParser parser(option);
43 map<string, string> parameters = parser. getParameters();
45 ValidParameters validParameter;
46 map<string, string>::iterator it;
48 //check to make sure all parameters are valid for command
49 for (it = parameters.begin(); it != parameters.end(); it++) {
50 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
53 //if the user changes the input directory command factory will send this info to us in the output parameter
54 string inputDir = validParameter.validFile(parameters, "inputdir", false);
55 if (inputDir == "not found"){ inputDir = ""; }
58 it = parameters.find("phylip");
59 //user has given a template file
60 if(it != parameters.end()){
61 path = m->hasPath(it->second);
62 //if the user has not given a path then, add inputdir. else leave path alone.
63 if (path == "") { parameters["phylip"] = inputDir + it->second; }
66 it = parameters.find("column");
67 //user has given a template file
68 if(it != parameters.end()){
69 path = m->hasPath(it->second);
70 //if the user has not given a path then, add inputdir. else leave path alone.
71 if (path == "") { parameters["column"] = inputDir + it->second; }
74 it = parameters.find("name");
75 //user has given a template file
76 if(it != parameters.end()){
77 path = m->hasPath(it->second);
78 //if the user has not given a path then, add inputdir. else leave path alone.
79 if (path == "") { parameters["name"] = inputDir + it->second; }
83 format = globaldata->getFormat();
86 phylipfile = validParameter.validFile(parameters, "phylip", true);
87 if (phylipfile == "not open") { abort = true; }
88 else if (phylipfile == "not found") { phylipfile = ""; }
89 else { globaldata->newRead(); format = "phylip"; globaldata->setPhylipFile(phylipfile); }
91 columnfile = validParameter.validFile(parameters, "column", true);
92 if (columnfile == "not open") { abort = true; }
93 else if (columnfile == "not found") { columnfile = ""; }
94 else { globaldata->newRead(); format = "column"; globaldata->setColumnFile(columnfile); }
96 namefile = validParameter.validFile(parameters, "name", true);
97 if (namefile == "not open") { abort = true; }
98 else if (namefile == "not found") { namefile = ""; }
99 else { globaldata->setNameFile(namefile); }
101 //error checking on files
102 if ((globaldata->getSharedFile() == "") && ((phylipfile == "") && (columnfile == ""))) { m->mothurOut("You must run the read.otu command or provide a distance file before running the tree.shared command."); m->mothurOutEndLine(); abort = true; }
103 else if ((phylipfile != "") && (columnfile != "")) { m->mothurOut("When running the tree.shared command with a distance file you may not use both the column and the phylip parameters."); m->mothurOutEndLine(); abort = true; }
105 if (columnfile != "") {
106 if (namefile == "") { m->mothurOut("You need to provide a namefile if you are going to use the column format."); m->mothurOutEndLine(); abort = true; }
109 //check for optional parameter and set defaults
110 // ...at some point should added some additional type checking...
111 label = validParameter.validFile(parameters, "label", false);
112 if (label == "not found") { label = ""; }
114 if(label != "all") { m->splitAtDash(label, labels); allLines = 0; }
115 else { allLines = 1; }
118 //if the user has not specified any labels use the ones from read.otu
120 allLines = globaldata->allLines;
121 labels = globaldata->labels;
124 groups = validParameter.validFile(parameters, "groups", false);
125 if (groups == "not found") { groups = ""; }
127 m->splitAtDash(groups, Groups);
128 globaldata->Groups = Groups;
131 calc = validParameter.validFile(parameters, "calc", false);
132 if (calc == "not found") { calc = "jclass-thetayc"; }
134 if (calc == "default") { calc = "jclass-thetayc"; }
136 m->splitAtDash(calc, Estimators);
139 temp = validParameter.validFile(parameters, "precision", false); if (temp == "not found") { temp = "100"; }
140 convert(temp, precision);
142 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; }
143 convert(temp, cutoff);
144 cutoff += (5 / (precision * 10.0));
146 //if the user changes the output directory command factory will send this info to us in the output parameter
147 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
149 outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it
153 if (abort == false) {
155 validCalculator = new ValidCalculators();
157 if (format == "sharedfile") {
159 for (i=0; i<Estimators.size(); i++) {
160 if (validCalculator->isValidCalculator("treegroup", Estimators[i]) == true) {
161 if (Estimators[i] == "jabund") {
162 treeCalculators.push_back(new JAbund());
163 }else if (Estimators[i] == "sorabund") {
164 treeCalculators.push_back(new SorAbund());
165 }else if (Estimators[i] == "jclass") {
166 treeCalculators.push_back(new Jclass());
167 }else if (Estimators[i] == "sorclass") {
168 treeCalculators.push_back(new SorClass());
169 }else if (Estimators[i] == "jest") {
170 treeCalculators.push_back(new Jest());
171 }else if (Estimators[i] == "sorest") {
172 treeCalculators.push_back(new SorEst());
173 }else if (Estimators[i] == "thetayc") {
174 treeCalculators.push_back(new ThetaYC());
175 }else if (Estimators[i] == "thetan") {
176 treeCalculators.push_back(new ThetaN());
177 }else if (Estimators[i] == "morisitahorn") {
178 treeCalculators.push_back(new MorHorn());
179 }else if (Estimators[i] == "braycurtis") {
180 treeCalculators.push_back(new BrayCurtis());
189 catch(exception& e) {
190 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
195 //**********************************************************************************************************************
197 void TreeGroupCommand::help(){
199 m->mothurOut("The tree.shared command creates a .tre to represent the similiarity between groups or sequences.\n");
200 m->mothurOut("The tree.shared command can only be executed after a successful read.otu command or by providing a distance file.\n");
201 m->mothurOut("The tree.shared command parameters are groups, calc, phylip, column, name, cutoff, precision and label.\n");
202 m->mothurOut("The groups parameter allows you to specify which of the groups in your groupfile you would like included used.\n");
203 m->mothurOut("The group names are separated by dashes. The label allow you to select what distance levels you would like trees created for, and are also separated by dashes.\n");
204 m->mothurOut("The phylip or column parameter are required if you do not run the read.otu command first, and only one may be used. If you use a column file the name filename is required. \n");
205 m->mothurOut("If you do not provide a cutoff value 10.00 is assumed. If you do not provide a precision value then 100 is assumed.\n");
206 m->mothurOut("The tree.shared command should be in the following format: tree.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n");
207 m->mothurOut("Example tree.shared(groups=A-B-C, calc=jabund-sorabund).\n");
208 m->mothurOut("The default value for groups is all the groups in your groupfile.\n");
209 m->mothurOut("The default value for calc is jclass-thetayc.\n");
210 m->mothurOut("The tree.shared command outputs a .tre file for each calculator you specify at each distance you choose.\n");
211 validCalculator->printCalc("treegroup", cout);
212 m->mothurOut("Or the tree.shared command can be in the following format: tree.shared(phylip=yourPhylipFile).\n");
213 m->mothurOut("Example tree.shared(phylip=abrecovery.dist).\n");
214 m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n");
216 catch(exception& e) {
217 m->errorOut(e, "TreeGroupCommand", "help");
223 //**********************************************************************************************************************
225 TreeGroupCommand::~TreeGroupCommand(){
226 globaldata->Groups.clear();
227 if (abort == false) {
229 if (format == "sharedfile") { delete read; delete input; globaldata->ginput = NULL; }
230 else { delete readMatrix; delete matrix; delete list; }
231 delete tmap; globaldata->gTreemap = NULL;
232 delete validCalculator;
237 //**********************************************************************************************************************
239 int TreeGroupCommand::execute(){
242 if (abort == true) { return 0; }
244 if (format == "sharedfile") {
245 //if the users entered no valid calculators don't execute command
246 if (treeCalculators.size() == 0) { m->mothurOut("You have given no valid calculators."); m->mothurOutEndLine(); return 0; }
248 if (globaldata->gGroupmap != NULL) { delete globaldata->gGroupmap; globaldata->gGroupmap = NULL; }
250 read = new ReadOTUFile(globaldata->inputFileName);
251 read->read(&*globaldata);
253 input = globaldata->ginput;
254 lookup = input->getSharedRAbundVectors();
255 lastLabel = lookup[0]->getLabel();
257 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); return 0; }
259 //used in tree constructor
260 globaldata->runParse = false;
262 //create treemap class from groupmap for tree class to use
263 tmap = new TreeMap();
264 tmap->makeSim(globaldata->gGroupmap);
265 globaldata->gTreemap = tmap;
267 //clear globaldatas old tree names if any
268 globaldata->Treenames.clear();
270 //fills globaldatas tree names
271 globaldata->Treenames = globaldata->Groups;
273 if (m->control_pressed) { return 0; }
278 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
281 filename = globaldata->inputFileName;
283 if (format == "column") { readMatrix = new ReadColumnMatrix(filename); }
284 else if (format == "phylip") { readMatrix = new ReadPhylipMatrix(filename); }
286 readMatrix->setCutoff(cutoff);
289 nameMap = new NameAssignment(namefile);
296 readMatrix->read(nameMap);
297 list = readMatrix->getListVector();
298 matrix = readMatrix->getMatrix();
301 tmap = new TreeMap();
303 if (m->control_pressed) { return 0; }
306 globaldata->gTreemap = tmap;
308 globaldata->Groups = tmap->namesOfGroups;
310 //clear globaldatas old tree names if any
311 globaldata->Treenames.clear();
313 //fills globaldatas tree names
314 globaldata->Treenames = globaldata->Groups;
316 //used in tree constructor
317 globaldata->runParse = false;
319 if (m->control_pressed) { return 0; }
323 if (m->control_pressed) { return 0; }
325 //create a new filename
326 outputFile = outputDir + m->getRootName(m->getSimpleName(globaldata->inputFileName)) + "tre";
327 outputNames.push_back(outputFile);
331 if (m->control_pressed) { return 0; }
333 m->mothurOut("Tree complete. "); m->mothurOutEndLine();
337 //reset groups parameter
338 globaldata->Groups.clear();
340 m->mothurOutEndLine();
341 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
342 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
343 m->mothurOutEndLine();
347 catch(exception& e) {
348 m->errorOut(e, "TreeGroupCommand", "execute");
352 //**********************************************************************************************************************
354 int TreeGroupCommand::createTree(){
359 //do merges and create tree structure by setting parents and children
360 //there are numGroups - 1 merges to do
361 for (int i = 0; i < (numGroups - 1); i++) {
362 float largest = -1000.0;
364 if (m->control_pressed) { delete t; return 1; }
367 //find largest value in sims matrix by searching lower triangle
368 for (int j = 1; j < simMatrix.size(); j++) {
369 for (int k = 0; k < j; k++) {
370 if (simMatrix[j][k] > largest) { largest = simMatrix[j][k]; row = j; column = k; }
374 //set non-leaf node info and update leaves to know their parents
376 t->tree[numGroups + i].setChildren(index[row], index[column]);
379 t->tree[index[row]].setParent(numGroups + i);
380 t->tree[index[column]].setParent(numGroups + i);
382 //blength = distance / 2;
383 float blength = ((1.0 - largest) / 2);
386 t->tree[index[row]].setBranchLength(blength - t->tree[index[row]].getLengthToLeaves());
387 t->tree[index[column]].setBranchLength(blength - t->tree[index[column]].getLengthToLeaves());
389 //set your length to leaves to your childs length plus branchlength
390 t->tree[numGroups + i].setLengthToLeaves(t->tree[index[row]].getLengthToLeaves() + t->tree[index[row]].getBranchLength());
394 index[row] = numGroups+i;
395 index[column] = numGroups+i;
397 //remove highest value that caused the merge.
398 simMatrix[row][column] = -1000.0;
399 simMatrix[column][row] = -1000.0;
401 //merge values in simsMatrix
402 for (int n = 0; n < simMatrix.size(); n++) {
403 //row becomes merge of 2 groups
404 simMatrix[row][n] = (simMatrix[row][n] + simMatrix[column][n]) / 2;
405 simMatrix[n][row] = simMatrix[row][n];
407 simMatrix[column][n] = -1000.0;
408 simMatrix[n][column] = -1000.0;
412 //adjust tree to make sure root to tip length is .5
413 int root = t->findRoot();
414 t->tree[root].setBranchLength((0.5 - t->tree[root].getLengthToLeaves()));
419 if (m->control_pressed) { delete t; return 1; }
422 t->createNewickFile(outputFile);
427 if (m->control_pressed) { remove(outputFile.c_str()); outputNames.pop_back(); return 1; }
432 catch(exception& e) {
433 m->errorOut(e, "TreeGroupCommand", "createTree");
437 /***********************************************************/
438 void TreeGroupCommand::printSims(ostream& out) {
441 //output column headers
443 //for (int i = 0; i < lookup.size(); i++) { out << lookup[i]->getGroup() << '\t'; }
447 for (int m = 0; m < simMatrix.size(); m++) {
448 //out << lookup[m]->getGroup() << '\t';
449 for (int n = 0; n < simMatrix.size(); n++) {
450 out << simMatrix[m][n] << '\t';
456 catch(exception& e) {
457 m->errorOut(e, "TreeGroupCommand", "printSims");
461 /***********************************************************/
462 int TreeGroupCommand::makeSimsDist() {
464 numGroups = list->size();
468 for (int g = 0; g < numGroups; g++) { index[g] = g; }
470 //initialize simMatrix
472 simMatrix.resize(numGroups);
473 for (int k = 0; k < simMatrix.size(); k++) {
474 for (int j = 0; j < simMatrix.size(); j++) {
475 simMatrix[k].push_back(0.0);
479 //go through sparse matrix and fill sims
480 //go through each cell in the sparsematrix
481 for(MatData currentCell = matrix->begin(); currentCell != matrix->end(); currentCell++){
482 //similairity = -(distance-1)
483 simMatrix[currentCell->row][currentCell->column] = -(currentCell->dist -1.0);
484 simMatrix[currentCell->column][currentCell->row] = -(currentCell->dist -1.0);
486 if (m->control_pressed) { return 1; }
492 catch(exception& e) {
493 m->errorOut(e, "TreeGroupCommand", "makeSimsDist");
498 /***********************************************************/
499 int TreeGroupCommand::makeSimsShared() {
501 set<string> processedLabels;
502 set<string> userLabels = labels;
504 //as long as you are not at the end of the file or done wih the lines you want
505 while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) {
506 if (m->control_pressed) { for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; } return 1; }
508 if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){
509 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
512 processedLabels.insert(lookup[0]->getLabel());
513 userLabels.erase(lookup[0]->getLabel());
516 if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) {
517 string saveLabel = lookup[0]->getLabel();
519 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
520 lookup = input->getSharedRAbundVectors(lastLabel);
522 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
525 processedLabels.insert(lookup[0]->getLabel());
526 userLabels.erase(lookup[0]->getLabel());
528 //restore real lastlabel to save below
529 lookup[0]->setLabel(saveLabel);
532 lastLabel = lookup[0]->getLabel();
534 //get next line to process
535 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
536 lookup = input->getSharedRAbundVectors();
539 if (m->control_pressed) { for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; } return 1; }
541 //output error messages about any remaining user labels
542 set<string>::iterator it;
543 bool needToRun = false;
544 for (it = userLabels.begin(); it != userLabels.end(); it++) {
545 m->mothurOut("Your file does not include the label " + *it);
546 if (processedLabels.count(lastLabel) != 1) {
547 m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine();
550 m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine();
554 //run last label if you need to
555 if (needToRun == true) {
556 for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
557 lookup = input->getSharedRAbundVectors(lastLabel);
559 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
561 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
564 for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; }
568 catch(exception& e) {
569 m->errorOut(e, "TreeGroupCommand", "makeSimsShared");
574 /***********************************************************/
575 int TreeGroupCommand::process(vector<SharedRAbundVector*> thisLookup) {
578 vector<SharedRAbundVector*> subset;
579 numGroups = thisLookup.size();
581 //for each calculator
582 for(int i = 0 ; i < treeCalculators.size(); i++) {
583 //initialize simMatrix
585 simMatrix.resize(numGroups);
586 for (int k = 0; k < simMatrix.size(); k++) {
587 for (int j = 0; j < simMatrix.size(); j++) {
588 simMatrix[k].push_back(0.0);
594 for (int g = 0; g < numGroups; g++) { index[g] = g; }
596 //create a new filename
597 outputFile = outputDir + m->getRootName(m->getSimpleName(globaldata->inputFileName)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".tre";
598 outputNames.push_back(outputFile);
600 for (int k = 0; k < thisLookup.size(); k++) {
601 for (int l = k; l < thisLookup.size(); l++) {
602 if (k != l) { //we dont need to similiarity of a groups to itself
603 //get estimated similarity between 2 groups
605 subset.clear(); //clear out old pair of sharedrabunds
606 //add new pair of sharedrabunds
607 subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]);
609 data = treeCalculators[i]->getValues(subset); //saves the calculator outputs
610 //cout << thisLookup[k]->getGroup() << '\t' << thisLookup[l]->getGroup() << '\t' << (1.0 - data[0]) << endl;
611 if (m->control_pressed) { return 1; }
613 //save values in similarity matrix
614 simMatrix[k][l] = data[0];
615 simMatrix[l][k] = data[0];
620 if (m->control_pressed) { return 1; }
621 //creates tree from similarity matrix and write out file
624 if (m->control_pressed) { return 1; }
630 catch(exception& e) {
631 m->errorOut(e, "TreeGroupCommand", "process");
635 /***********************************************************/