5 * Created by Sarah Westcott on 1/22/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
12 /*****************************************************************/
13 Tree::Tree(int num, TreeMap* t) : tmap(t) {
15 m = MothurOut::getInstance();
18 numNodes = 2*numLeaves - 1;
20 tree.resize(numNodes);
23 m->errorOut(e, "Tree", "Tree - numNodes");
27 /*****************************************************************/
28 Tree::Tree(string g) { //do not use tree generated by this its just to extract the treenames, its a chicken before the egg thing that needs to be revisited.
30 m = MothurOut::getInstance();
31 parseTreeFile(); m->runParse = false;
34 m->errorOut(e, "Tree", "Tree - just parse");
38 /*****************************************************************/
39 Tree::Tree(TreeMap* t) : tmap(t) {
41 m = MothurOut::getInstance();
43 if (m->runParse == true) { parseTreeFile(); m->runParse = false; }
44 //for(int i = 0; i < globaldata->Treenames.size(); i++) { cout << i << '\t' << globaldata->Treenames[i] << endl; }
45 numLeaves = m->Treenames.size();
46 numNodes = 2*numLeaves - 1;
48 tree.resize(numNodes);
50 //initialize groupNodeInfo
51 for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
52 groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
55 //initialize tree with correct number of nodes, name and group info.
56 for (int i = 0; i < numNodes; i++) {
57 //initialize leaf nodes
58 if (i <= (numLeaves-1)) {
59 tree[i].setName(m->Treenames[i]);
62 string group = tmap->getGroup(m->Treenames[i]);
64 vector<string> tempGroups; tempGroups.push_back(group);
65 tree[i].setGroup(tempGroups);
66 groupNodeInfo[group].push_back(i);
68 //set pcount and pGroup for groupname to 1.
69 tree[i].pcount[group] = 1;
70 tree[i].pGroups[group] = 1;
72 //Treemap knows name, group and index to speed up search
73 tmap->setIndex(m->Treenames[i], i);
75 //intialize non leaf nodes
76 }else if (i > (numLeaves-1)) {
78 vector<string> tempGroups;
79 tree[i].setGroup(tempGroups);
85 m->errorOut(e, "Tree", "Tree");
89 /*****************************************************************/
90 Tree::Tree(TreeMap* t, vector< vector<double> >& sims) : tmap(t) {
92 m = MothurOut::getInstance();
94 if (m->runParse == true) { parseTreeFile(); m->runParse = false; }
95 numLeaves = m->Treenames.size();
96 numNodes = 2*numLeaves - 1;
98 tree.resize(numNodes);
100 //initialize groupNodeInfo
101 for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
102 groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
105 //initialize tree with correct number of nodes, name and group info.
106 for (int i = 0; i < numNodes; i++) {
107 //initialize leaf nodes
108 if (i <= (numLeaves-1)) {
109 tree[i].setName(m->Treenames[i]);
112 string group = tmap->getGroup(m->Treenames[i]);
114 vector<string> tempGroups; tempGroups.push_back(group);
115 tree[i].setGroup(tempGroups);
116 groupNodeInfo[group].push_back(i);
118 //set pcount and pGroup for groupname to 1.
119 tree[i].pcount[group] = 1;
120 tree[i].pGroups[group] = 1;
122 //Treemap knows name, group and index to speed up search
123 tmap->setIndex(m->Treenames[i], i);
125 //intialize non leaf nodes
126 }else if (i > (numLeaves-1)) {
128 vector<string> tempGroups;
129 tree[i].setGroup(tempGroups);
133 //build tree from matrix
135 map<int, int> indexes; //maps row in simMatrix to vector index in the tree
136 for (int g = 0; g < numLeaves; g++) { indexes[g] = g; }
138 //do merges and create tree structure by setting parents and children
139 //there are numGroups - 1 merges to do
140 for (int i = 0; i < (numLeaves - 1); i++) {
141 float largest = -1000.0;
143 if (m->control_pressed) { break; }
146 //find largest value in sims matrix by searching lower triangle
147 for (int j = 1; j < sims.size(); j++) {
148 for (int k = 0; k < j; k++) {
149 if (sims[j][k] > largest) { largest = sims[j][k]; row = j; column = k; }
153 //set non-leaf node info and update leaves to know their parents
155 tree[numLeaves + i].setChildren(indexes[row], indexes[column]);
158 tree[indexes[row]].setParent(numLeaves + i);
159 tree[indexes[column]].setParent(numLeaves + i);
161 //blength = distance / 2;
162 float blength = ((1.0 - largest) / 2);
165 tree[indexes[row]].setBranchLength(blength - tree[indexes[row]].getLengthToLeaves());
166 tree[indexes[column]].setBranchLength(blength - tree[indexes[column]].getLengthToLeaves());
168 //set your length to leaves to your childs length plus branchlength
169 tree[numLeaves + i].setLengthToLeaves(tree[indexes[row]].getLengthToLeaves() + tree[indexes[row]].getBranchLength());
173 indexes[row] = numLeaves+i;
174 indexes[column] = numLeaves+i;
176 //remove highest value that caused the merge.
177 sims[row][column] = -1000.0;
178 sims[column][row] = -1000.0;
180 //merge values in simsMatrix
181 for (int n = 0; n < sims.size(); n++) {
182 //row becomes merge of 2 groups
183 sims[row][n] = (sims[row][n] + sims[column][n]) / 2;
184 sims[n][row] = sims[row][n];
186 sims[column][n] = -1000.0;
187 sims[n][column] = -1000.0;
191 //adjust tree to make sure root to tip length is .5
192 int root = findRoot();
193 tree[root].setBranchLength((0.5 - tree[root].getLengthToLeaves()));
196 catch(exception& e) {
197 m->errorOut(e, "Tree", "Tree");
201 /*****************************************************************/
203 /*****************************************************************/
204 void Tree::addNamesToCounts(map<string, string> nameMap) {
206 //ex. seq1 seq2,seq3,se4
212 //before this function seq1.pcount = pasture -> 1
213 //after seq1.pcount = pasture -> 2, forest -> 1, ocean -> 1
215 //before this function seq1.pgroups = pasture -> 1
216 //after seq1.pgroups = pasture -> 1 since that is the dominant group
219 //go through each leaf and update its pcounts and pgroups
223 for (int i = 0; i < numLeaves; i++) {
225 string name = tree[i].getName();
227 map<string, string>::iterator itNames = nameMap.find(name);
229 if (itNames == nameMap.end()) { m->mothurOut(name + " is not in your name file, please correct."); m->mothurOutEndLine(); exit(1); }
231 vector<string> dupNames;
232 m->splitAtComma(nameMap[name], dupNames);
234 map<string, int>::iterator itCounts;
236 set<string> groupsAddedForThisNode;
237 for (int j = 0; j < dupNames.size(); j++) {
239 string group = tmap->getGroup(dupNames[j]);
241 if (dupNames[j] != name) {//you already added yourself in the constructor
243 if (groupsAddedForThisNode.count(group) == 0) { groupNodeInfo[group].push_back(i); groupsAddedForThisNode.insert(group); } //if you have not already added this node for this group, then add it
246 itCounts = tree[i].pcount.find(group);
247 if (itCounts == tree[i].pcount.end()) { //new group, add it
248 tree[i].pcount[group] = 1;
250 tree[i].pcount[group]++;
254 itCounts = tree[i].pGroups.find(group);
255 if (itCounts == tree[i].pGroups.end()) { //new group, add it
256 tree[i].pGroups[group] = 1;
258 tree[i].pGroups[group]++;
262 if(tree[i].pGroups[group] > maxPars){
263 maxPars = tree[i].pGroups[group];
265 }else { groupsAddedForThisNode.insert(group); } //add it so you don't add it to groupNodeInfo again
268 if (maxPars > 1) { //then we have some more dominant groups
269 //erase all the groups that are less than maxPars because you found a more dominant group.
270 for(it=tree[i].pGroups.begin();it!=tree[i].pGroups.end();){
271 if(it->second < maxPars){
272 tree[i].pGroups.erase(it++);
275 //set one remaining groups to 1
276 for(it=tree[i].pGroups.begin();it!=tree[i].pGroups.end();it++){
277 tree[i].pGroups[it->first] = 1;
281 //update groups to reflect all the groups this node represents
282 vector<string> nodeGroups;
283 map<string, int>::iterator itGroups;
284 for (itGroups = tree[i].pcount.begin(); itGroups != tree[i].pcount.end(); itGroups++) {
285 nodeGroups.push_back(itGroups->first);
287 tree[i].setGroup(nodeGroups);
293 //cout << "addNamesToCounts\t" << (B - A) / CLOCKS_PER_SEC << endl;
296 catch(exception& e) {
297 m->errorOut(e, "Tree", "addNamesToCounts");
301 /*****************************************************************/
302 int Tree::getIndex(string searchName) {
304 //Treemap knows name, group and index to speed up search
305 // getIndex function will return the vector index or -1 if seq is not found.
306 int index = tmap->getIndex(searchName);
310 catch(exception& e) {
311 m->errorOut(e, "Tree", "getIndex");
315 /*****************************************************************/
317 void Tree::setIndex(string searchName, int index) {
319 //set index in treemap
320 tmap->setIndex(searchName, index);
322 catch(exception& e) {
323 m->errorOut(e, "Tree", "setIndex");
327 /*****************************************************************/
328 int Tree::assembleTree(map<string, string> nameMap) {
333 //if user has given a names file we want to include that info in the pgroups and pcount info.
334 if(nameMap.size() != 0) { addNamesToCounts(nameMap); }
336 //build the pGroups in non leaf nodes to be used in the parsimony calcs.
337 for (int i = numLeaves; i < numNodes; i++) {
338 if (m->control_pressed) { return 1; }
340 tree[i].pGroups = (mergeGroups(i));
341 tree[i].pcount = (mergeGcounts(i));
346 catch(exception& e) {
347 m->errorOut(e, "Tree", "assembleTree");
351 /*****************************************************************
352 int Tree::assembleTree(string n) {
355 //build the pGroups in non leaf nodes to be used in the parsimony calcs.
356 for (int i = numLeaves; i < numNodes; i++) {
357 if (m->control_pressed) { return 1; }
359 tree[i].pGroups = (mergeGroups(i));
360 tree[i].pcount = (mergeGcounts(i));
363 //cout << "assembleTree\t" << (B-A) / CLOCKS_PER_SEC << endl;
366 catch(exception& e) {
367 m->errorOut(e, "Tree", "assembleTree");
371 /*****************************************************************/
372 //assumes leaf node names are in groups and no names file - used by indicator command
373 void Tree::getSubTree(Tree* Ctree, vector<string> Groups) {
376 //copy Tree since we are going to destroy it
377 Tree* copy = new Tree(tmap);
378 copy->getCopy(Ctree);
379 map<string, string> empty;
380 copy->assembleTree(empty);
382 //we want to select some of the leaf nodes to create the output tree
383 //go through the input Tree starting at parents of leaves
384 for (int i = 0; i < numNodes; i++) {
386 //initialize leaf nodes
387 if (i <= (numLeaves-1)) {
388 tree[i].setName(Groups[i]);
391 string group = tmap->getGroup(Groups[i]);
392 vector<string> tempGroups; tempGroups.push_back(group);
393 tree[i].setGroup(tempGroups);
394 groupNodeInfo[group].push_back(i);
396 //set pcount and pGroup for groupname to 1.
397 tree[i].pcount[group] = 1;
398 tree[i].pGroups[group] = 1;
400 //Treemap knows name, group and index to speed up search
401 tmap->setIndex(Groups[i], i);
403 //intialize non leaf nodes
404 }else if (i > (numLeaves-1)) {
406 vector<string> tempGroups;
407 tree[i].setGroup(tempGroups);
411 set<int> removedLeaves;
412 for (int i = 0; i < copy->getNumLeaves(); i++) {
414 if (removedLeaves.count(i) == 0) {
417 int parent = copy->tree[i].getParent();
421 if (m->inUsersGroups(copy->tree[i].getName(), Groups)) {
422 //find my siblings name
423 int parentRC = copy->tree[parent].getRChild();
424 int parentLC = copy->tree[parent].getLChild();
426 //if I am the right child, then my sib is the left child
427 int sibIndex = parentRC;
428 if (parentRC == i) { sibIndex = parentLC; }
430 string sibsName = copy->tree[sibIndex].getName();
432 //if yes, is my sibling
433 if ((m->inUsersGroups(sibsName, Groups)) || (sibsName == "")) {
434 //we both are okay no trimming required
436 //i am, my sib is not, so remove sib by setting my parent to my grandparent
437 int grandparent = copy->tree[parent].getParent();
438 int grandparentLC = copy->tree[grandparent].getLChild();
439 int grandparentRC = copy->tree[grandparent].getRChild();
441 //whichever of my granparents children was my parent now equals me
442 if (grandparentLC == parent) { grandparentLC = i; }
443 else { grandparentRC = i; }
445 copy->tree[i].setParent(grandparent);
446 copy->tree[i].setBranchLength((copy->tree[i].getBranchLength()+copy->tree[parent].getBranchLength()));
447 if (grandparent != -1) {
448 copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
450 removedLeaves.insert(sibIndex);
453 //find my siblings name
454 int parentRC = copy->tree[parent].getRChild();
455 int parentLC = copy->tree[parent].getLChild();
457 //if I am the right child, then my sib is the left child
458 int sibIndex = parentRC;
459 if (parentRC == i) { sibIndex = parentLC; }
461 string sibsName = copy->tree[sibIndex].getName();
463 //if no is my sibling
464 if ((m->inUsersGroups(sibsName, Groups)) || (sibsName == "")) {
465 //i am not, but my sib is
466 int grandparent = copy->tree[parent].getParent();
467 int grandparentLC = copy->tree[grandparent].getLChild();
468 int grandparentRC = copy->tree[grandparent].getRChild();
470 //whichever of my granparents children was my parent now equals my sib
471 if (grandparentLC == parent) { grandparentLC = sibIndex; }
472 else { grandparentRC = sibIndex; }
474 copy->tree[sibIndex].setParent(grandparent);
475 copy->tree[sibIndex].setBranchLength((copy->tree[sibIndex].getBranchLength()+copy->tree[parent].getBranchLength()));
476 if (grandparent != -1) {
477 copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
479 removedLeaves.insert(i);
481 //neither of us are, so we want to eliminate ourselves and our parent
482 //so set our parents sib to our great-grandparent
483 int parent = copy->tree[i].getParent();
484 int grandparent = copy->tree[parent].getParent();
486 if (grandparent != -1) {
487 int greatgrandparent = copy->tree[grandparent].getParent();
488 int greatgrandparentLC, greatgrandparentRC;
489 if (greatgrandparent != -1) {
490 greatgrandparentLC = copy->tree[greatgrandparent].getLChild();
491 greatgrandparentRC = copy->tree[greatgrandparent].getRChild();
494 int grandparentLC = copy->tree[grandparent].getLChild();
495 int grandparentRC = copy->tree[grandparent].getRChild();
497 parentsSibIndex = grandparentLC;
498 if (grandparentLC == parent) { parentsSibIndex = grandparentRC; }
500 //whichever of my greatgrandparents children was my grandparent
501 if (greatgrandparentLC == grandparent) { greatgrandparentLC = parentsSibIndex; }
502 else { greatgrandparentRC = parentsSibIndex; }
504 copy->tree[parentsSibIndex].setParent(greatgrandparent);
505 copy->tree[parentsSibIndex].setBranchLength((copy->tree[parentsSibIndex].getBranchLength()+copy->tree[grandparent].getBranchLength()));
506 if (greatgrandparent != -1) {
507 copy->tree[greatgrandparent].setChildren(greatgrandparentLC, greatgrandparentRC);
510 copy->tree[parent].setParent(-1);
511 //cout << "issues with making subtree" << endl;
513 removedLeaves.insert(sibIndex);
514 removedLeaves.insert(i);
522 for (int i = 0; i < copy->getNumNodes(); i++) {
524 if (copy->tree[i].getParent() == -1) { root = i; break; }
527 int nextSpot = numLeaves;
528 populateNewTree(copy->tree, root, nextSpot);
532 catch(exception& e) {
533 m->errorOut(e, "Tree", "getSubTree");
537 /*****************************************************************/
538 //assumes nameMap contains unique names as key or is empty.
539 //assumes numLeaves defined in tree constructor equals size of seqsToInclude and seqsToInclude only contains unique seqs.
540 int Tree::getSubTree(Tree* copy, vector<string> seqsToInclude, map<string, string> nameMap) {
543 if (numLeaves != seqsToInclude.size()) { m->mothurOut("[ERROR]: numLeaves does not equal numUniques, cannot create subtree.\n"); m->control_pressed = true; return 0; }
545 getSubTree(copy, seqsToInclude);
546 if (nameMap.size() != 0) { addNamesToCounts(nameMap); }
548 //build the pGroups in non leaf nodes to be used in the parsimony calcs.
549 for (int i = numLeaves; i < numNodes; i++) {
550 if (m->control_pressed) { return 1; }
552 tree[i].pGroups = (mergeGroups(i));
553 tree[i].pcount = (mergeGcounts(i));
558 catch(exception& e) {
559 m->errorOut(e, "Tree", "getSubTree");
563 /*****************************************************************/
564 int Tree::populateNewTree(vector<Node>& oldtree, int node, int& index) {
567 if (oldtree[node].getLChild() != -1) {
568 int rc = populateNewTree(oldtree, oldtree[node].getLChild(), index);
569 int lc = populateNewTree(oldtree, oldtree[node].getRChild(), index);
571 tree[index].setChildren(lc, rc);
572 tree[rc].setParent(index);
573 tree[lc].setParent(index);
575 tree[index].setBranchLength(oldtree[node].getBranchLength());
576 tree[rc].setBranchLength(oldtree[oldtree[node].getLChild()].getBranchLength());
577 tree[lc].setBranchLength(oldtree[oldtree[node].getRChild()].getBranchLength());
580 }else { //you are a leaf
581 int indexInNewTree = tmap->getIndex(oldtree[node].getName());
582 return indexInNewTree;
585 catch(exception& e) {
586 m->errorOut(e, "Tree", "populateNewTree");
590 /*****************************************************************/
591 void Tree::getCopy(Tree* copy) {
594 //for each node in the tree copy its info
595 for (int i = 0; i < numNodes; i++) {
597 tree[i].setName(copy->tree[i].getName());
600 tree[i].setGroup(copy->tree[i].getGroup());
603 tree[i].setBranchLength(copy->tree[i].getBranchLength());
606 tree[i].setParent(copy->tree[i].getParent());
609 tree[i].setChildren(copy->tree[i].getLChild(), copy->tree[i].getRChild());
611 //copy index in node and tmap
612 tree[i].setIndex(copy->tree[i].getIndex());
613 setIndex(copy->tree[i].getName(), getIndex(copy->tree[i].getName()));
616 tree[i].pGroups = copy->tree[i].pGroups;
619 tree[i].pcount = copy->tree[i].pcount;
622 groupNodeInfo = copy->groupNodeInfo;
625 catch(exception& e) {
626 m->errorOut(e, "Tree", "getCopy");
630 /*****************************************************************/
631 //returns a map with a groupname and the number of times that group was seen in the children
632 //for instance if your children are white and black then it would return a map with 2 entries
633 // p[white] = 1 and p[black] = 1. Now go up a level and merge that with a node who has p[white] = 1
634 //and you get p[white] = 2, p[black] = 1, but you erase the p[black] because you have a p value higher than 1.
636 map<string, int> Tree::mergeGroups(int i) {
638 int lc = tree[i].getLChild();
639 int rc = tree[i].getRChild();
641 //set parsimony groups to left child
642 map<string,int> parsimony = tree[lc].pGroups;
646 //look at right child groups and update maxPars if right child has something higher for that group.
647 for(it=tree[rc].pGroups.begin();it!=tree[rc].pGroups.end();it++){
648 it2 = parsimony.find(it->first);
649 if (it2 != parsimony.end()) {
650 parsimony[it->first]++;
652 parsimony[it->first] = 1;
655 if(parsimony[it->first] > maxPars){
656 maxPars = parsimony[it->first];
660 // this is true if right child had a greater parsimony for a certain group
662 //erase all the groups that are only 1 because you found something with 2.
663 for(it=parsimony.begin();it!=parsimony.end();){
665 parsimony.erase(it++);
668 //set one remaining groups to 1
669 //so with our above example p[white] = 2 would be left and it would become p[white] = 1
670 for(it=parsimony.begin();it!=parsimony.end();it++){
671 parsimony[it->first] = 1;
678 catch(exception& e) {
679 m->errorOut(e, "Tree", "mergeGroups");
683 /*****************************************************************/
684 //returns a map with a groupname and the number of times that group was seen in the children
685 //for instance if your children are white and black then it would return a map with 2 entries
686 // p[white] = 1 and p[black] = 1. Now go up a level and merge that with a node who has p[white] = 1
687 //and you get p[white] = 2, p[black] = 1, but you erase the p[black] because you have a p value higher than 1.
689 map<string, int> Tree::mergeUserGroups(int i, vector<string> g) {
692 int lc = tree[i].getLChild();
693 int rc = tree[i].getRChild();
695 //loop through nodes groups removing the ones the user doesn't want
696 for(it=tree[lc].pGroups.begin();it!=tree[lc].pGroups.end();){
697 if (m->inUsersGroups(it->first, g) != true) {
698 tree[lc].pGroups.erase(it++);
702 //loop through nodes groups removing the ones the user doesn't want
703 for(it=tree[rc].pGroups.begin();it!=tree[rc].pGroups.end();){
704 if (m->inUsersGroups(it->first, g) != true) {
705 tree[rc].pGroups.erase(it++);
709 //set parsimony groups to left child
710 map<string,int> parsimony = tree[lc].pGroups;
714 //look at right child groups and update maxPars if right child has something higher for that group.
715 for(it=tree[rc].pGroups.begin();it!=tree[rc].pGroups.end();it++){
716 it2 = parsimony.find(it->first);
717 if (it2 != parsimony.end()) {
718 parsimony[it->first]++;
720 parsimony[it->first] = 1;
723 if(parsimony[it->first] > maxPars){
724 maxPars = parsimony[it->first];
728 // this is true if right child had a greater parsimony for a certain group
730 //erase all the groups that are only 1 because you found something with 2.
731 for(it=parsimony.begin();it!=parsimony.end();){
733 parsimony.erase(it++);
737 for(it=parsimony.begin();it!=parsimony.end();it++){
738 parsimony[it->first] = 1;
744 catch(exception& e) {
745 m->errorOut(e, "Tree", "mergeUserGroups");
751 /**************************************************************************************************/
753 map<string,int> Tree::mergeGcounts(int position) {
755 map<string,int>::iterator pos;
757 int lc = tree[position].getLChild();
758 int rc = tree[position].getRChild();
760 map<string,int> sum = tree[lc].pcount;
762 for(it=tree[rc].pcount.begin();it!=tree[rc].pcount.end();it++){
763 sum[it->first] += it->second;
767 catch(exception& e) {
768 m->errorOut(e, "Tree", "mergeGcounts");
772 /**************************************************************************************************/
773 void Tree::randomLabels(vector<string> g) {
776 //initialize groupNodeInfo
777 for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
778 groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
781 for(int i = 0; i < numLeaves; i++){
783 //get random index to switch with
784 z = int((float)(i+1) * (float)(rand()) / ((float)RAND_MAX+1.0));
786 //you only want to randomize the nodes that are from a group the user wants analyzed, so
787 //if either of the leaf nodes you are about to switch are not in the users groups then you don't want to switch them.
790 treez = m->inUsersGroups(tree[z].getGroup(), g);
791 treei = m->inUsersGroups(tree[i].getGroup(), g);
793 if ((treez == true) && (treei == true)) {
794 //switches node i and node z's info.
795 map<string,int> lib_hold = tree[z].pGroups;
796 tree[z].pGroups = (tree[i].pGroups);
797 tree[i].pGroups = (lib_hold);
799 vector<string> zgroup = tree[z].getGroup();
800 tree[z].setGroup(tree[i].getGroup());
801 tree[i].setGroup(zgroup);
803 string zname = tree[z].getName();
804 tree[z].setName(tree[i].getName());
805 tree[i].setName(zname);
807 map<string,int> gcount_hold = tree[z].pcount;
808 tree[z].pcount = (tree[i].pcount);
809 tree[i].pcount = (gcount_hold);
812 for (int k = 0; k < (tree[i].getGroup()).size(); k++) { groupNodeInfo[(tree[i].getGroup())[k]].push_back(i); }
813 for (int k = 0; k < (tree[z].getGroup()).size(); k++) { groupNodeInfo[(tree[z].getGroup())[k]].push_back(z); }
816 catch(exception& e) {
817 m->errorOut(e, "Tree", "randomLabels");
821 /**************************************************************************************************/
822 void Tree::randomBlengths() {
824 for(int i=numNodes-1;i>=0;i--){
825 int z = int((float)(i+1) * (float)(rand()) / ((float)RAND_MAX+1.0));
827 float bl_hold = tree[z].getBranchLength();
828 tree[z].setBranchLength(tree[i].getBranchLength());
829 tree[i].setBranchLength(bl_hold);
832 catch(exception& e) {
833 m->errorOut(e, "Tree", "randomBlengths");
837 /*************************************************************************************************/
838 void Tree::assembleRandomUnifracTree(vector<string> g) {
840 map<string, string> empty;
843 /*************************************************************************************************/
844 void Tree::assembleRandomUnifracTree(string groupA, string groupB) {
845 vector<string> temp; temp.push_back(groupA); temp.push_back(groupB);
847 map<string, string> empty;
851 /*************************************************************************************************/
852 //for now it's just random topology but may become random labels as well later that why this is such a simple function now...
853 void Tree::assembleRandomTree() {
855 map<string, string> empty;
858 /**************************************************************************************************/
860 void Tree::randomTopology() {
862 for(int i=0;i<numNodes;i++){
863 tree[i].setParent(-1);
865 for(int i=numLeaves;i<numNodes;i++){
866 tree[i].setChildren(-1, -1);
869 for(int i=numLeaves;i<numNodes;i++){
871 int rnd_index1, rnd_index2;
873 rnd_index1 = (int)(((double)rand() / (double) RAND_MAX)*i);
874 if(tree[rnd_index1].getParent() == -1){escape = 1;}
879 rnd_index2 = (int)(((double)rand() / (double) RAND_MAX)*i);
880 if(rnd_index2 != rnd_index1 && tree[rnd_index2].getParent() == -1){
885 tree[i].setChildren(rnd_index1,rnd_index2);
886 tree[i].setParent(-1);
887 tree[rnd_index1].setParent(i);
888 tree[rnd_index2].setParent(i);
891 catch(exception& e) {
892 m->errorOut(e, "Tree", "randomTopology");
896 /*****************************************************************/
897 void Tree::print(ostream& out) {
899 int root = findRoot();
900 printBranch(root, out, "branch");
903 catch(exception& e) {
904 m->errorOut(e, "Tree", "print");
908 /*****************************************************************/
909 void Tree::print(ostream& out, map<string, string> nameMap) {
911 int root = findRoot();
912 printBranch(root, out, nameMap);
915 catch(exception& e) {
916 m->errorOut(e, "Tree", "print");
920 /*****************************************************************/
921 void Tree::print(ostream& out, string mode) {
923 int root = findRoot();
924 printBranch(root, out, mode);
927 catch(exception& e) {
928 m->errorOut(e, "Tree", "print");
932 /*****************************************************************/
933 // This prints out the tree in Newick form.
934 void Tree::createNewickFile(string f) {
936 int root = findRoot();
940 m->openOutputFile(filename, out);
942 printBranch(root, out, "branch");
944 // you are at the end of the tree
948 catch(exception& e) {
949 m->errorOut(e, "Tree", "createNewickFile");
954 /*****************************************************************/
955 //This function finds the index of the root node.
957 int Tree::findRoot() {
959 for (int i = 0; i < numNodes; i++) {
961 if (tree[i].getParent() == -1) { return i; }
962 //cout << "i = " << i << endl;
963 //cout << "i's parent = " << tree[i].getParent() << endl;
967 catch(exception& e) {
968 m->errorOut(e, "Tree", "findRoot");
972 /*****************************************************************/
973 void Tree::printBranch(int node, ostream& out, map<string, string> names) {
976 // you are not a leaf
977 if (tree[node].getLChild() != -1) {
979 printBranch(tree[node].getLChild(), out, names);
981 printBranch(tree[node].getRChild(), out, names);
984 //if there is a branch length then print it
985 if (tree[node].getBranchLength() != -1) {
986 out << ":" << tree[node].getBranchLength();
989 }else { //you are a leaf
990 map<string, string>::iterator itNames = names.find(tree[node].getName());
992 string outputString = "";
993 if (itNames != names.end()) {
995 vector<string> dupNames;
996 m->splitAtComma((itNames->second), dupNames);
998 if (dupNames.size() == 1) {
999 outputString += tree[node].getName();
1000 if (tree[node].getBranchLength() != -1) {
1001 outputString += ":" + toString(tree[node].getBranchLength());
1004 outputString += "(";
1006 for (int u = 0; u < dupNames.size()-1; u++) {
1007 outputString += dupNames[u];
1009 if (tree[node].getBranchLength() != -1) {
1010 outputString += ":" + toString(0.0);
1012 outputString += ",";
1015 outputString += dupNames[dupNames.size()-1];
1016 if (tree[node].getBranchLength() != -1) {
1017 outputString += ":" + toString(0.0);
1020 outputString += ")";
1021 if (tree[node].getBranchLength() != -1) {
1022 outputString += ":" + toString(tree[node].getBranchLength());
1026 outputString = tree[node].getName();
1027 //if there is a branch length then print it
1028 if (tree[node].getBranchLength() != -1) {
1029 outputString += ":" + toString(tree[node].getBranchLength());
1032 m->mothurOut("[ERROR]: " + tree[node].getName() + " is not in your namefile, please correct."); m->mothurOutEndLine();
1035 out << outputString;
1039 catch(exception& e) {
1040 m->errorOut(e, "Tree", "printBranch");
1044 /*****************************************************************/
1045 void Tree::printBranch(int node, ostream& out, string mode) {
1048 // you are not a leaf
1049 if (tree[node].getLChild() != -1) {
1051 printBranch(tree[node].getLChild(), out, mode);
1053 printBranch(tree[node].getRChild(), out, mode);
1055 if (mode == "branch") {
1056 //if there is a branch length then print it
1057 if (tree[node].getBranchLength() != -1) {
1058 out << ":" << tree[node].getBranchLength();
1060 }else if (mode == "boot") {
1061 //if there is a label then print it
1062 if (tree[node].getLabel() != -1) {
1063 out << tree[node].getLabel();
1065 }else if (mode == "both") {
1066 if (tree[node].getLabel() != -1) {
1067 out << tree[node].getLabel();
1069 //if there is a branch length then print it
1070 if (tree[node].getBranchLength() != -1) {
1071 out << ":" << tree[node].getBranchLength();
1074 }else { //you are a leaf
1075 string leafGroup = tmap->getGroup(tree[node].getName());
1077 if (mode == "branch") {
1079 //if there is a branch length then print it
1080 if (tree[node].getBranchLength() != -1) {
1081 out << ":" << tree[node].getBranchLength();
1083 }else if (mode == "boot") {
1085 //if there is a label then print it
1086 if (tree[node].getLabel() != -1) {
1087 out << tree[node].getLabel();
1089 }else if (mode == "both") {
1090 out << tree[node].getName();
1091 if (tree[node].getLabel() != -1) {
1092 out << tree[node].getLabel();
1094 //if there is a branch length then print it
1095 if (tree[node].getBranchLength() != -1) {
1096 out << ":" << tree[node].getBranchLength();
1102 catch(exception& e) {
1103 m->errorOut(e, "Tree", "printBranch");
1107 /*****************************************************************/
1108 void Tree::printBranch(int node, ostream& out, string mode, vector<Node>& theseNodes) {
1111 // you are not a leaf
1112 if (theseNodes[node].getLChild() != -1) {
1114 printBranch(theseNodes[node].getLChild(), out, mode);
1116 printBranch(theseNodes[node].getRChild(), out, mode);
1118 if (mode == "branch") {
1119 //if there is a branch length then print it
1120 if (theseNodes[node].getBranchLength() != -1) {
1121 out << ":" << theseNodes[node].getBranchLength();
1123 }else if (mode == "boot") {
1124 //if there is a label then print it
1125 if (theseNodes[node].getLabel() != -1) {
1126 out << theseNodes[node].getLabel();
1128 }else if (mode == "both") {
1129 if (theseNodes[node].getLabel() != -1) {
1130 out << theseNodes[node].getLabel();
1132 //if there is a branch length then print it
1133 if (theseNodes[node].getBranchLength() != -1) {
1134 out << ":" << theseNodes[node].getBranchLength();
1137 }else { //you are a leaf
1138 string leafGroup = tmap->getGroup(theseNodes[node].getName());
1140 if (mode == "branch") {
1142 //if there is a branch length then print it
1143 if (theseNodes[node].getBranchLength() != -1) {
1144 out << ":" << theseNodes[node].getBranchLength();
1146 }else if (mode == "boot") {
1148 //if there is a label then print it
1149 if (theseNodes[node].getLabel() != -1) {
1150 out << theseNodes[node].getLabel();
1152 }else if (mode == "both") {
1153 out << theseNodes[node].getName();
1154 if (theseNodes[node].getLabel() != -1) {
1155 out << theseNodes[node].getLabel();
1157 //if there is a branch length then print it
1158 if (theseNodes[node].getBranchLength() != -1) {
1159 out << ":" << theseNodes[node].getBranchLength();
1165 catch(exception& e) {
1166 m->errorOut(e, "Tree", "printBranch");
1170 /*****************************************************************/
1172 void Tree::printTree() {
1174 for(int i=0;i<numNodes;i++){
1176 tree[i].printNode();
1181 /*****************************************************************/
1182 //this code is a mess and should be rethought...-slw
1183 void Tree::parseTreeFile() {
1185 //only takes names from the first tree and assumes that all trees use the same names.
1187 string filename = m->getTreeFile();
1188 ifstream filehandle;
1189 m->openInputFile(filename, filehandle);
1194 //ifyou are not a nexus file
1195 if((c = filehandle.peek()) != '#') {
1196 while((c = filehandle.peek()) != ';') {
1197 while ((c = filehandle.peek()) != ';') {
1198 // get past comments
1205 if((c == '(') && (comment != 1)){ break; }
1209 done = readTreeString(filehandle);
1210 if (done == 0) { break; }
1212 //ifyou are a nexus file
1213 }else if((c = filehandle.peek()) == '#') {
1216 // get past comments
1217 while(holder != "translate" && holder != "Translate"){
1218 if(holder == "[" || holder == "[!"){
1224 filehandle >> holder;
1226 //if there is no translate then you must read tree string otherwise use translate to get names
1227 if((holder == "tree") && (comment != 1)){
1228 //pass over the "tree rep.6878900 = "
1229 while (((c = filehandle.get()) != '(') && ((c = filehandle.peek()) != EOF)) {;}
1231 if(c == EOF) { break; }
1232 filehandle.putback(c); //put back first ( of tree.
1233 done = readTreeString(filehandle);
1238 if (done == 0) { break; }
1241 //use nexus translation rather than parsing tree to save time
1242 if((holder == "translate") || (holder == "Translate")) {
1244 string number, name, h;
1245 h = ""; // so it enters the loop the first time
1246 while((h != ";") && (number != ";")) {
1247 filehandle >> number;
1250 //c = , until done with translation then c = ;
1251 h = name.substr(name.length()-1, name.length());
1252 name.erase(name.end()-1); //erase the comma
1253 m->Treenames.push_back(number);
1255 if(number == ";") { m->Treenames.pop_back(); } //in case ';' from translation is on next line instead of next to last name
1260 //for (int i = 0; i < globaldata->Treenames.size(); i++) {
1261 //cout << globaldata->Treenames[i] << endl; }
1262 //cout << globaldata->Treenames.size() << endl;
1264 catch(exception& e) {
1265 m->errorOut(e, "Tree", "parseTreeFile");
1269 /*******************************************************/
1271 /*******************************************************/
1272 int Tree::readTreeString(ifstream& filehandle) {
1277 while((c = filehandle.peek()) != ';') {
1279 //cout << " at beginning of while " << k << endl;
1281 //to pass over labels in trees
1283 while((c!=',') && (c != -1) && (c!= ':') && (c!=';')){ c=filehandle.get(); }
1284 filehandle.putback(c);
1286 if(c == ';') { return 0; }
1287 if(c == -1) { return 0; }
1289 if((c != '(') && (c != ')') && (c != ',') && (c != ':') && (c != '\n') && (c != '\t') && (c != 32)) { //32 is space
1291 c = filehandle.get();
1293 //cout << k << endl;
1294 while ((c != '(') && (c != ')') && (c != ',') && (c != ':') && (c != '\n') && (c != 32) && (c != '\t')) {
1296 c = filehandle.get();
1298 //cout << " in name while " << k << endl;
1301 //cout << "name = " << name << endl;
1302 m->Treenames.push_back(name);
1303 filehandle.putback(c);
1305 //cout << " after putback" << k << endl;
1308 if(c == ':') { //read until you reach the end of the branch length
1309 while ((c != '(') && (c != ')') && (c != ',') && (c != ';') && (c != '\n') && (c != '\t') && (c != 32)) {
1310 c = filehandle.get();
1312 //cout << " in branch while " << k << endl;
1314 filehandle.putback(c);
1317 c = filehandle.get();
1319 //cout << " here after get " << k << endl;
1320 if(c == ';') { return 0; }
1321 if(c == ')') { filehandle.putback(c); }
1323 //cout << k << endl;
1328 catch(exception& e) {
1329 m->errorOut(e, "Tree", "readTreeString");
1334 /*******************************************************/
1336 /*******************************************************/