5 * Created by Sarah Westcott on 3/9/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 /* This class is designed to implement an integral form of the Cramer-von Mises statistic.
11 you may refer to the "Integration of Microbial Ecology and Statistics: A Test To Compare Gene Libraries"
12 paper in Applied and Environmental Microbiology, Sept. 2004, p. 5485-5492 0099-2240/04/$8.00+0
13 DOI: 10.1128/AEM.70.9.5485-5492.2004 Copyright 2004 American Society for Microbiology for more information. */
16 #include "libshuffcommand.h"
18 #include "slibshuff.h"
19 #include "dlibshuff.h"
23 //**********************************************************************************************************************
24 vector<string> LibShuffCommand::setParameters(){
26 CommandParameter pphylip("phylip", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pphylip);
27 CommandParameter pgroup("group", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pgroup);
28 CommandParameter pgroups("groups", "String", "", "", "", "", "",false,false); parameters.push_back(pgroups);
29 CommandParameter piters("iters", "Number", "", "10000", "", "", "",false,false); parameters.push_back(piters);
30 CommandParameter pstep("step", "Number", "", "0.01", "", "", "",false,false); parameters.push_back(pstep);
31 CommandParameter pcutoff("cutoff", "Number", "", "1.0", "", "", "",false,false); parameters.push_back(pcutoff);
32 CommandParameter pform("form", "Multiple", "discrete-integral", "integral", "", "", "",false,false); parameters.push_back(pform);
33 CommandParameter psim("sim", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(psim);
34 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
35 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
37 vector<string> myArray;
38 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
42 m->errorOut(e, "LibShuffCommand", "setParameters");
46 //**********************************************************************************************************************
47 string LibShuffCommand::getHelpString(){
49 string helpString = "";
50 helpString += "The libshuff command parameters are phylip, group, sim, groups, iters, step, form and cutoff. phylip and group parameters are required, unless you have valid current files.\n";
51 helpString += "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";
52 helpString += "The group names are separated by dashes. The iters parameter allows you to specify how many random matrices you would like compared to your matrix.\n";
53 helpString += "The step parameter allows you to specify change in distance you would like between each output if you are using the discrete form.\n";
54 helpString += "The form parameter allows you to specify if you would like to analyze your matrix using the discrete or integral form. Your options are integral or discrete.\n";
55 helpString += "The libshuff command should be in the following format: libshuff(groups=yourGroups, iters=yourIters, cutOff=yourCutOff, form=yourForm, step=yourStep).\n";
56 helpString += "Example libshuff(groups=A-B-C, iters=500, form=discrete, step=0.01, cutOff=2.0).\n";
57 helpString += "The default value for groups is all the groups in your groupfile, iters is 10000, cutoff is 1.0, form is integral and step is 0.01.\n";
58 helpString += "The libshuff command output two files: .coverage and .slsummary their descriptions are in the manual.\n";
59 helpString += "Note: No spaces between parameter labels (i.e. iters), '=' and parameters (i.e.yourIters).\n";
63 m->errorOut(e, "LibShuffCommand", "getHelpString");
67 //**********************************************************************************************************************
68 LibShuffCommand::LibShuffCommand(){
70 abort = true; calledHelp = true;
72 vector<string> tempOutNames;
73 outputTypes["coverage"] = tempOutNames;
74 outputTypes["libshuffsummary"] = tempOutNames;
77 m->errorOut(e, "LibShuffCommand", "LibShuffCommand");
82 //**********************************************************************************************************************
83 LibShuffCommand::LibShuffCommand(string option) {
85 abort = false; calledHelp = false;
87 //allow user to run help
88 if(option == "help") { help(); abort = true; calledHelp = true; }
91 vector<string> myArray = setParameters();
93 OptionParser parser(option);
94 map<string, string> parameters = parser.getParameters();
95 map<string,string>::iterator it;
97 ValidParameters validParameter;
99 //check to make sure all parameters are valid for command
100 for (it = parameters.begin(); it != parameters.end(); it++) {
101 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
104 //initialize outputTypes
105 vector<string> tempOutNames;
106 outputTypes["coverage"] = tempOutNames;
107 outputTypes["libshuffsummary"] = tempOutNames;
109 string inputDir = validParameter.validFile(parameters, "inputdir", false);
110 if (inputDir == "not found"){ inputDir = ""; }
113 it = parameters.find("phylip");
114 //user has given a template file
115 if(it != parameters.end()){
116 path = m->hasPath(it->second);
117 //if the user has not given a path then, add inputdir. else leave path alone.
118 if (path == "") { parameters["phylip"] = inputDir + it->second; }
121 it = parameters.find("group");
122 //user has given a template file
123 if(it != parameters.end()){
124 path = m->hasPath(it->second);
125 //if the user has not given a path then, add inputdir. else leave path alone.
126 if (path == "") { parameters["group"] = inputDir + it->second; }
130 //check for required parameters
131 phylipfile = validParameter.validFile(parameters, "phylip", true);
132 if (phylipfile == "not open") { phylipfile = ""; abort = true; }
133 else if (phylipfile == "not found") {
134 phylipfile = m->getPhylipFile();
135 if (phylipfile != "") { m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); }
137 m->mothurOut("You must provide a phylip file."); m->mothurOutEndLine();
142 //check for required parameters
143 groupfile = validParameter.validFile(parameters, "group", true);
144 if (groupfile == "not open") { groupfile = ""; abort = true; }
145 else if (groupfile == "not found") {
146 groupfile = m->getGroupFile();
147 if (groupfile != "") { m->mothurOut("Using " + groupfile + " as input file for the group parameter."); m->mothurOutEndLine(); }
149 m->mothurOut("You must provide a group file."); m->mothurOutEndLine();
154 //if the user changes the output directory command factory will send this info to us in the output parameter
155 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
157 outputDir += m->hasPath(phylipfile); //if user entered a file with a path then preserve it
160 //check for optional parameter and set defaults
161 // ...at some point should added some additional type checking...
162 groups = validParameter.validFile(parameters, "groups", false);
163 if (groups == "not found") { groups = ""; savegroups = groups; }
166 m->splitAtDash(groups, Groups);
171 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "10000"; }
172 convert(temp, iters);
174 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "1.0"; }
175 convert(temp, cutOff);
177 temp = validParameter.validFile(parameters, "step", false); if (temp == "not found") { temp = "0.01"; }
180 temp = validParameter.validFile(parameters, "sim", false); if (temp == "not found") { temp = "F"; }
181 sim = m->isTrue(temp);
183 userform = validParameter.validFile(parameters, "form", false); if (userform == "not found") { userform = "integral"; }
188 catch(exception& e) {
189 m->errorOut(e, "LibShuffCommand", "LibShuffCommand");
193 //**********************************************************************************************************************
195 int LibShuffCommand::execute(){
198 if (abort == true) { if (calledHelp) { return 0; } return 2; }
201 groupMap = new GroupMap(groupfile);
202 int error = groupMap->readMap();
203 if (error == 1) { delete groupMap; return 0; }
206 m->openInputFile(phylipfile, in);
207 matrix = new FullMatrix(in, groupMap, sim); //reads the matrix file
210 if (m->control_pressed) { delete groupMap; delete matrix; return 0; }
212 //if files don't match...
213 if (matrix->getNumSeqs() < groupMap->getNumSeqs()) {
214 m->mothurOut("Your distance file contains " + toString(matrix->getNumSeqs()) + " sequences, and your group file contains " + toString(groupMap->getNumSeqs()) + " sequences."); m->mothurOutEndLine();
215 //create new group file
216 if(outputDir == "") { outputDir += m->hasPath(groupfile); }
218 string newGroupFile = outputDir + m->getRootName(m->getSimpleName(groupfile)) + "editted.groups";
219 outputNames.push_back(newGroupFile);
221 m->openOutputFile(newGroupFile, outGroups);
223 for (int i = 0; i < matrix->getNumSeqs(); i++) {
224 if (m->control_pressed) { delete groupMap; delete matrix; outGroups.close(); remove(newGroupFile.c_str()); return 0; }
226 Names temp = matrix->getRowInfo(i);
227 outGroups << temp.seqName << '\t' << temp.groupName << endl;
231 m->mothurOut(newGroupFile + " is a new group file containing only the sequence that are in your distance file. I will read this file instead."); m->mothurOutEndLine();
235 groupfile = newGroupFile;
237 groupMap = new GroupMap(groupfile);
240 if (m->control_pressed) { delete groupMap; delete matrix; remove(newGroupFile.c_str()); return 0; }
244 setGroups(); //set the groups to be analyzed and sorts them
246 /********************************************************************************************/
247 //this is needed because when we read the matrix we sort it into groups in alphabetical order
248 //the rest of the command and the classes used in this command assume specific order
249 /********************************************************************************************/
250 matrix->setGroups(groupMap->namesOfGroups);
252 for (int i = 0; i < groupMap->namesOfGroups.size(); i++) { sizes.push_back(groupMap->getNumSeqs(groupMap->namesOfGroups[i])); }
253 matrix->setSizes(sizes);
256 if(userform == "discrete"){
257 form = new DLibshuff(matrix, iters, step, cutOff);
260 form = new SLibshuff(matrix, iters, cutOff);
263 savedDXYValues = form->evaluateAll();
264 savedMinValues = form->getSavedMins();
266 if (m->control_pressed) { delete form; m->Groups.clear(); delete matrix; delete groupMap; return 0; }
268 pValueCounts.resize(numGroups);
269 for(int i=0;i<numGroups;i++){
270 pValueCounts[i].assign(numGroups, 0);
273 if (m->control_pressed) { outputTypes.clear(); delete form; m->Groups.clear(); delete matrix; delete groupMap; return 0; }
275 Progress* reading = new Progress();
277 for(int i=0;i<numGroups-1;i++) {
278 for(int j=i+1;j<numGroups;j++) {
280 if (m->control_pressed) { outputTypes.clear(); delete form; m->Groups.clear(); delete matrix; delete groupMap; delete reading; return 0; }
282 reading->newLine(groupNames[i]+'-'+groupNames[j], iters);
283 int spoti = groupMap->groupIndex[groupNames[i]]; //neccessary in case user selects groups so you know where they are in the matrix
284 int spotj = groupMap->groupIndex[groupNames[j]];
286 for(int p=0;p<iters;p++) {
288 if (m->control_pressed) { outputTypes.clear(); delete form; m->Groups.clear(); delete matrix; delete groupMap; delete reading; return 0; }
290 form->randomizeGroups(spoti,spotj);
291 if(form->evaluatePair(spoti,spotj) >= savedDXYValues[spoti][spotj]) { pValueCounts[i][j]++; }
292 if(form->evaluatePair(spotj,spoti) >= savedDXYValues[spotj][spoti]) { pValueCounts[j][i]++; }
294 if (m->control_pressed) { outputTypes.clear(); delete form; m->Groups.clear(); delete matrix; delete groupMap; delete reading; return 0; }
298 form->resetGroup(spoti);
299 form->resetGroup(spotj);
303 if (m->control_pressed) { outputTypes.clear(); delete form; m->Groups.clear(); delete matrix; delete groupMap; delete reading; return 0; }
308 m->mothurOutEndLine();
312 //clear out users groups
316 delete matrix; delete groupMap;
318 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
320 m->mothurOutEndLine();
321 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
322 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
323 m->mothurOutEndLine();
327 catch(exception& e) {
328 m->errorOut(e, "LibShuffCommand", "execute");
333 //**********************************************************************************************************************
335 int LibShuffCommand::printCoverageFile() {
339 summaryFile = outputDir + m->getRootName(m->getSimpleName(phylipfile)) + "libshuff.coverage";
340 m->openOutputFile(summaryFile, outCov);
341 outputNames.push_back(summaryFile); outputTypes["coverage"].push_back(summaryFile);
342 outCov.setf(ios::fixed, ios::floatfield); outCov.setf(ios::showpoint);
343 //cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint);
345 map<double,vector<int> > allDistances;
346 map<double,vector<int> >::iterator it;
348 vector<vector<int> > indices(numGroups);
349 int numIndices = numGroups * numGroups;
352 for(int i=0;i<numGroups;i++){
353 indices[i].assign(numGroups,0);
354 for(int j=0;j<numGroups;j++){
355 indices[i][j] = index++;
357 int spoti = groupMap->groupIndex[groupNames[i]]; //neccessary in case user selects groups so you know where they are in the matrix
358 int spotj = groupMap->groupIndex[groupNames[j]];
360 for(int k=0;k<savedMinValues[spoti][spotj].size();k++){
362 if(m->control_pressed) { outCov.close(); return 0; }
364 if(allDistances[savedMinValues[spoti][spotj][k]].size() != 0){
365 allDistances[savedMinValues[spoti][spotj][k]][indices[i][j]]++;
368 allDistances[savedMinValues[spoti][spotj][k]].assign(numIndices, 0);
369 allDistances[savedMinValues[spoti][spotj][k]][indices[i][j]] = 1;
374 it=allDistances.begin();
376 //cout << setprecision(8);
378 vector<int> prevRow = it->second;
381 for(;it!=allDistances.end();it++){
382 for(int i=0;i<it->second.size();i++){
383 it->second[i] += prevRow[i];
385 prevRow = it->second;
388 vector<int> lastRow = allDistances.rbegin()->second;
389 outCov << setprecision(8);
392 for (int i = 0; i < numGroups; i++){
393 outCov << '\t' << groupNames[i];
395 for (int i=0;i<numGroups;i++){
396 for(int j=i+1;j<numGroups;j++){
397 if(m->control_pressed) { outCov.close(); return 0; }
398 outCov << '\t' << groupNames[i] << '-' << groupNames[j] << '\t';
399 outCov << groupNames[j] << '-' << groupNames[i];
404 for(it=allDistances.begin();it!=allDistances.end();it++){
405 outCov << it->first << '\t';
406 for(int i=0;i<numGroups;i++){
407 outCov << it->second[indices[i][i]]/(float)lastRow[indices[i][i]] << '\t';
409 for(int i=0;i<numGroups;i++){
410 for(int j=i+1;j<numGroups;j++){
411 if(m->control_pressed) { outCov.close(); return 0; }
413 outCov << it->second[indices[i][j]]/(float)lastRow[indices[i][j]] << '\t';
414 outCov << it->second[indices[j][i]]/(float)lastRow[indices[j][i]] << '\t';
423 catch(exception& e) {
424 m->errorOut(e, "LibShuffCommand", "printCoverageFile");
429 //**********************************************************************************************************************
431 int LibShuffCommand::printSummaryFile() {
435 summaryFile = outputDir + m->getRootName(m->getSimpleName(phylipfile)) + "libshuff.summary";
436 m->openOutputFile(summaryFile, outSum);
437 outputNames.push_back(summaryFile); outputTypes["libshuffsummary"].push_back(summaryFile);
439 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
440 cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint);
442 cout << setw(20) << left << "Comparison" << '\t' << setprecision(8) << "dCXYScore" << '\t' << "Significance" << endl;
443 m->mothurOutJustToLog("Comparison\tdCXYScore\tSignificance"); m->mothurOutEndLine();
444 outSum << setw(20) << left << "Comparison" << '\t' << setprecision(8) << "dCXYScore" << '\t' << "Significance" << endl;
446 int precision = (int)log10(iters);
447 for(int i=0;i<numGroups;i++){
448 for(int j=i+1;j<numGroups;j++){
449 if(m->control_pressed) { outSum.close(); return 0; }
451 int spoti = groupMap->groupIndex[groupNames[i]]; //neccessary in case user selects groups so you know where they are in the matrix
452 int spotj = groupMap->groupIndex[groupNames[j]];
454 if(pValueCounts[i][j]){
455 cout << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[spoti][spotj] << '\t' << setprecision(precision) << pValueCounts[i][j]/(float)iters << endl;
456 m->mothurOutJustToLog(groupNames[i]+"-"+groupNames[j] + "\t" + toString(savedDXYValues[spoti][spotj]) + "\t" + toString((pValueCounts[i][j]/(float)iters))); m->mothurOutEndLine();
457 outSum << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[spoti][spotj] << '\t' << setprecision(precision) << pValueCounts[i][j]/(float)iters << endl;
460 cout << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[spoti][spotj] << '\t' << '<' <<setprecision(precision) << 1/(float)iters << endl;
461 m->mothurOutJustToLog(groupNames[i]+"-"+groupNames[j] + "\t" + toString(savedDXYValues[spoti][spotj]) + "\t" + toString((1/(float)iters))); m->mothurOutEndLine();
462 outSum << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[spoti][spotj] << '\t' << '<' <<setprecision(precision) << 1/(float)iters << endl;
464 if(pValueCounts[j][i]){
465 cout << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[spotj][spoti] << '\t' << setprecision (precision) << pValueCounts[j][i]/(float)iters << endl;
466 m->mothurOutJustToLog(groupNames[j]+"-"+groupNames[i] + "\t" + toString(savedDXYValues[spotj][spoti]) + "\t" + toString((pValueCounts[j][i]/(float)iters))); m->mothurOutEndLine();
467 outSum << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[spotj][spoti] << '\t' << setprecision (precision) << pValueCounts[j][i]/(float)iters << endl;
470 cout << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[spotj][spoti] << '\t' << '<' <<setprecision (precision) << 1/(float)iters << endl;
471 m->mothurOutJustToLog(groupNames[j]+"-"+groupNames[i] + "\t" + toString(savedDXYValues[spotj][spoti]) + "\t" + toString((1/(float)iters))); m->mothurOutEndLine();
472 outSum << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[spotj][spoti] << '\t' << '<' <<setprecision (precision) << 1/(float)iters << endl;
480 catch(exception& e) {
481 m->errorOut(e, "LibShuffCommand", "printSummaryFile");
486 //**********************************************************************************************************************
488 void LibShuffCommand::setGroups() {
490 //if the user has not entered specific groups to analyze then do them all
491 if (m->Groups.size() == 0) {
492 numGroups = groupMap->getNumGroups();
493 for (int i=0; i < numGroups; i++) {
494 m->Groups.push_back(groupMap->namesOfGroups[i]);
497 if (savegroups != "all") {
498 //check that groups are valid
499 for (int i = 0; i < m->Groups.size(); i++) {
500 if (groupMap->isValidGroup(m->Groups[i]) != true) {
501 m->mothurOut(m->Groups[i] + " is not a valid group, and will be disregarded."); m->mothurOutEndLine();
502 // erase the invalid group from globaldata->Groups
503 m->Groups.erase(m->Groups.begin()+i);
507 //if the user only entered invalid groups
508 if ((m->Groups.size() == 0) || (m->Groups.size() == 1)) {
509 numGroups = groupMap->getNumGroups();
510 for (int i=0; i < numGroups; i++) {
511 m->Groups.push_back(groupMap->namesOfGroups[i]);
513 m->mothurOut("When using the groups parameter you must have at least 2 valid groups. I will run the command using all the groups in your groupfile."); m->mothurOutEndLine();
514 } else { numGroups = m->Groups.size(); }
515 } else { //users wants all groups
516 numGroups = groupMap->getNumGroups();
518 for (int i=0; i < numGroups; i++) {
519 m->Groups.push_back(groupMap->namesOfGroups[i]);
524 //sort so labels match
525 sort(m->Groups.begin(), m->Groups.end());
528 sort(groupMap->namesOfGroups.begin(), groupMap->namesOfGroups.end());
530 for (int i = 0; i < groupMap->namesOfGroups.size(); i++) { groupMap->groupIndex[groupMap->namesOfGroups[i]] = i; }
532 groupNames = m->Groups;
535 catch(exception& e) {
536 m->errorOut(e, "LibShuffCommand", "setGroups");
541 /***********************************************************/