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"
21 //**********************************************************************************************************************
23 LibShuffCommand::LibShuffCommand(string option){
25 srand( (unsigned)time( NULL ) );
27 globaldata = GlobalData::getInstance();
32 //allow user to run help
33 if(option == "help") { help(); abort = true; }
36 //valid paramters for this command
37 string Array[] = {"iters","groups","step","form","cutoff"};
38 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
40 parser = new OptionParser();
41 parser->parse(option, parameters); delete parser;
43 ValidParameters* validParameter = new ValidParameters();
45 //check to make sure all parameters are valid for command
46 for (it = parameters.begin(); it != parameters.end(); it++) {
47 if (validParameter->isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
50 //make sure the user has already run the read.dist command
51 if ((globaldata->gMatrix == NULL) || (globaldata->gGroupmap == NULL)) {
52 cout << "You must read in a matrix and groupfile using the read.dist command, before you use the libshuff command. " << endl; abort = true;;
55 //check for optional parameter and set defaults
56 // ...at some point should added some additional type checking...
57 groups = validParameter->validFile(parameters, "groups", false);
58 if (groups == "not found") { groups = ""; savegroups = groups; }
61 splitAtDash(groups, Groups);
62 globaldata->Groups = Groups;
66 temp = validParameter->validFile(parameters, "iters", false); if (temp == "not found") { temp = "10000"; }
69 temp = validParameter->validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "1.0"; }
70 convert(temp, cutOff);
72 temp = validParameter->validFile(parameters, "step", false); if (temp == "not found") { temp = "0.01"; }
75 userform = validParameter->validFile(parameters, "form", false); if (userform == "not found") { userform = "integral"; }
77 delete validParameter;
81 matrix = globaldata->gMatrix; //get the distance matrix
82 setGroups(); //set the groups to be analyzed
84 if(userform == "discrete"){
85 form = new DLibshuff(matrix, iters, step, cutOff);
88 form = new SLibshuff(matrix, iters, cutOff);
96 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function LibShuffCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
100 cout << "An unknown error has occurred in the LibShuffCommand class function LibShuffCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
105 //**********************************************************************************************************************
107 void LibShuffCommand::help(){
109 cout << "The libshuff command can only be executed after a successful read.dist command including a groupfile." << "\n";
110 cout << "The libshuff command parameters are groups, iters, step, form and cutoff. No parameters are required." << "\n";
111 cout << "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";
112 cout << "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";
113 cout << "The step parameter allows you to specify change in distance you would like between each output if you are using the discrete form." << "\n";
114 cout << "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";
115 cout << "The libshuff command should be in the following format: libshuff(groups=yourGroups, iters=yourIters, cutOff=yourCutOff, form=yourForm, step=yourStep)." << "\n";
116 cout << "Example libshuff(groups=A-B-C, iters=500, form=discrete, step=0.01, cutOff=2.0)." << "\n";
117 cout << "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";
118 cout << "The libshuff command output two files: .coverage and .slsummary their descriptions are in the manual." << "\n";
119 cout << "Note: No spaces between parameter labels (i.e. iters), '=' and parameters (i.e.yourIters)." << "\n" << "\n";
121 catch(exception& e) {
122 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
126 cout << "An unknown error has occurred in the LibShuffCommand class function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
131 //**********************************************************************************************************************
133 int LibShuffCommand::execute(){
136 if (abort == true) { return 0; }
138 savedDXYValues = form->evaluateAll();
139 savedMinValues = form->getSavedMins();
141 pValueCounts.resize(numGroups);
142 for(int i=0;i<numGroups;i++){
143 pValueCounts[i].assign(numGroups, 0);
146 Progress* reading = new Progress();
148 for(int i=0;i<numGroups-1;i++) {
149 for(int j=i+1;j<numGroups;j++) {
150 reading->newLine(groupNames[i]+'-'+groupNames[j], iters);
151 for(int p=0;p<iters;p++) {
152 form->randomizeGroups(i,j);
153 if(form->evaluatePair(i,j) >= savedDXYValues[i][j]) { pValueCounts[i][j]++; }
154 if(form->evaluatePair(j,i) >= savedDXYValues[j][i]) { pValueCounts[j][i]++; }
168 //clear out users groups
169 globaldata->Groups.clear();
172 //delete globaldata's copy of the gmatrix to free up memory
173 delete globaldata->gMatrix; globaldata->gMatrix = NULL;
177 catch(exception& e) {
178 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
182 cout << "An unknown error has occurred in the LibShuffCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
187 //**********************************************************************************************************************
189 void LibShuffCommand::printCoverageFile() {
193 summaryFile = getRootName(globaldata->getPhylipFile()) + "libshuff.coverage";
194 openOutputFile(summaryFile, outCov);
195 outCov.setf(ios::fixed, ios::floatfield); outCov.setf(ios::showpoint);
196 cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint);
198 map<double,vector<int> > allDistances;
199 map<double,vector<int> >::iterator it;
201 vector<vector<int> > indices(numGroups);
202 int numIndices = numGroups * numGroups;
205 for(int i=0;i<numGroups;i++){
206 indices[i].assign(numGroups,0);
207 for(int j=0;j<numGroups;j++){
208 indices[i][j] = index++;
209 for(int k=0;k<savedMinValues[i][j].size();k++){
210 if(allDistances[savedMinValues[i][j][k]].size() != 0){
211 allDistances[savedMinValues[i][j][k]][indices[i][j]]++;
214 allDistances[savedMinValues[i][j][k]].assign(numIndices, 0);
215 allDistances[savedMinValues[i][j][k]][indices[i][j]] = 1;
220 it=allDistances.begin();
222 cout << setprecision(8);
224 vector<int> prevRow = it->second;
227 for(it;it!=allDistances.end();it++){
228 for(int i=0;i<it->second.size();i++){
229 it->second[i] += prevRow[i];
231 prevRow = it->second;
234 vector<int> lastRow = allDistances.rbegin()->second;
235 outCov << setprecision(8);
238 for (int i = 0; i < numGroups; i++){
239 outCov << '\t' << groupNames[i];
241 for (int i=0;i<numGroups;i++){
242 for(int j=i+1;j<numGroups;j++){
243 outCov << '\t' << groupNames[i] << '-' << groupNames[j] << '\t';
244 outCov << groupNames[j] << '-' << groupNames[i];
249 for(it=allDistances.begin();it!=allDistances.end();it++){
250 outCov << it->first << '\t';
251 for(int i=0;i<numGroups;i++){
252 outCov << it->second[indices[i][i]]/(float)lastRow[indices[i][i]] << '\t';
254 for(int i=0;i<numGroups;i++){
255 for(int j=i+1;j<numGroups;j++){
256 outCov << it->second[indices[i][j]]/(float)lastRow[indices[i][j]] << '\t';
257 outCov << it->second[indices[j][i]]/(float)lastRow[indices[j][i]] << '\t';
264 catch(exception& e) {
265 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function printCoverageFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
269 cout << "An unknown error has occurred in the LibShuffCommand class function printCoverageFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
274 //**********************************************************************************************************************
276 void LibShuffCommand::printSummaryFile() {
280 summaryFile = getRootName(globaldata->getPhylipFile()) + "libshuff.summary";
281 openOutputFile(summaryFile, outSum);
283 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
284 cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint);
286 cout << setw(20) << left << "Comparison" << '\t' << setprecision(8) << "dCXYScore" << '\t' << "Significance" << endl;
287 outSum << setw(20) << left << "Comparison" << '\t' << setprecision(8) << "dCXYScore" << '\t' << "Significance" << endl;
289 int precision = (int)log10(iters);
290 for(int i=0;i<numGroups;i++){
291 for(int j=i+1;j<numGroups;j++){
292 if(pValueCounts[i][j]){
293 cout << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[i][j] << '\t' << setprecision(precision) << pValueCounts[i][j]/(float)iters << endl;
294 outSum << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[i][j] << '\t' << setprecision(precision) << pValueCounts[i][j]/(float)iters << endl;
297 cout << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[i][j] << '\t' << '<' <<setprecision(precision) << 1/(float)iters << endl;
298 outSum << setw(20) << left << groupNames[i]+'-'+groupNames[j] << '\t' << setprecision(8) << savedDXYValues[i][j] << '\t' << '<' <<setprecision(precision) << 1/(float)iters << endl;
300 if(pValueCounts[j][i]){
301 cout << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[j][i] << '\t' << setprecision (precision) << pValueCounts[j][i]/(float)iters << endl;
302 outSum << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[j][i] << '\t' << setprecision (precision) << pValueCounts[j][i]/(float)iters << endl;
305 cout << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[j][i] << '\t' << '<' <<setprecision (precision) << 1/(float)iters << endl;
306 outSum << setw(20) << left << groupNames[j]+'-'+groupNames[i] << '\t' << setprecision(8) << savedDXYValues[j][i] << '\t' << '<' <<setprecision (precision) << 1/(float)iters << endl;
313 catch(exception& e) {
314 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function printSummaryFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
318 cout << "An unknown error has occurred in the LibShuffCommand class function printSummaryFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
323 //**********************************************************************************************************************
325 void LibShuffCommand::setGroups() {
327 //if the user has not entered specific groups to analyze then do them all
328 if (globaldata->Groups.size() == 0) {
329 numGroups = globaldata->gGroupmap->getNumGroups();
330 for (int i=0; i < numGroups; i++) {
331 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
334 if (savegroups != "all") {
335 //check that groups are valid
336 for (int i = 0; i < globaldata->Groups.size(); i++) {
337 if (globaldata->gGroupmap->isValidGroup(globaldata->Groups[i]) != true) {
338 cout << globaldata->Groups[i] << " is not a valid group, and will be disregarded." << endl;
339 // erase the invalid group from globaldata->Groups
340 globaldata->Groups.erase(globaldata->Groups.begin()+i);
344 //if the user only entered invalid groups
345 if ((globaldata->Groups.size() == 0) || (globaldata->Groups.size() == 1)) {
346 numGroups = globaldata->gGroupmap->getNumGroups();
347 for (int i=0; i < numGroups; i++) {
348 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
350 cout << "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." << endl;
351 } else { numGroups = globaldata->Groups.size(); }
352 } else { //users wants all groups
353 numGroups = globaldata->gGroupmap->getNumGroups();
354 globaldata->Groups.clear();
355 for (int i=0; i < numGroups; i++) {
356 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
361 //sort so labels match
362 sort(globaldata->Groups.begin(), globaldata->Groups.end());
365 sort(globaldata->gGroupmap->namesOfGroups.begin(), globaldata->gGroupmap->namesOfGroups.end());
367 groupNames = globaldata->Groups;
369 // number of comparisons i.e. with groups A,B,C = AA, AB, AC, BA, BB, BC...;
370 // for (int i=0; i<numGroups; i++) {
371 // for (int l = 0; l < numGroups; l++) {
372 // //set group comparison labels
373 // groupComb.push_back(globaldata->Groups[i] + "-" + globaldata->Groups[l]);
377 catch(exception& e) {
378 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function setGroups. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
382 cout << "An unknown error has occurred in the LibShuffCommand class function setGroups. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
387 /***********************************************************/