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 //**********************************************************************************************************************
21 LibShuffCommand::LibShuffCommand(){
23 globaldata = GlobalData::getInstance();
24 convert(globaldata->getCutOff(), cutOff);
25 convert(globaldata->getIters(), iters);
26 convert(globaldata->getStep(), step);
27 form = globaldata->getForm();
28 matrix = globaldata->gMatrix;
29 coverageFile = getRootName(globaldata->getPhylipFile()) + "coverage";
30 summaryFile = getRootName(globaldata->getPhylipFile()) + "slsummary";
31 openOutputFile(coverageFile, out);
32 openOutputFile(summaryFile, outSum);
34 //set the groups to be analyzed
37 //file headers for coverage file
39 for (int i = 0; i < groupComb.size(); i++) {
40 out << "C" + groupComb[i] << '\t';
43 for (int i = 0; i < numGroups; i++) {
44 for (int j = 0; j < numGroups; j++) {
45 //don't output AA to AA
47 out << "Delta" + globaldata->Groups[i] + "-" + globaldata->Groups[j] << '\t';
53 numComp = numGroups*numGroups;
55 coverage = new Coverage();
59 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function LibShuffCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
63 cout << "An unknown error has occurred in the LibShuffCommand class function LibShuffCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
69 //**********************************************************************************************************************
71 LibShuffCommand::~LibShuffCommand(){
75 //**********************************************************************************************************************
77 int LibShuffCommand::execute(){
79 //deltaValues[0] = scores for the difference between AA and AB.
80 //cValues[0][0][0] = AA at distance 0.0, cValues[0][0][1] = AB at distance 0.0, cValues[0][0][2] = AC at distance 0.0, cValues[0][1][0] = BA at distance 0.0, cValues[0][1][1] = BB...
82 reading = new Progress("Comparing to random:", iters);
84 sumDelta.resize(numComp-numGroups, 0.0);
89 if (form != "discrete") { matrix->getDist(dist); }
98 /*****************************/
99 //get values for users matrix
100 /*****************************/
102 //clear out old Values
104 deltaValues.resize(dist.size());
106 coverage->getValues(matrix, cValues, dist, "user");
108 float distDiff = dist[0];
110 //loop through each distance and load rsumdelta
111 for (int p = 0; p < cValues.size(); p++) {
114 for (int i = 0; i < numGroups; i++) {
115 for (int j = 0; j < numGroups; j++) {
116 //don't save AA to AA
119 deltaValues[p].push_back(((cValues[p][i][i]-cValues[p][i][j]) * (cValues[p][i][i]-cValues[p][i][j])) * distDiff);
120 sumDelta[count] += deltaValues[p][count];
125 if (p < cValues.size() - 1) {
126 distDiff = dist[p+1] - dist[p];
127 //cout << distDiff << endl;
133 /*******************************************************************************/
134 //create and score random matrixes finding the sumDelta values for summary file
135 /******************************************************************************/
137 //initialize rsumDelta
138 rsumDelta.resize(numComp-numGroups);
139 for (int l = 0; l < rsumDelta.size(); l++) {
140 for (int w = 0; w < iters; w++) {
141 rsumDelta[l].push_back(0.0);
146 for (int m = 0; m < iters; m++) {
147 //generate random matrix in getValues
148 //values for random matrix
150 coverage->getValues(matrix, cValues, dist, "random");
154 //loop through each distance and load rsumdelta
155 for (int p = 0; p < cValues.size(); p++) {
158 for (int i = 0; i < numGroups; i++) {
159 for (int j = 0; j < numGroups; j++) {
160 //don't save AA to AA
163 rsumDelta[count][m] += (((cValues[p][i][i]-cValues[p][i][j]) * (cValues[p][i][i]-cValues[p][i][j])) * distDiff);
169 if (p < cValues.size() - 1) {
170 distDiff = dist[p+1] - dist[p];
174 //clear out old Values
183 /**********************************************************/
184 //find the signifigance of the user matrix' sumdelta values
185 /**********************************************************/
187 for (int t = 0; t < rsumDelta.size(); t++) {
189 sort(rsumDelta[t].begin(), rsumDelta[t].end());
191 //the index of the score higher than yours is returned
192 //so if you have 1000 random matrices the index returned is 100
193 //then there are 900 matrices with a score greater then you.
194 //giving you a signifigance of 0.900
195 int index = findIndex(sumDelta[t], t);
197 //the signifigance is the number of trees with the users score or higher
198 sumDeltaSig.push_back((iters-index)/(float)iters);
204 //clear out users groups
205 globaldata->Groups.clear();
209 catch(exception& e) {
210 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
214 cout << "An unknown error has occurred in the LibShuffCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
218 //**********************************************************************************************************************
219 void LibShuffCommand::printCoverageFile() {
222 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
224 //loop through each distance
225 for (int p = 0; p < cValues.size(); p++) {
226 out << setprecision(6) << dist[p] << '\t';
227 //print out coverage values
228 for (int i = 0; i < numGroups; i++) {
229 for (int j = 0; j < numGroups; j++) {
230 out << cValues[p][i][j] << '\t';
234 for (int h = 0; h < deltaValues[p].size(); h++) {
235 out << deltaValues[p][h] << '\t';
242 catch(exception& e) {
243 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function printCoverageFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
247 cout << "An unknown error has occurred in the LibShuffCommand class function printCoverageFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
251 //**********************************************************************************************************************
252 void LibShuffCommand::printSummaryFile() {
255 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
257 for (int i = 0; i < numGroups; i++) {
258 for (int j = 0; j < numGroups; j++) {
259 //don't output AA to AA
261 outSum << "Delta " + globaldata->Groups[i] + "-" + globaldata->Groups[j] << '\t'<< "DeltaSig " + globaldata->Groups[i] + "-" + globaldata->Groups[j] << '\t';
262 cout << "Delta " + globaldata->Groups[i] + "-" + globaldata->Groups[j] << '\t'<< "DeltaSig " + globaldata->Groups[i] + "-" + globaldata->Groups[j] << '\t';
269 //print out delta values
270 for (int i = 0; i < sumDelta.size(); i++) {
271 if (sumDeltaSig[i] > (1/(float)iters)) {
272 outSum << setprecision(6) << sumDelta[i] << '\t' << setprecision(globaldata->getIters().length()) << sumDeltaSig[i] << '\t';
273 cout << setprecision(6) << sumDelta[i] << '\t' << setprecision(globaldata->getIters().length()) << sumDeltaSig[i] << '\t';
275 outSum << setprecision(6) << sumDelta[i] << '\t' << setprecision(globaldata->getIters().length()) << "<" << (1/float(iters)) << '\t';
276 cout << setprecision(6) << sumDelta[i] << '\t' << setprecision(globaldata->getIters().length()) << "<" << (1/float(iters)) << '\t';
284 catch(exception& e) {
285 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function printSummaryFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
289 cout << "An unknown error has occurred in the LibShuffCommand class function printSummaryFile. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
294 //**********************************************************************************************************************
295 void LibShuffCommand::setGroups() {
297 //if the user has not entered specific groups to analyze then do them all
298 if (globaldata->Groups.size() == 0) {
299 numGroups = globaldata->gGroupmap->getNumGroups();
300 for (int i=0; i < numGroups; i++) {
301 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
304 if (globaldata->getGroups() != "all") {
305 //check that groups are valid
306 for (int i = 0; i < globaldata->Groups.size(); i++) {
307 if (globaldata->gGroupmap->isValidGroup(globaldata->Groups[i]) != true) {
308 cout << globaldata->Groups[i] << " is not a valid group, and will be disregarded." << endl;
309 // erase the invalid group from globaldata->Groups
310 globaldata->Groups.erase (globaldata->Groups.begin()+i);
314 //if the user only entered invalid groups
315 if ((globaldata->Groups.size() == 0) || (globaldata->Groups.size() == 1)) {
316 numGroups = globaldata->gGroupmap->getNumGroups();
317 for (int i=0; i < numGroups; i++) {
318 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
320 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;
321 }else { numGroups = globaldata->Groups.size(); }
322 }else { //users wants all groups
323 numGroups = globaldata->gGroupmap->getNumGroups();
324 globaldata->Groups.clear();
325 for (int i=0; i < numGroups; i++) {
326 globaldata->Groups.push_back(globaldata->gGroupmap->namesOfGroups[i]);
331 //sort so labels match
332 sort(globaldata->Groups.begin(), globaldata->Groups.end());
334 // number of comparisons i.e. with groups A,B,C = AA, AB, AC, BA, BB, BC...;
335 for (int i=0; i<numGroups; i++) {
336 for (int l = 0; l < numGroups; l++) {
337 //set group comparison labels
338 groupComb.push_back(globaldata->Groups[i] + "-" + globaldata->Groups[l]);
342 catch(exception& e) {
343 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function setGroups. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
347 cout << "An unknown error has occurred in the LibShuffCommand class function setGroups. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
351 /***********************************************************/
352 int LibShuffCommand::findIndex(float score, int index) {
354 for (int i = 0; i < rsumDelta[index].size(); i++) {
355 if (rsumDelta[index][i] >= score) { return i; }
357 return rsumDelta[index].size();
359 catch(exception& e) {
360 cout << "Standard Error: " << e.what() << " has occurred in the LibShuffCommand class Function findIndex. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
364 cout << "An unknown error has occurred in the LibShuffCommand class function findIndex. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
369 /***********************************************************/