X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=corraxescommand.cpp;h=c27eb4b9ac335112f19244c708cee9bb8630feda;hb=d205e70ae86dbee2efc2df02f2717975854de6ba;hp=82cfa3683cbd0801da4a9805a567c84f67d3c427;hpb=ca9ac1d80c62f57270b0dcd49410ebe08a8aecd6;p=mothur.git diff --git a/corraxescommand.cpp b/corraxescommand.cpp index 82cfa36..c27eb4b 100644 --- a/corraxescommand.cpp +++ b/corraxescommand.cpp @@ -9,6 +9,7 @@ #include "corraxescommand.h" #include "sharedutilities.h" +#include "linearalgebra.h" //********************************************************************************************************************** vector CorrAxesCommand::setParameters(){ @@ -46,7 +47,7 @@ string CorrAxesCommand::getHelpString(){ helpString += "The corr.axes command should be in the following format: corr.axes(axes=yourPcoaFile, shared=yourSharedFile, method=yourMethod).\n"; helpString += "Example corr.axes(axes=genus.pool.thetayc.genus.lt.pcoa, shared=genus.pool.shared, method=kendall).\n"; helpString += "The corr.axes command outputs a .corr.axes file.\n"; - helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n"; + helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n"; return helpString; } catch(exception& e) { @@ -74,6 +75,7 @@ CorrAxesCommand::CorrAxesCommand(string option) { //allow user to run help if(option == "help") { help(); abort = true; calledHelp = true; } + else if(option == "citation") { citation(); abort = true; calledHelp = true;} else { vector myArray = setParameters(); @@ -139,12 +141,12 @@ CorrAxesCommand::CorrAxesCommand(string option) { sharedfile = validParameter.validFile(parameters, "shared", true); if (sharedfile == "not open") { abort = true; } else if (sharedfile == "not found") { sharedfile = ""; } - else { inputFileName = sharedfile; } + else { inputFileName = sharedfile; m->setSharedFile(sharedfile); } relabundfile = validParameter.validFile(parameters, "relabund", true); if (relabundfile == "not open") { abort = true; } else if (relabundfile == "not found") { relabundfile = ""; } - else { inputFileName = relabundfile; } + else { inputFileName = relabundfile; m->setRelAbundFile(relabundfile); } metadatafile = validParameter.validFile(parameters, "metadata", true); if (metadatafile == "not open") { abort = true; } @@ -157,7 +159,7 @@ CorrAxesCommand::CorrAxesCommand(string option) { pickedGroups = true; m->splitAtDash(groups, Groups); } - m->Groups = Groups; + m->setGroups(Groups); outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = m->hasPath(inputFileName); } @@ -186,7 +188,7 @@ CorrAxesCommand::CorrAxesCommand(string option) { } string temp; temp = validParameter.validFile(parameters, "numaxes", false); if (temp == "not found"){ temp = "3"; } - convert(temp, numaxes); + m->mothurConvert(temp, numaxes); method = validParameter.validFile(parameters, "method", false); if (method == "not found"){ method = "pearson"; } @@ -274,7 +276,7 @@ int CorrAxesCommand::execute(){ if (metadatafile == "") { out << "OTU"; } else { out << "Feature"; } - for (int i = 0; i < numaxes; i++) { out << '\t' << "axis" << (i+1); } + for (int i = 0; i < numaxes; i++) { out << '\t' << "axis" << (i+1) << "\tp-value"; } out << "\tlength" << endl; if (method == "pearson") { calcPearson(axes, out); } @@ -303,6 +305,8 @@ int CorrAxesCommand::execute(){ int CorrAxesCommand::calcPearson(map >& axes, ofstream& out) { try { + LinearAlgebra linear; + //find average of each axis - X vector averageAxes; averageAxes.resize(numaxes, 0.0); for (map >::iterator it = axes.begin(); it != axes.end(); it++) { @@ -317,7 +321,7 @@ int CorrAxesCommand::calcPearson(map >& axes, ofstream& ou //for each otu for (int i = 0; i < lookupFloat[0]->getNumBins(); i++) { - if (metadatafile == "") { out << i+1; } + if (metadatafile == "") { out << m->currentBinLabels[i]; } else { out << metadataLabels[i]; } //find the averages this otu - Y @@ -348,8 +352,15 @@ int CorrAxesCommand::calcPearson(map >& axes, ofstream& ou double denom = (sqrt(denomTerm1) * sqrt(denomTerm2)); r = numerator / denom; + + if (isnan(r) || isinf(r)) { r = 0.0; } + rValues[k] = r; out << '\t' << r; + + double sig = linear.calcPearsonSig(lookupFloat.size(), r); + + out << '\t' << sig; } double sum = 0; @@ -370,6 +381,9 @@ int CorrAxesCommand::calcPearson(map >& axes, ofstream& ou int CorrAxesCommand::calcSpearman(map >& axes, ofstream& out) { try { + LinearAlgebra linear; + vector sf; + //format data vector< map > tableX; tableX.resize(numaxes); map::iterator itTable; @@ -409,6 +423,7 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o vector ties; int rankTotal = 0; + double sfTemp = 0.0; for (int j = 0; j < scores[i].size(); j++) { rankTotal += (j+1); ties.push_back(scores[i][j]); @@ -420,6 +435,8 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o float thisrank = rankTotal / (float) ties.size(); rankAxes[ties[k].name].push_back(thisrank); } + int t = ties.size(); + sfTemp += (t*t*t-t); ties.clear(); rankTotal = 0; } @@ -432,13 +449,14 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o } } } + sf.push_back(sfTemp); } //for each otu for (int i = 0; i < lookupFloat[0]->getNumBins(); i++) { - if (metadatafile == "") { out << i+1; } + if (metadatafile == "") { out << m->currentBinLabels[i]; } else { out << metadataLabels[i]; } //find the ranks of this otu - Y @@ -466,6 +484,7 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o sort(otuScores.begin(), otuScores.end(), compareSpearman); + double sg = 0.0; map rankOtus; vector ties; int rankTotal = 0; @@ -480,6 +499,8 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o float thisrank = rankTotal / (float) ties.size(); rankOtus[ties[k].name] = thisrank; } + int t = ties.size(); + sg += (t*t*t-t); ties.clear(); rankTotal = 0; } @@ -514,9 +535,15 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o p = (SX2 + SY2 - di) / (2.0 * sqrt((SX2*SY2))); + if (isnan(p) || isinf(p)) { p = 0.0; } + out << '\t' << p; pValues[j] = p; + + double sig = linear.calcSpearmanSig(n, sf[j], sg, di); + out << '\t' << sig; + } double sum = 0; @@ -537,6 +564,8 @@ int CorrAxesCommand::calcSpearman(map >& axes, ofstream& o int CorrAxesCommand::calcKendall(map >& axes, ofstream& out) { try { + LinearAlgebra linear; + //format data vector< vector > scores; scores.resize(numaxes); for (map >::iterator it = axes.begin(); it != axes.end(); it++) { @@ -580,7 +609,7 @@ int CorrAxesCommand::calcKendall(map >& axes, ofstream& ou //for each otu for (int i = 0; i < lookupFloat[0]->getNumBins(); i++) { - if (metadatafile == "") { out << i+1; } + if (metadatafile == "") { out << m->currentBinLabels[i]; } else { out << metadataLabels[i]; } //find the ranks of this otu - Y @@ -650,10 +679,14 @@ int CorrAxesCommand::calcKendall(map >& axes, ofstream& ou } double p = (numCoor - numDisCoor) / (float) count; - + if (isnan(p) || isinf(p)) { p = 0.0; } + out << '\t' << p; pValues[j] = p; - + + double sig = linear.calcKendallSig(scores[j].size(), p); + + out << '\t' << sig; } double sum = 0; @@ -758,6 +791,8 @@ int CorrAxesCommand::eliminateZeroOTUS(vector& thisloo } //for each bin + vector newBinLabels; + string snumBins = toString(thislookup[0]->getNumBins()); for (int i = 0; i < thislookup[0]->getNumBins(); i++) { if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; } @@ -772,12 +807,25 @@ int CorrAxesCommand::eliminateZeroOTUS(vector& thisloo for (int j = 0; j < thislookup.size(); j++) { newLookup[j]->push_back(thislookup[j]->getAbundance(i), thislookup[j]->getGroup()); } + + //if there is a bin label use it otherwise make one + string binLabel = "Otu"; + string sbinNumber = toString(i+1); + if (sbinNumber.length() < snumBins.length()) { + int diff = snumBins.length() - sbinNumber.length(); + for (int h = 0; h < diff; h++) { binLabel += "0"; } + } + binLabel += sbinNumber; + if (i < m->currentBinLabels.size()) { binLabel = m->currentBinLabels[i]; } + + newBinLabels.push_back(binLabel); } } for (int j = 0; j < thislookup.size(); j++) { delete thislookup[j]; } thislookup = newLookup; + m->currentBinLabels = newBinLabels; return 0; @@ -862,14 +910,14 @@ int CorrAxesCommand::getMetadata(){ //read the first label, because it refers to the groups string columnLabel; iss >> columnLabel; m->gobble(iss); - + //save names of columns you are reading while (!iss.eof()) { iss >> columnLabel; m->gobble(iss); metadataLabels.push_back(columnLabel); } int count = metadataLabels.size(); - + //read rest of file while (!in.eof()) { @@ -878,7 +926,7 @@ int CorrAxesCommand::getMetadata(){ string group = ""; in >> group; m->gobble(in); groupNames.push_back(group); - + SharedRAbundFloatVector* tempLookup = new SharedRAbundFloatVector(); tempLookup->setGroup(group); tempLookup->setLabel("1"); @@ -886,7 +934,6 @@ int CorrAxesCommand::getMetadata(){ for (int i = 0; i < count; i++) { float temp = 0.0; in >> temp; - tempLookup->push_back(temp, group); } @@ -899,12 +946,13 @@ int CorrAxesCommand::getMetadata(){ //remove any groups the user does not want, and set globaldata->groups with only valid groups SharedUtil* util; util = new SharedUtil(); - - util->setGroups(m->Groups, groupNames); + Groups = m->getGroups(); + util->setGroups(Groups, groupNames); + m->setGroups(Groups); for (int i = 0; i < lookupFloat.size(); i++) { //if this sharedrabund is not from a group the user wants then delete it. - if (util->isValidGroup(lookupFloat[i]->getGroup(), m->Groups) == false) { + if (util->isValidGroup(lookupFloat[i]->getGroup(), m->getGroups()) == false) { delete lookupFloat[i]; lookupFloat[i] = NULL; lookupFloat.erase(lookupFloat.begin()+i); i--;