#include "corraxescommand.h"
#include "sharedutilities.h"
+#include "linearalgebra.h"
//**********************************************************************************************************************
vector<string> CorrAxesCommand::setParameters(){
try {
- CommandParameter paxes("axes", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(paxes);
- CommandParameter pshared("shared", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none",false,false); parameters.push_back(pshared);
- CommandParameter prelabund("relabund", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none",false,false); parameters.push_back(prelabund);
- CommandParameter pmetadata("metadata", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none",false,false); parameters.push_back(pmetadata);
- CommandParameter pnumaxes("numaxes", "Number", "", "3", "", "", "",false,false); parameters.push_back(pnumaxes);
- CommandParameter plabel("label", "String", "", "", "", "", "",false,false); parameters.push_back(plabel);
- CommandParameter pgroups("groups", "String", "", "", "", "", "",false,false); parameters.push_back(pgroups);
- CommandParameter pmethod("method", "Multiple", "pearson-spearman-kendall", "pearson", "", "", "",false,false); parameters.push_back(pmethod);
- CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
- CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
+ CommandParameter paxes("axes", "InputTypes", "", "", "none", "none", "none","corraxes",false,true,true); parameters.push_back(paxes);
+ CommandParameter pshared("shared", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none","",false,false,true); parameters.push_back(pshared);
+ CommandParameter prelabund("relabund", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none","",false,false,true); parameters.push_back(prelabund);
+ CommandParameter pmetadata("metadata", "InputTypes", "", "", "SharedRelMeta", "SharedRelMeta", "none","",false,false); parameters.push_back(pmetadata);
+ CommandParameter pnumaxes("numaxes", "Number", "", "3", "", "", "","",false,false); parameters.push_back(pnumaxes);
+ CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
+ CommandParameter pgroups("groups", "String", "", "", "", "", "","",false,false); parameters.push_back(pgroups);
+ CommandParameter pmethod("method", "Multiple", "pearson-spearman-kendall", "pearson", "", "", "","",false,false); parameters.push_back(pmethod);
+ CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir);
+ CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir);
vector<string> myArray;
for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
exit(1);
}
}
+//**********************************************************************************************************************
+string CorrAxesCommand::getOutputPattern(string type) {
+ try {
+ string pattern = "";
+
+ if (type == "corraxes") { pattern = "[filename],[tag],corr.axes"; }
+ else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
+
+ return pattern;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "CorrAxesCommand", "getOutputPattern");
+ exit(1);
+ }
+}
+
//**********************************************************************************************************************
CorrAxesCommand::CorrAxesCommand(){
try {
abort = true; calledHelp = true;
setParameters();
vector<string> tempOutNames;
- outputTypes["corr.axes"] = tempOutNames;
+ outputTypes["corraxes"] = tempOutNames;
}
catch(exception& e) {
m->errorOut(e, "CorrAxesCommand", "CorrAxesCommand");
}
vector<string> tempOutNames;
- outputTypes["corr.axes"] = tempOutNames;
+ outputTypes["corraxes"] = tempOutNames;
//if the user changes the input directory command factory will send this info to us in the output parameter
string inputDir = validParameter.validFile(parameters, "inputdir", false);
/*************************************************************************************/
// calc the r values //
/************************************************************************************/
-
- string outputFileName = outputDir + m->getRootName(m->getSimpleName(inputFileName)) + method + ".corr.axes";
- outputNames.push_back(outputFileName); outputTypes["corr.axes"].push_back(outputFileName);
+ map<string, string> variables;
+ variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputFileName));
+ variables["[tag]"] = method;
+ string outputFileName = getOutputFileName("corraxes", variables);
+ outputNames.push_back(outputFileName); outputTypes["corraxes"].push_back(outputFileName);
ofstream out;
m->openOutputFile(outputFileName, out);
out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
int CorrAxesCommand::calcPearson(map<string, vector<float> >& axes, ofstream& out) {
try {
+ LinearAlgebra linear;
+
//find average of each axis - X
vector<float> averageAxes; averageAxes.resize(numaxes, 0.0);
for (map<string, vector<float> >::iterator it = axes.begin(); it != axes.end(); it++) {
//for each otu
for (int i = 0; i < lookupFloat[0]->getNumBins(); i++) {
- if (metadatafile == "") { out << i+1; }
+ if (metadatafile == "") { out << m->currentSharedBinLabels[i]; }
else { out << metadataLabels[i]; }
//find the averages this otu - Y
rValues[k] = r;
out << '\t' << r;
- //signifigance calc - http://faculty.vassar.edu/lowry/ch4apx.html
- double temp = (1- (r*r)) / (double) (lookupFloat.size()-2);
- temp = sqrt(temp);
- double sig = r / temp;
- if (isnan(sig) || isinf(sig)) { sig = 0.0; }
+ double sig = linear.calcPearsonSig(lookupFloat.size(), r);
out << '\t' << sig;
}
int CorrAxesCommand::calcSpearman(map<string, vector<float> >& axes, ofstream& out) {
try {
+ LinearAlgebra linear;
+ vector<double> sf;
+
//format data
vector< map<float, int> > tableX; tableX.resize(numaxes);
map<float, int>::iterator itTable;
vector<spearmanRank> 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]);
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;
}
}
}
}
+ 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->currentSharedBinLabels[i]; }
else { out << metadataLabels[i]; }
//find the ranks of this otu - Y
sort(otuScores.begin(), otuScores.end(), compareSpearman);
+ double sg = 0.0;
map<string, float> rankOtus;
vector<spearmanRank> ties;
int rankTotal = 0;
float thisrank = rankTotal / (float) ties.size();
rankOtus[ties[k].name] = thisrank;
}
+ int t = ties.size();
+ sg += (t*t*t-t);
ties.clear();
rankTotal = 0;
}
pValues[j] = p;
- //signifigance calc - http://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient
- double temp = (lookupFloat.size()-2) / (double) (1- (p*p));
- temp = sqrt(temp);
- double sig = p*temp;
- if (isnan(sig) || isinf(sig)) { sig = 0.0; }
-
+ double sig = linear.calcSpearmanSig(n, sf[j], sg, di);
out << '\t' << sig;
}
int CorrAxesCommand::calcKendall(map<string, vector<float> >& axes, ofstream& out) {
try {
+ LinearAlgebra linear;
+
//format data
vector< vector<spearmanRank> > scores; scores.resize(numaxes);
for (map<string, vector<float> >::iterator it = axes.begin(); it != axes.end(); it++) {
//for each otu
for (int i = 0; i < lookupFloat[0]->getNumBins(); i++) {
- if (metadatafile == "") { out << i+1; }
+ if (metadatafile == "") { out << m->currentSharedBinLabels[i]; }
else { out << metadataLabels[i]; }
//find the ranks of this otu - Y
out << '\t' << p;
pValues[j] = p;
- //calc signif - zA - http://en.wikipedia.org/wiki/Kendall_tau_rank_correlation_coefficient#Significance_tests
- double numer = 3.0 * (numCoor - numDisCoor);
- int n = scores[j].size();
- double denom = n * (n-1) * (2*n + 5) / (double) 2.0;
- denom = sqrt(denom);
- double sig = numer / denom;
-
- if (isnan(sig) || isinf(sig)) { sig = 0.0; }
+ double sig = linear.calcKendallSig(scores[j].size(), p);
out << '\t' << sig;
}
for (int h = 0; h < diff; h++) { binLabel += "0"; }
}
binLabel += sbinNumber;
- if (i < m->currentBinLabels.size()) { binLabel = m->currentBinLabels[i]; }
+ if (i < m->currentSharedBinLabels.size()) { binLabel = m->currentSharedBinLabels[i]; }
newBinLabels.push_back(binLabel);
}
for (int j = 0; j < thislookup.size(); j++) { delete thislookup[j]; }
thislookup = newLookup;
- m->currentBinLabels = newBinLabels;
+ m->currentSharedBinLabels = newBinLabels;
return 0;
m->openInputFile(metadatafile, in);
string headerLine = m->getline(in); m->gobble(in);
- istringstream iss (headerLine,istringstream::in);
-
- //read the first label, because it refers to the groups
- string columnLabel;
- iss >> columnLabel; m->gobble(iss);
+ vector<string> pieces = m->splitWhiteSpace(headerLine);
//save names of columns you are reading
- while (!iss.eof()) {
- iss >> columnLabel; m->gobble(iss);
- metadataLabels.push_back(columnLabel);
+ for (int i = 1; i < pieces.size(); i++) {
+ metadataLabels.push_back(pieces[i]);
}
int count = metadataLabels.size();
projects / mothur.git / blobdiff