#include "inputdata.h"
//**********************************************************************************************************************
-vector<string> PCACommand::getValidParameters(){
+vector<string> PCACommand::setParameters(){
try {
- string Array[] = {"label", "groups","metric","outputdir","inputdir"};
- vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ CommandParameter pshared("shared", "InputTypes", "", "", "LRSS", "LRSS", "none","pca-loadings",false,false,true); parameters.push_back(pshared);
+ CommandParameter prelabund("relabund", "InputTypes", "", "", "LRSS", "LRSS", "none","pca-loadings",false,false,true); parameters.push_back(prelabund);
+ CommandParameter pgroups("groups", "String", "", "", "", "", "","",false,false); parameters.push_back(pgroups);
+ CommandParameter pmetric("metric", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pmetric);
+ CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
+ 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); }
return myArray;
}
catch(exception& e) {
- m->errorOut(e, "PCACommand", "getValidParameters");
+ m->errorOut(e, "PCACommand", "setParameters");
exit(1);
}
}
//**********************************************************************************************************************
-PCACommand::PCACommand(){
+string PCACommand::getHelpString(){
try {
- abort = true; calledHelp = true;
- vector<string> tempOutNames;
- outputTypes["pca"] = tempOutNames;
- outputTypes["loadings"] = tempOutNames;
+ string helpString = "";
+ helpString += "The pca command parameters are shared, relabund, label, groups and metric. shared or relabund is required unless you have a valid current file.";
+ helpString += "The label parameter is used to analyze specific labels in your input. Default is the first label in your shared or relabund file. Multiple labels may be separated by dashes.\n";
+ helpString += "The groups parameter allows you to specify which groups you would like analyzed. Groupnames are separated by dashes.\n";
+ helpString += "The metric parameter allows you to indicate if would like the pearson correlation coefficient calculated. Default=True";
+ helpString += "Example pca(groups=yourGroups).\n";
+ helpString += "Example pca(groups=A-B-C).\n";
+ helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n";
+ return helpString;
}
catch(exception& e) {
- m->errorOut(e, "PCACommand", "PCACommand");
+ m->errorOut(e, "PCACommand", "getHelpString");
exit(1);
}
}
//**********************************************************************************************************************
-vector<string> PCACommand::getRequiredParameters(){
- try {
- vector<string> myArray;
- return myArray;
- }
- catch(exception& e) {
- m->errorOut(e, "PCACommand", "getRequiredParameters");
- exit(1);
- }
+string PCACommand::getOutputPattern(string type) {
+ try {
+ string pattern = "";
+
+ if (type == "pca") { pattern = "[filename],[distance],pca.axes"; }
+ else if (type == "loadings") { pattern = "[filename],[distance],pca.loadings"; }
+ else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
+
+ return pattern;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "PCACommand", "getOutputPattern");
+ exit(1);
+ }
}
+
//**********************************************************************************************************************
-vector<string> PCACommand::getRequiredFiles(){
+PCACommand::PCACommand(){
try {
- string Array[] = {"shared","relabund","or"};
- vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
- return myArray;
+ abort = true; calledHelp = true;
+ setParameters();
+ vector<string> tempOutNames;
+ outputTypes["pca"] = tempOutNames;
+ outputTypes["loadings"] = tempOutNames;
}
catch(exception& e) {
- m->errorOut(e, "PCACommand", "getRequiredFiles");
+ m->errorOut(e, "PCACommand", "PCACommand");
exit(1);
}
}
try {
abort = false; calledHelp = false;
- globaldata = GlobalData::getInstance();
-
//allow user to run help
if(option == "help") { help(); abort = true; calledHelp = true; }
+ else if(option == "citation") { citation(); abort = true; calledHelp = true;}
else {
- //valid paramters for this command
- string Array[] = {"label","groups","metric","outputdir", "inputdir"};
- vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+ vector<string> myArray = setParameters();
OptionParser parser(option);
map<string, string> parameters = parser. getParameters();
for (it = parameters.begin(); it != parameters.end(); it++) {
if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
}
- //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); if (inputDir == "not found"){ inputDir = ""; }
-
+
//initialize outputTypes
vector<string> tempOutNames;
outputTypes["pca"] = tempOutNames;
outputTypes["loadings"] = tempOutNames;
- //make sure the user has already run the read.otu command
- if ((globaldata->getSharedFile() == "") && (globaldata->getRelAbundFile() == "")) {
- m->mothurOut("You must read a list and a group, shared or relabund file before you can use the pca command."); m->mothurOutEndLine(); abort = true;
+ //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);
+ if (inputDir == "not found"){ inputDir = ""; }
+ else {
+ string path;
+ it = parameters.find("shared");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = m->hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["shared"] = inputDir + it->second; }
+ }
+
+ it = parameters.find("relabund");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = m->hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["relabund"] = inputDir + it->second; }
+ }
}
- if (globaldata->getSharedFile() != "") { mode = "shared"; inputFile = globaldata->getSharedFile(); }
- if (globaldata->getRelAbundFile() != "") { mode = "relabund"; inputFile = globaldata->getRelAbundFile(); }
+ //check for required parameters
+ sharedfile = validParameter.validFile(parameters, "shared", true);
+ if (sharedfile == "not open") { sharedfile = ""; abort = true; }
+ else if (sharedfile == "not found") { sharedfile = ""; }
+ else { mode = "sharedfile"; inputFile = sharedfile; m->setSharedFile(sharedfile); }
+
+ relabundfile = validParameter.validFile(parameters, "relabund", true);
+ if (relabundfile == "not open") { relabundfile = ""; abort = true; }
+ else if (relabundfile == "not found") { relabundfile = ""; }
+ else { mode = "relabund"; inputFile = relabundfile; m->setRelAbundFile(relabundfile); }
+
+ if ((sharedfile == "") && (relabundfile == "")) {
+ //is there are current file available for any of these?
+ //give priority to shared, then list, then rabund, then sabund
+ //if there is a current shared file, use it
+ sharedfile = m->getSharedFile();
+ if (sharedfile != "") { inputFile = sharedfile; mode = "sharedfile"; m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); }
+ else {
+ relabundfile = m->getRelAbundFile();
+ if (relabundfile != "") { inputFile = relabundfile; mode = "relabund"; m->mothurOut("Using " + relabundfile + " as input file for the relabund parameter."); m->mothurOutEndLine(); }
+ else {
+ m->mothurOut("No valid current files. You must provide a relabund or shared file."); m->mothurOutEndLine();
+ abort = true;
+ }
+ }
+ }
+
//if the user changes the output directory command factory will send this info to us in the output parameter
outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
outputDir = "";
metric = m->isTrue(temp);
label = validParameter.validFile(parameters, "label", false);
- if (label == "not found") { label = ""; labels = globaldata->labels; if(labels.size() == 0) { m->mothurOut("You did not provide a label, I will use the first label in your inputfile."); m->mothurOutEndLine(); } }
+ if (label == "not found") { label = ""; if(labels.size() == 0) { m->mothurOut("You did not provide a label, I will use the first label in your inputfile."); m->mothurOutEndLine(); } }
else { m->splitAtDash(label, labels); }
groups = validParameter.validFile(parameters, "groups", false);
if (groups == "not found") { groups = ""; }
else { m->splitAtDash(groups, Groups); }
- globaldata->Groups = Groups;
+ m->setGroups(Groups);
}
}
}
//**********************************************************************************************************************
-void PCACommand::help(){
- try {
- m->mothurOut("The pca command can only be run after a successful read.otu command of a shared or relabund file."); m->mothurOutEndLine();
- m->mothurOut("The pca command parameters are label, groups and metric. No parameters are required."); m->mothurOutEndLine();
- m->mothurOut("The label parameter is used to analyze specific labels in your input. Default is the first label in your shared or relabund file. Multiple labels may be separated by dashes.\n");
- m->mothurOut("The groups parameter allows you to specify which groups you would like analyzed. Groupnames are separated by dashes.\n");
- m->mothurOut("The metric parameter allows indicate you if would like the pearson correlation coefficient calculated. Default=True"); m->mothurOutEndLine();
- m->mothurOut("Example pca(groups=yourGroups).\n");
- m->mothurOut("Example pca(groups=A-B-C).\n");
- m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n");
- }
- catch(exception& e) {
- m->errorOut(e, "PCACommand", "help");
- exit(1);
- }
-}
-//**********************************************************************************************************************
-PCACommand::~PCACommand(){}
-//**********************************************************************************************************************
int PCACommand::execute(){
try {
//get first line of shared file
vector< vector<double> > matrix;
InputData* input;
- if (mode == "shared") {
+ if (mode == "sharedfile") {
input = new InputData(inputFile, "sharedfile");
}else if (mode == "relabund") {
input = new InputData(inputFile, "relabund");
//as long as you are not at the end of the file or done wih the lines you want
while((lookupFloat[0] != NULL) && (userLabels.size() != 0)) {
- if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete input; for (int i = 0; i < lookupFloat.size(); i++) { delete lookupFloat[i]; } lookupFloat.clear(); return 0; }
+ if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } delete input; for (int i = 0; i < lookupFloat.size(); i++) { delete lookupFloat[i]; } lookupFloat.clear(); return 0; }
if(labels.count(lookupFloat[0]->getLabel()) == 1){
processedLabels.insert(lookupFloat[0]->getLabel());
}
- if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete input; for (int i = 0; i < lookupFloat.size(); i++) { delete lookupFloat[i]; } lookupFloat.clear(); return 0; }
+ if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } delete input; for (int i = 0; i < lookupFloat.size(); i++) { delete lookupFloat[i]; } lookupFloat.clear(); return 0; }
//output error messages about any remaining user labels
set<string>::iterator it;
for (int i = 0; i < lookupFloat.size(); i++) { if (lookupFloat[i] != NULL) { delete lookupFloat[i]; } } lookupFloat.clear();
delete input;
- if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+ if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
m->mothurOutEndLine();
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
exit(1);
}
}
+
/**********************************************************************************************************************
vector< vector<double> > PCACommand::createMatrix(vector<SharedRAbundFloatVector*> lookupFloat){
try {
}
}*/
//**********************************************************************************************************************
+
int PCACommand::process(vector<SharedRAbundFloatVector*>& lookupFloat){
try {
m->mothurOut("\nProcessing " + lookupFloat[0]->getLabel()); m->mothurOutEndLine();
+
+ int numOTUs = lookupFloat[0]->getNumBins();
+ int numSamples = lookupFloat.size();
- vector< vector<double> > matrix; matrix.resize(lookupFloat.size());
-
- ofstream out;
- string temp = outputDir + "matrix.transpose.out";
- m->openOutputFile(temp, out);
- out << "matrix" << endl;
+ vector< vector<double> > matrix(numSamples);
+ vector<double> colMeans(numOTUs);
- //fill matrix with shared files relative abundances
+ //fill matrix with shared relative abundances, re-center
for (int i = 0; i < lookupFloat.size(); i++) {
- for (int j = 0; j < lookupFloat[i]->getNumBins(); j++) {
- matrix[i].push_back(lookupFloat[i]->getAbundance(j));
- out << lookupFloat[i]->getAbundance(j) << '\t';
- }
- out << endl;
- }
- out << endl << endl << "transpose" << endl;
- vector< vector<double> > transposeMatrix; transposeMatrix.resize(matrix[0].size());
- for (int i = 0; i < transposeMatrix.size(); i++) {
- for (int j = 0; j < matrix.size(); j++) {
- transposeMatrix[i].push_back(matrix[j][i]);
- out << matrix[j][i] << '\t';
+ matrix[i].resize(numOTUs, 0);
+
+ for (int j = 0; j < numOTUs; j++) {
+ matrix[i][j] = lookupFloat[i]->getAbundance(j);
+ colMeans[j] += matrix[i][j];
}
- out << endl;
}
- matrix = linearCalc.matrix_mult(matrix, transposeMatrix);
+
+ for(int j=0;j<numOTUs;j++){
+ colMeans[j] = colMeans[j] / (double)numSamples;
+ }
- out << endl << endl << "matrix mult" << endl;
- for (int i = 0; i < matrix.size(); i++) {
- for (int j = 0; j < matrix[i].size(); j++) {
- out << matrix[i][j] << '\t';
+ vector<vector<double> > centered = matrix;
+ for(int i=0;i<numSamples;i++){
+ for(int j=0;j<numOTUs;j++){
+ centered[i][j] = centered[i][j] - colMeans[j];
}
- out << endl;
}
- out.close();
+
+
+ vector< vector<double> > transpose(numOTUs);
+ for (int i = 0; i < numOTUs; i++) {
+ transpose[i].resize(numSamples, 0);
+ for (int j = 0; j < numSamples; j++) {
+ transpose[i][j] = centered[j][i];
+ }
+ }
+
+ vector<vector<double> > crossProduct = linearCalc.matrix_mult(transpose, centered);
- double offset = 0.0000;
vector<double> d;
vector<double> e;
- vector<vector<double> > G = matrix;
- //vector<vector<double> > copy_G;
-
- for(int count=0;count<2;count++){
- linearCalc.recenter(offset, matrix, G); if (m->control_pressed) { return 0; }
- linearCalc.tred2(G, d, e); if (m->control_pressed) { return 0; }
- linearCalc.qtli(d, e, G); if (m->control_pressed) { return 0; }
- offset = d[d.size()-1];
- if(offset > 0.0) break;
- }
+
+ linearCalc.tred2(crossProduct, d, e); if (m->control_pressed) { return 0; }
+ linearCalc.qtli(d, e, crossProduct); if (m->control_pressed) { return 0; }
+
+ vector<vector<double> > X = linearCalc.matrix_mult(centered, crossProduct);
if (m->control_pressed) { return 0; }
string fbase = outputDir + m->getRootName(m->getSimpleName(inputFile));
- string outputFileName = fbase + lookupFloat[0]->getLabel();
- output(outputFileName, globaldata->Groups, G, d);
+ //string outputFileName = fbase + lookupFloat[0]->getLabel();
+ output(fbase, lookupFloat[0]->getLabel(), m->getGroups(), X, d);
if (metric) {
+ vector<vector<double> > observedEuclideanDistance = linearCalc.getObservedEuclideanDistance(centered);
+
for (int i = 1; i < 4; i++) {
- vector< vector<double> > EuclidDists = linearCalc.calculateEuclidianDistance(G, i); //G is the pca file
-
- if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
-
- double corr = linearCalc.calcPearson(EuclidDists, matrix); //G is the pca file, D is the users distance matrix
-
- m->mothurOut("Pearson's coefficient using " + toString(i) + " axis: " + toString(corr)); m->mothurOutEndLine();
+ vector< vector<double> > PCAEuclidDists = linearCalc.calculateEuclidianDistance(X, i); //G is the pca file
+ if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
+
+ double corr = linearCalc.calcPearson(PCAEuclidDists, observedEuclideanDistance);
+
m->mothurOut("Rsq " + toString(i) + " axis: " + toString(corr * corr)); m->mothurOutEndLine();
- if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+ if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
}
}
}
/*********************************************************************************************************************************/
-void PCACommand::output(string fnameRoot, vector<string> name_list, vector<vector<double> >& G, vector<double> d) {
+void PCACommand::output(string fbase, string label, vector<string> name_list, vector<vector<double> >& G, vector<double> d) {
try {
- int rank = name_list.size();
+
+ int numEigenValues = d.size();
double dsum = 0.0000;
- for(int i=0;i<rank;i++){
+ for(int i=0;i<numEigenValues;i++){
dsum += d[i];
- for(int j=0;j<rank;j++){
- if(d[j] >= 0) { G[i][j] *= pow(d[j],0.5); }
- else { G[i][j] = 0.00000; }
- }
}
- ofstream pcaData((fnameRoot+".pca.axes").c_str(), ios::trunc);
+ ofstream pcaData;
+ map<string, string> variables;
+ variables["[filename]"] = fbase;
+ variables["[distance]"] = label;
+ string pcaFileName = getOutputFileName("pca",variables);
+ m->openOutputFile(pcaFileName, pcaData);
pcaData.setf(ios::fixed, ios::floatfield);
pcaData.setf(ios::showpoint);
- outputNames.push_back(fnameRoot+".pca.axes");
- outputTypes["pca"].push_back(fnameRoot+".pca.axes");
+ outputNames.push_back(pcaFileName);
+ outputTypes["pca"].push_back(pcaFileName);
- ofstream pcaLoadings((fnameRoot+".pca.loadings").c_str(), ios::trunc);
+ ofstream pcaLoadings;
+ string loadingsFilename = getOutputFileName("loadings",variables);
+ m->openOutputFile(loadingsFilename, pcaLoadings);
pcaLoadings.setf(ios::fixed, ios::floatfield);
pcaLoadings.setf(ios::showpoint);
- outputNames.push_back(fnameRoot+".pca.loadings");
- outputTypes["loadings"].push_back(fnameRoot+".pca.loadings");
+ outputNames.push_back(loadingsFilename);
+ outputTypes["loadings"].push_back(loadingsFilename);
pcaLoadings << "axis\tloading\n";
- for(int i=0;i<rank;i++){
+ for(int i=0;i<numEigenValues;i++){
pcaLoadings << i+1 << '\t' << d[i] * 100.0 / dsum << endl;
}
pcaData << "group";
- for(int i=0;i<rank;i++){
+ for(int i=0;i<numEigenValues;i++){
pcaData << '\t' << "axis" << i+1;
}
pcaData << endl;
- for(int i=0;i<rank;i++){
+ for(int i=0;i<name_list.size();i++){
pcaData << name_list[i] << '\t';
- for(int j=0;j<rank;j++){
+ for(int j=0;j<numEigenValues;j++){
pcaData << G[i][j] << '\t';
}
pcaData << endl;