]> git.donarmstrong.com Git - mothur.git/blobdiff - pcoacommand.cpp
pca command
[mothur.git] / pcoacommand.cpp
index ee0a5462de87049f6b026546492203be630ea8d7..d1d919c93d5f43086658a86cf79f11a09de8ae9f 100644 (file)
@@ -8,22 +8,22 @@
  *
  */
 
-#include "pcacommand.h"
+#include "pcoacommand.h"
 
 //**********************************************************************************************************************
-vector<string> PCACommand::getValidParameters(){       
+vector<string> PCOACommand::getValidParameters(){      
        try {
-               string Array[] =  {"phylip", "outputdir","inputdir"};
+               string Array[] =  {"phylip", "metric","outputdir","inputdir"};
                vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
                return myArray;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "getValidParameters");
+               m->errorOut(e, "PCOACommand", "getValidParameters");
                exit(1);
        }
 }
 //**********************************************************************************************************************
-PCACommand::PCACommand(){      
+PCOACommand::PCOACommand(){    
        try {
                abort = true;
                //initialize outputTypes
@@ -32,36 +32,36 @@ PCACommand::PCACommand(){
                outputTypes["loadings"] = tempOutNames;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "PCACommand");
+               m->errorOut(e, "PCOACommand", "PCOACommand");
                exit(1);
        }
 }
 //**********************************************************************************************************************
-vector<string> PCACommand::getRequiredParameters(){    
+vector<string> PCOACommand::getRequiredParameters(){   
        try {
                string Array[] =  {"phylip"};
                vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
                return myArray;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "getRequiredParameters");
+               m->errorOut(e, "PCOACommand", "getRequiredParameters");
                exit(1);
        }
 }
 //**********************************************************************************************************************
-vector<string> PCACommand::getRequiredFiles(){ 
+vector<string> PCOACommand::getRequiredFiles(){        
        try {
                vector<string> myArray;
                return myArray;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "getRequiredFiles");
+               m->errorOut(e, "PCOACommand", "getRequiredFiles");
                exit(1);
        }
 }
 //**********************************************************************************************************************
 
-PCACommand::PCACommand(string option)  {
+PCOACommand::PCOACommand(string option)  {
        try {
                abort = false;
                
@@ -70,7 +70,7 @@ PCACommand::PCACommand(string option)  {
                
                else {
                        //valid paramters for this command
-                       string Array[] =  {"phylip","outputdir", "inputdir"};
+                       string Array[] =  {"phylip","metric","outputdir", "inputdir"};
                        vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
                        
                        OptionParser parser(option);
@@ -115,30 +115,37 @@ PCACommand::PCACommand(string option)  {
                        }
                        
                        //error checking on files       
-                       if (phylipfile == "")   { m->mothurOut("You must provide a distance file before running the pca command."); m->mothurOutEndLine(); abort = true; }              
+                       if (phylipfile == "")   { m->mothurOut("You must provide a distance file before running the pcoa command."); m->mothurOutEndLine(); abort = true; }             
+               
+                       string temp = validParameter.validFile(parameters, "metric", false);    if (temp == "not found"){       temp = "T";                             }
+                       metric = m->isTrue(temp); 
                }
 
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "PCACommand");
+               m->errorOut(e, "PCOACommand", "PCOACommand");
                exit(1);
        }
 }
 //**********************************************************************************************************************
-void PCACommand::help(){
+void PCOACommand::help(){
        try {
        
-               m->mothurOut("The pca command..."); m->mothurOutEndLine();
+               m->mothurOut("The pcoa command parameters are phylip and metric"); m->mothurOutEndLine();
+               m->mothurOut("The phylip parameter allows you to enter your distance file."); m->mothurOutEndLine();
+               m->mothurOut("The metric parameter allows indicate you if would like the pearson correlation coefficient calculated. Default=True"); m->mothurOutEndLine();
+               m->mothurOut("Example pcoa(phylip=yourDistanceFile).\n");
+               m->mothurOut("Note: No spaces between parameter labels (i.e. phylip), '=' and parameters (i.e.yourDistanceFile).\n\n");
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "help");
+               m->errorOut(e, "PCOACommand", "help");
                exit(1);
        }
 }
 //**********************************************************************************************************************
-PCACommand::~PCACommand(){}
+PCOACommand::~PCOACommand(){}
 //**********************************************************************************************************************
-int PCACommand::execute(){
+int PCOACommand::execute(){
        try {
        
                if (abort == true) { return 0; }
@@ -162,14 +169,13 @@ int PCACommand::execute(){
                vector<double> e;
                vector<vector<double> > G = D;
                vector<vector<double> > copy_G;
-               //int rank = D.size();
-               
-               m->mothurOut("\nProcessing...\n");
+                               
+               m->mothurOut("\nProcessing...\n\n");
                
                for(int count=0;count<2;count++){
-                       recenter(offset, D, G);         if (m->control_pressed) { return 0; }
-                       tred2(G, d, e);                         if (m->control_pressed) { return 0; }
-                       qtli(d, e, G);                          if (m->control_pressed) { return 0; }
+                       recenter(offset, D, 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;
                } 
@@ -180,6 +186,22 @@ int PCACommand::execute(){
                
                if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0; }
                
+               if (metric) {   
+                       
+                       for (int i = 1; i < 4; i++) {
+                                                       
+                               vector< vector<double> > EuclidDists = linearCalc.calculateEuclidianDistance(G, i); //G is the pcoa 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, D); //G is the pcoa file, D is the users distance matrix
+                               
+                               m->mothurOut("Pearson's coefficient using " + toString(i) + " axis: " + toString(corr)); m->mothurOutEndLine();
+                               
+                               if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0; }
+                       }
+               }
+               
                m->mothurOutEndLine();
                m->mothurOut("Output File Names: "); m->mothurOutEndLine();
                for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
@@ -188,34 +210,27 @@ int PCACommand::execute(){
                return 0;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "execute");
+               m->errorOut(e, "PCOACommand", "execute");
                exit(1);
        }
 }
 /*********************************************************************************************************************************/
 
-inline double SIGN(const double a, const double b)
-{
-    return b>=0 ? (a>=0 ? a:-a) : (a>=0 ? -a:a);
-}
-
-/*********************************************************************************************************************************/
-
-void PCACommand::get_comment(istream& f, char begin, char end){
+void PCOACommand::get_comment(istream& f, char begin, char end){
        try {
                char d=f.get();
                while(d != end){        d = f.get();    }
                d = f.peek();
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "get_comment");
+               m->errorOut(e, "PCOACommand", "get_comment");
                exit(1);
        }
 }      
 
 /*********************************************************************************************************************************/
 
-int PCACommand::read_phylip(istream& f, int square_m, vector<string>& name_list, vector<vector<double> >& d){
+int PCOACommand::read_phylip(istream& f, int square_m, vector<string>& name_list, vector<vector<double> >& d){
        try {
                //     int count1=0;
                //     int count2=0;
@@ -262,7 +277,7 @@ int PCACommand::read_phylip(istream& f, int square_m, vector<string>& name_list,
                return 0;
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "read_phylip");
+               m->errorOut(e, "PCOACommand", "read_phylip");
                exit(1);
        }
 
@@ -270,7 +285,7 @@ int PCACommand::read_phylip(istream& f, int square_m, vector<string>& name_list,
 
 /*********************************************************************************************************************************/
 
-void PCACommand::read(string fname, vector<string>& names, vector<vector<double> >& D){
+void PCOACommand::read(string fname, vector<string>& names, vector<vector<double> >& D){
        try {
                ifstream f;
                m->openInputFile(fname, f);
@@ -304,47 +319,14 @@ void PCACommand::read(string fname, vector<string>& names, vector<vector<double>
                read_phylip(f, q, names, D);
        }
                catch(exception& e) {
-               m->errorOut(e, "PCACommand", "read");
+               m->errorOut(e, "PCOACommand", "read");
                exit(1);
        }
 }
 
 /*********************************************************************************************************************************/
 
-double PCACommand::pythag(double a, double b)  {       return(pow(a*a+b*b,0.5));       }
-
-/*********************************************************************************************************************************/
-
-void PCACommand::matrix_mult(vector<vector<double> > first, vector<vector<double> > second, vector<vector<double> >& product){
-       try {
-               int first_rows = first.size();
-               int first_cols = first[0].size();
-               int second_cols = second[0].size();
-               
-               product.resize(first_rows);
-               for(int i=0;i<first_rows;i++){
-                       product[i].resize(second_cols);
-               }
-               
-               for(int i=0;i<first_rows;i++){
-                       for(int j=0;j<second_cols;j++){
-                               product[i][j] = 0.0;
-                               for(int k=0;k<first_cols;k++){
-                                       product[i][j] += first[i][k] * second[k][j];
-                               }
-                       }
-               }
-       }
-       catch(exception& e) {
-               m->errorOut(e, "PCACommand", "matrix_mult");
-               exit(1);
-       }
-
-}
-
-/*********************************************************************************************************************************/
-
-void PCACommand::recenter(double offset, vector<vector<double> > D, vector<vector<double> >& G){
+void PCOACommand::recenter(double offset, vector<vector<double> > D, vector<vector<double> >& G){
        try {
                int rank = D.size();
                
@@ -366,104 +348,11 @@ void PCACommand::recenter(double offset, vector<vector<double> > D, vector<vecto
                        }
                }
                
-               matrix_mult(C,A,A);
-               matrix_mult(A,C,G);
-       }
-       catch(exception& e) {
-               m->errorOut(e, "PCACommand", "recenter");
-               exit(1);
-       }
-
-}
-
-/*********************************************************************************************************************************/
-
-//  This function is taken from Numerical Recipes in C++ by Press et al., 2nd edition, pg. 479
-
-void PCACommand::tred2(vector<vector<double> >& a, vector<double>& d, vector<double>& e){
-       try {
-               double scale, hh, h, g, f;
-               
-               int n = a.size();
-               
-               d.resize(n);
-               e.resize(n);
-               
-               for(int i=n-1;i>0;i--){
-                       int l=i-1;
-                       h = scale = 0.0000;
-                       if(l>0){
-                               for(int k=0;k<l+1;k++){
-                                       scale += fabs(a[i][k]);
-                               }
-                               if(scale == 0.0){
-                                       e[i] = a[i][l];
-                               }
-                               else{
-                                       for(int k=0;k<l+1;k++){
-                                               a[i][k] /= scale;
-                                               h += a[i][k] * a[i][k];
-                                       }
-                                       f = a[i][l];
-                                       g = (f >= 0.0 ? -sqrt(h) : sqrt(h));
-                                       e[i] = scale * g;
-                                       h -= f * g;
-                                       a[i][l] = f - g;
-                                       f = 0.0;
-                                       for(int j=0;j<l+1;j++){
-                                               a[j][i] = a[i][j] / h;
-                                               g = 0.0;
-                                               for(int k=0;k<j+1;k++){
-                                                       g += a[j][k] * a[i][k];
-                                               }
-                                               for(int k=j+1;k<l+1;k++){
-                                                       g += a[k][j] * a[i][k];
-                                               }
-                                               e[j] = g / h;
-                                               f += e[j] * a[i][j];
-                                       }
-                                       hh = f / (h + h);
-                                       for(int j=0;j<l+1;j++){
-                                               f = a[i][j];
-                                               e[j] = g = e[j] - hh * f;
-                                               for(int k=0;k<j+1;k++){
-                                                       a[j][k] -= (f * e[k] + g * a[i][k]);
-                                               }
-                                       }
-                               }
-                       }
-                       else{
-                               e[i] = a[i][l];
-                       }
-                       
-                       d[i] = h;
-               }
-               
-               d[0] = 0.0000;
-               e[0] = 0.0000;
-               
-               for(int i=0;i<n;i++){
-                       int l = i;
-                       if(d[i] != 0.0){
-                               for(int j=0;j<l;j++){
-                                       g = 0.0000;
-                                       for(int k=0;k<l;k++){
-                                               g += a[i][k] * a[k][j];
-                                       }
-                                       for(int k=0;k<l;k++){
-                                               a[k][j] -= g * a[k][i];
-                                       }
-                               }
-                       }
-                       d[i] = a[i][i];
-                       a[i][i] = 1.0000;
-                       for(int j=0;j<l;j++){
-                               a[j][i] = a[i][j] = 0.0;
-                       }
-               }
+               A = linearCalc.matrix_mult(C,A);
+               G = linearCalc.matrix_mult(A,C);
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "tred2");
+               m->errorOut(e, "PCOACommand", "recenter");
                exit(1);
        }
 
@@ -471,90 +360,7 @@ void PCACommand::tred2(vector<vector<double> >& a, vector<double>& d, vector<dou
 
 /*********************************************************************************************************************************/
 
-//  This function is taken from Numerical Recipes in C++ by Press et al., 2nd edition, pg. 479
-
-void PCACommand::qtli(vector<double>& d, vector<double>& e, vector<vector<double> >& z) {
-       try {
-               int m, i, iter;
-               double s, r, p, g, f, dd, c, b;
-               
-               int n = d.size();
-               for(int i=1;i<=n;i++){
-                       e[i-1] = e[i];
-               }
-               e[n-1] = 0.0000;
-               
-               for(int l=0;l<n;l++){
-                       iter = 0;
-                       do {
-                               for(m=l;m<n-1;m++){
-                                       dd = fabs(d[m]) + fabs(d[m+1]);
-                                       if(fabs(e[m])+dd == dd) break;
-                               }
-                               if(m != l){
-                                       if(iter++ == 30) cerr << "Too many iterations in tqli\n";
-                                       g = (d[l+1]-d[l]) / (2.0 * e[l]);
-                                       r = pythag(g, 1.0);
-                                       g = d[m] - d[l] + e[l] / (g + SIGN(r,g));
-                                       s = c = 1.0;
-                                       p = 0.0000;
-                                       for(i=m-1;i>=l;i--){
-                                               f = s * e[i];
-                                               b = c * e[i];
-                                               e[i+1] = (r=pythag(f,g));
-                                               if(r==0.0){
-                                                       d[i+1] -= p;
-                                                       e[m] = 0.0000;
-                                                       break;
-                                               }
-                                               s = f / r;
-                                               c = g / r;
-                                               g = d[i+1] - p;
-                                               r = (d[i] - g) * s + 2.0 * c * b;
-                                               d[i+1] = g + ( p = s * r);
-                                               g = c * r - b;
-                                               for(int k=0;k<n;k++){
-                                                       f = z[k][i+1];
-                                                       z[k][i+1] = s * z[k][i] + c * f;
-                                                       z[k][i] = c * z[k][i] - s * f;
-                                               }
-                                       }
-                                       if(r == 0.00 && i >= l) continue;
-                                       d[l] -= p;
-                                       e[l] = g;
-                                       e[m] = 0.0;
-                               }
-                       } while (m != l);
-               }
-               
-               int k;
-               for(int i=0;i<n;i++){
-                       p=d[k=i];
-                       for(int j=i;j<n;j++){
-                               if(d[j] >= p){
-                                       p=d[k=j];
-                               }
-                       }
-                       if(k!=i){
-                               d[k]=d[i];
-                               d[i]=p;
-                               for(int j=0;j<n;j++){
-                                       p=z[j][i];
-                                       z[j][i] = z[j][k];
-                                       z[j][k] = p;
-                               }
-                       }
-               }
-       }
-       catch(exception& e) {
-               m->errorOut(e, "PCACommand", "qtli");
-               exit(1);
-       }
-}
-
-/*********************************************************************************************************************************/
-
-void PCACommand::output(string fnameRoot, vector<string> name_list, vector<vector<double> > G, vector<double> d) {
+void PCOACommand::output(string fnameRoot, vector<string> name_list, vector<vector<double> >& G, vector<double> d) {
        try {
                int rank = name_list.size();
                double dsum = 0.0000;
@@ -598,7 +404,7 @@ void PCACommand::output(string fnameRoot, vector<string> name_list, vector<vecto
                }
        }
        catch(exception& e) {
-               m->errorOut(e, "PCACommand", "output");
+               m->errorOut(e, "PCOACommand", "output");
                exit(1);
        }
 }