CommandParameter pclass("class", "String", "", "", "", "", "","",false,false); parameters.push_back(pclass);
CommandParameter psubclass("subclass", "String", "", "", "", "", "","",false,false); parameters.push_back(psubclass);
CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
- CommandParameter pclasses("classes", "String", "", "", "", "", "","",false,false); parameters.push_back(pclasses);
- CommandParameter palpha("anova_alpha", "Number", "", "0.05", "", "", "","",false,false); parameters.push_back(palpha);
- CommandParameter pwalpha("wilcoxon_alpha", "Number", "", "0.05", "", "", "","",false,false); parameters.push_back(pwalpha);
+ //CommandParameter pclasses("classes", "String", "", "", "", "", "","",false,false); parameters.push_back(pclasses);
+ CommandParameter palpha("aalpha", "Number", "", "0.05", "", "", "","",false,false); parameters.push_back(palpha);
+ CommandParameter pwalpha("walpha", "Number", "", "0.05", "", "", "","",false,false); parameters.push_back(pwalpha);
+
+ CommandParameter plda("lda", "Number", "", "2.0", "", "", "","",false,false); parameters.push_back(plda);
+ CommandParameter pwilc("wilc", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pwilc);
+ CommandParameter pnormmillion("norm", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pnormmillion);
+ CommandParameter piters("iters", "Number", "", "30", "", "", "","",false,false); parameters.push_back(piters);
+ //CommandParameter pwilcsamename("wilcsamename", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pwilcsamename);
+ CommandParameter pcurv("curv", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pcurv);
+ CommandParameter pfiters("fboots", "Number", "", "0.67", "", "", "","",false,false); parameters.push_back(pfiters);
+ CommandParameter pstrict("strict", "Multiple", "0-1-2", "0", "", "", "","",false,false); parameters.push_back(pstrict);
+ CommandParameter pminc("minc", "Number", "", "10", "", "", "","",false,false); parameters.push_back(pminc);
+ CommandParameter pmulticlass_strat("multiclass", "Multiple", "onevone-onevall", "onevall", "", "", "","",false,false); parameters.push_back(pmulticlass_strat);
+ //CommandParameter psubject("subject", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(psubject);
+
+
+ //not used in their current code, but in parameters
+ //CommandParameter pnlogs("nlogs", "Number", "", "3", "", "", "","",false,false); parameters.push_back(pnlogs);
+ //CommandParameter pranktec("ranktec", "Multiple", "lda-svm", "lda", "", "", "","",false,false); parameters.push_back(pranktec); // svm not implemented in their source yet.
+ //CommandParameter psvmnorm("svmnorm", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(psvmnorm); //not used because svm not implemented yet.
+
+
//every command must have inputdir and outputdir. This allows mothur users to redirect input and output files.
CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir);
CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir);
try {
string helpString = "";
helpString += "The lefse command allows you to ....\n";
- helpString += "The lefse command parameters are: shared, design, class, subclass, label, classes, wilcoxon_alpha, anova_alpha.\n";
+ helpString += "The lefse command parameters are: shared, design, class, subclass, label, walpha, aalpha, lda, wilc, iters, curv, fboots, strict, minc, multiclass and norm.\n";
helpString += "The class parameter is used to indicate the which category you would like used for the Kruskal Wallis analysis. If none is provided first category is used.\n";
- helpString += "The subclass parameter is used to indicate the .....If none is provided second category is used, or if only one category subclass is ignored. \n";
- helpString += "The classes parameter is used to indicate the classes you would like to use. Clases should be inputted in the following format: classes=label<value1|value2|value3>-label<value1|value2>. For example to include groups from treatment with value early or late and age= young or old. class=treatment<Early|Late>-age<young|old>.\n";
+ helpString += "The subclass parameter is used to indicate the .....If none is provided, second category is used, or if only one category subclass is ignored. \n";
+ helpString += "The aalpha parameter is used to set the alpha value for the Krukal Wallis Anova test Default=0.05. \n";
+ helpString += "The walpha parameter is used to set the alpha value for the Wilcoxon test. Default=0.05. \n";
+ helpString += "The lda parameter is used to set the threshold on the absolute value of the logarithmic LDA score. Default=2.0. \n";
+ helpString += "The wilc parameter is used to indicate whether to perform the Wilcoxon test. Default=T. \n";
+ helpString += "The iters parameter is used to set the number of bootstrap iteration for LDA. Default=30. \n";
+ //helpString += "The wilcsamename parameter is used to indicate whether perform the wilcoxon test only among the subclasses with the same name. Default=F. \n";
+ helpString += "The curv parameter is used to set whether perform the wilcoxon testing the Curtis's approach [BETA VERSION] Default=F. \n";
+ helpString += "The norm parameter is used to multiply relative abundances by 1000000. Recommended when very low values are present. Default=T. \n";
+ helpString += "The fboots parameter is used to set the subsampling fraction value for each bootstrap iteration. Default=0.67. \n";
+ helpString += "The strict parameter is used to set the multiple testing correction options. 0 no correction (more strict, default), 1 correction for independent comparisons, 2 correction for independent comparison. Options = 0,1,2. Default=0. \n";
+ helpString += "The minc parameter is used to minimum number of samples per subclass for performing wilcoxon test. Default=10. \n";
+ helpString += "The multiclass parameter is used to (for multiclass tasks) set whether the test is performed in a one-against-one ( onevone - more strict!) or in a one-against-all setting ( onevall - less strict). Default=onevall. \n";
+ //helpString += "The classes parameter is used to indicate the classes you would like to use. Classes should be inputted in the following format: classes=label<value1|value2|value3>-label<value1|value2>. For example to include groups from treatment with value early or late and age= young or old. class=treatment<Early|Late>-age<young|old>.\n";
helpString += "The label parameter is used to indicate which distances in the shared file you would like to use. labels are separated by dashes.\n";
helpString += "The lefse command should be in the following format: lefse(shared=final.an.shared, design=final.design, class=treatment, subclass=age).\n";
return helpString;
string pattern = "";
if (type == "summary") { pattern = "[filename],[distance],lefse_summary"; }
- else if (type == "kruskall-wallis") { pattern = "[filename],[distance],kruskall_wallis"; }
- else if (type == "wilcoxon") { pattern = "[filename],[distance],wilcoxon"; }
else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
return pattern;
setParameters();
vector<string> tempOutNames;
outputTypes["summary"] = tempOutNames;
- outputTypes["kruskall-wallis"] = tempOutNames;
- outputTypes["wilcoxon"] = tempOutNames;
}
catch(exception& e) {
m->errorOut(e, "LefseCommand", "LefseCommand");
vector<string> tempOutNames;
outputTypes["summary"] = tempOutNames;
- outputTypes["kruskall-wallis"] = tempOutNames;
- outputTypes["wilcoxon"] = 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);
if (mclass == "not found") { mclass = ""; }
subclass = validParameter.validFile(parameters, "subclass", false);
- if (subclass == "not found") { subclass = ""; }
+ if (subclass == "not found") { subclass = mclass; }
- classes = validParameter.validFile(parameters, "classes", false);
- if (classes == "not found") { classes = ""; }
-
- string temp = validParameter.validFile(parameters, "anova_alpha", false);
+ string temp = validParameter.validFile(parameters, "aalpha", false);
if (temp == "not found") { temp = "0.05"; }
m->mothurConvert(temp, anovaAlpha);
- temp = validParameter.validFile(parameters, "wilcoxon_alpha", false);
+ temp = validParameter.validFile(parameters, "walpha", false);
if (temp == "not found") { temp = "0.05"; }
m->mothurConvert(temp, wilcoxonAlpha);
+
+ temp = validParameter.validFile(parameters, "wilc", false);
+ if (temp == "not found") { temp = "T"; }
+ wilc = m->isTrue(temp);
+
+ temp = validParameter.validFile(parameters, "norm", false);
+ if (temp == "not found") { temp = "T"; }
+ normMillion = m->isTrue(temp);
+
+ //temp = validParameter.validFile(parameters, "subject", false);
+ //if (temp == "not found") { temp = "F"; }
+ //subject = m->isTrue(temp);
+ temp = validParameter.validFile(parameters, "lda", false);
+ if (temp == "not found") { temp = "2.0"; }
+ m->mothurConvert(temp, ldaThreshold);
+
+ temp = validParameter.validFile(parameters, "iters", false);
+ if (temp == "not found") { temp = "30"; }
+ m->mothurConvert(temp, iters);
+
+ temp = validParameter.validFile(parameters, "fboots", false);
+ if (temp == "not found") { temp = "0.67"; }
+ m->mothurConvert(temp, fBoots);
+
+ //temp = validParameter.validFile(parameters, "wilcsamename", false);
+ //if (temp == "not found") { temp = "F"; }
+ //wilcsamename = m->isTrue(temp);
+
+ temp = validParameter.validFile(parameters, "curv", false);
+ if (temp == "not found") { temp = "F"; }
+ curv = m->isTrue(temp);
+
+ temp = validParameter.validFile(parameters, "strict", false);
+ if (temp == "not found"){ temp = "0"; }
+ if ((temp != "0") && (temp != "1") && (temp != "2")) { m->mothurOut("Invalid strict option: choices are 0, 1 or 2."); m->mothurOutEndLine(); abort=true; }
+ else { m->mothurConvert(temp, strict); }
+
+ temp = validParameter.validFile(parameters, "minc", false);
+ if (temp == "not found") { temp = "10"; }
+ m->mothurConvert(temp, minC);
+
+ multiClassStrat = validParameter.validFile(parameters, "multiclass", false);
+ if (multiClassStrat == "not found"){ multiClassStrat = "onevall"; }
+ if ((multiClassStrat != "onevall") && (multiClassStrat != "onevone")) { m->mothurOut("Invalid multiclass option: choices are onevone or onevall."); m->mothurOutEndLine(); abort=true; }
}
}
int LefseCommand::execute(){
try {
+ srand(1982);
+ //for reading lefse formatted file and running in mothur for testing - pass number of rows used for design file
+ if (false) { makeShared(1); exit(1); }
if (abort == true) { if (calledHelp) { return 0; } return 2; }
DesignMap designMap(designfile);
- //if user set classes set groups=those classes
- if (classes != "") {
- map<string, vector<string> > thisClasses = m->parseClasses(classes);
- vector<string> groups = designMap.getNamesUnique(thisClasses);
- if (groups.size() != 0) { m->setGroups(groups); }
- else { m->mothurOut("[ERROR]: no groups meet your classes requirement, quitting.\n"); return 0; }
- }
//if user did not select class use first column
- if (mclass == "") { mclass = designMap.getDefaultClass(); m->mothurOut("\nYou did not provide a class, using " + mclass +".\n\n"); }
+ if (mclass == "") { mclass = designMap.getDefaultClass(); m->mothurOut("\nYou did not provide a class, using " + mclass +".\n\n"); if (subclass == "") { subclass = mclass; } }
InputData input(sharedfile, "sharedfile");
- vector<SharedRAbundVector*> lookup = input.getSharedRAbundVectors();
+ vector<SharedRAbundFloatVector*> lookup = input.getSharedRAbundFloatVectors();
string lastLabel = lookup[0]->getLabel();
//if the users enters label "0.06" and there is no "0.06" in their file use the next lowest label.
string saveLabel = lookup[0]->getLabel();
for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
- lookup = input.getSharedRAbundVectors(lastLabel);
+ lookup = input.getSharedRAbundFloatVectors(lastLabel);
m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
process(lookup, designMap);
if (m->control_pressed) { return 0; }
//get next line to process
- lookup = input.getSharedRAbundVectors();
+ lookup = input.getSharedRAbundFloatVectors();
}
if (m->control_pressed) { return 0; }
//run last label if you need to
if (needToRun == true) {
for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
- lookup = input.getSharedRAbundVectors(lastLabel);
+ lookup = input.getSharedRAbundFloatVectors(lastLabel);
m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
process(lookup, designMap);
m->mothurOut("Output File Names: "); m->mothurOutEndLine();
for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
m->mothurOutEndLine();
+ srand(time(NULL));
return 0;
}
}
//**********************************************************************************************************************
-int LefseCommand::process(vector<SharedRAbundVector*>& lookup, DesignMap& designMap) {
+int LefseCommand::process(vector<SharedRAbundFloatVector*>& lookup, DesignMap& designMap) {
try {
+ vector<string> classes;
+ vector<string> subclasses;
+ map<string, string> subclass2Class;
+ map<string, set<string> > class2SubClasses; //maps class name to vector of its subclasses
+ map<string, vector<int> > subClass2GroupIndex; //maps subclass name to vector of indexes in lookup from that subclass. old -> 1,2,3 means groups in location 1,2,3 of lookup are from old. Saves time below.
+ map<string, vector<int> > class2GroupIndex; //maps subclass name to vector of indexes in lookup from that class. old -> 1,2,3 means groups in location 1,2,3 of lookup are from old. Saves time below.
+ if (normMillion) { normalize(lookup); }
+ for (int j = 0; j < lookup.size(); j++) {
+ string group = lookup[j]->getGroup();
+ string treatment = designMap.get(group, mclass); //get value for this group in this category
+ string thisSub = designMap.get(group, subclass);
+ map<string, string>::iterator it = subclass2Class.find(thisSub);
+ if (it == subclass2Class.end()) {
+ subclass2Class[thisSub] = treatment;
+ vector<int> temp; temp.push_back(j);
+ subClass2GroupIndex[thisSub] = temp;
+ }
+ else {
+ if (it->second != treatment) {
+ //m->mothurOut("[WARNING]: subclass " + thisSub + " has members in " + it->second + " and " + treatment + ". Subclass members must be from the same class for Wilcoxon. Changing " + thisSub + " to " + treatment + "_" + thisSub + ".\n");
+ thisSub = treatment + "_" + thisSub;
+ subclass2Class[thisSub] = treatment;
+ vector<int> temp; temp.push_back(j);
+ subClass2GroupIndex[thisSub] = temp;
+ }else { subClass2GroupIndex[thisSub].push_back(j); }
+ }
+
+ map<string, set<string> >::iterator itClass = class2SubClasses.find(treatment);
+ if (itClass == class2SubClasses.end()) {
+ set<string> temp; temp.insert(thisSub);
+ class2SubClasses[treatment] = temp;
+ vector<int> temp2; temp2.push_back(j);
+ class2GroupIndex[treatment] = temp2;
+ classes.push_back(treatment);
+ }else{
+ class2SubClasses[treatment].insert(thisSub);
+ class2GroupIndex[treatment].push_back(j);
+ }
+ }
+ //sort classes so order is right
+ sort(classes.begin(), classes.end());
+
+ vector< vector<double> > means = getMeans(lookup, class2GroupIndex); //[numOTUs][classes] - classes in same order as class2GroupIndex
+
//run kruskal wallis on each otu
- vector<int> significantOtuLabels = runKruskalWallis(lookup, designMap);
+ map<int, double> significantOtuLabels = runKruskalWallis(lookup, designMap);
+
+ int numSigBeforeWilcox = significantOtuLabels.size();
+
+ if (m->debug) { m->mothurOut("[DEBUG]: completed Kruskal Wallis\n"); }
//check for subclass
- if (subclass != "") { significantOtuLabels = runWilcoxon(lookup, designMap, significantOtuLabels); }
+ string wilcoxString = "";
+ if ((subclass != "") && wilc) { significantOtuLabels = runWilcoxon(lookup, designMap, significantOtuLabels, class2SubClasses, subClass2GroupIndex, subclass2Class); wilcoxString += " ( " + toString(numSigBeforeWilcox) + " ) before internal wilcoxon"; }
+
+ int numSigAfterWilcox = significantOtuLabels.size();
+
+ if (m->debug) { m->mothurOut("[DEBUG]: completed Wilcoxon\n"); }
+
+ m->mothurOut("\nNumber of significantly discriminative features: " + toString(numSigAfterWilcox) + wilcoxString + ".\n");
+
+ map<int, double> sigOTUSLDA;
+ if (numSigAfterWilcox > 0) {
+ sigOTUSLDA = testLDA(lookup, significantOtuLabels, class2GroupIndex, subClass2GroupIndex);
+ m->mothurOut("Number of discriminative features with abs LDA score > " + toString(ldaThreshold) + " : " + toString(significantOtuLabels.size()) + ".\n");
+ }
+ else { m->mothurOut("No features with significant differences between the classes.\n"); }
+
+ if (m->debug) { m->mothurOut("[DEBUG]: completed lda\n"); }
+
+ printResults(means, significantOtuLabels, sigOTUSLDA, lookup[0]->getLabel(), classes);
return 0;
}
}
}
//**********************************************************************************************************************
-
-vector<int> LefseCommand::runKruskalWallis(vector<SharedRAbundVector*>& lookup, DesignMap& designMap) {
+int LefseCommand::normalize(vector<SharedRAbundFloatVector*>& lookup) {
try {
- map<string, string> variables;
- variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile));
- variables["[distance]"] = lookup[0]->getLabel();
- string outputFileName = getOutputFileName("kruskall-wallis",variables);
+ vector<double> mul;
+ for (int i = 0; i < lookup.size(); i++) {
+ double sum = 0.0;
+ for (int j = 0; j < lookup[i]->getNumBins(); j++) { sum += lookup[i]->getAbundance(j); }
+ mul.push_back(1000000.0/sum);
+ }
- ofstream out;
- m->openOutputFile(outputFileName, out);
- outputNames.push_back(outputFileName); outputTypes["kruskall-wallis"].push_back(outputFileName);
- out << "OTULabel\tKW\tPvalue\n";
+ for (int i = 0; i < lookup.size(); i++) {
+ for (int j = 0; j < lookup[i]->getNumBins(); j++) {
+ lookup[i]->set(j, lookup[i]->getAbundance(j)*mul[i], lookup[i]->getGroup());
+ }
+ }
- vector<int> significantOtuLabels;
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "normalize");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+map<int, double> LefseCommand::runKruskalWallis(vector<SharedRAbundFloatVector*>& lookup, DesignMap& designMap) {
+ try {
+ map<int, double> significantOtuLabels;
int numBins = lookup[0]->getNumBins();
//sanity check to make sure each treatment has a group in the shared file
set<string> treatments;
double pValue = 0.0;
double H = linear.calcKruskalWallis(values, pValue);
-
- //output H and signifigance
- out << m->currentBinLabels[i] << '\t' << H << '\t' << pValue << endl;
-
- if (pValue < anovaAlpha) { significantOtuLabels.push_back(i); }
+
+ if (pValue < anovaAlpha) { significantOtuLabels[i] = pValue; }
}
- out.close();
return significantOtuLabels;
}
}
//**********************************************************************************************************************
//assumes not neccessarily paired
-vector<int> LefseCommand::runWilcoxon(vector<SharedRAbundVector*>& lookup, DesignMap& designMap, vector<int> bins) {
+map<int, double> LefseCommand::runWilcoxon(vector<SharedRAbundFloatVector*>& lookup, DesignMap& designMap, map<int, double> bins, map<string, set<string> >& class2SubClasses, map<string, vector<int> >& subClass2GroupIndex, map<string, string> subclass2Class) {
try {
- LinearAlgebra linear;
- vector<int> significantOtuLabels;
+ map<int, double> significantOtuLabels;
+ map<int, double>::iterator it;
//if it exists and meets the following requirements run Wilcoxon
/*
1. Subclass members all belong to same main class
anything else
*/
- vector<string> subclasses;
- map<string, string> subclass2Class;
- map<string, int> subclassCounts;
- map<string, vector<int> > subClass2GroupIndex; //maps subclass name to vector of indexes in lookup from that subclass. old -> 1,2,3 means groups in location 1,2,3 of lookup are from old. Saves time below.
- bool error = false;
- for (int j = 0; j < lookup.size(); j++) {
- string group = lookup[j]->getGroup();
- string treatment = designMap.get(group, mclass); //get value for this group in this category
- string thisSub = designMap.get(group, subclass);
- map<string, string>::iterator it = subclass2Class.find(thisSub);
- if (it == subclass2Class.end()) {
- subclass2Class[thisSub] = treatment;
- subclassCounts[thisSub] = 1;
- vector<int> temp; temp.push_back(j);
- subClass2GroupIndex[thisSub] = temp;
+
+ int numBins = lookup[0]->getNumBins();
+ for (int i = 0; i < numBins; i++) {
+ if (m->control_pressed) { break; }
+
+ it = bins.find(i);
+ if (it != bins.end()) { //flagged in Kruskal Wallis
+
+ vector<float> abunds; for (int j = 0; j < lookup.size(); j++) { abunds.push_back(lookup[j]->getAbundance(i)); }
+
+ bool sig = testOTUWilcoxon(class2SubClasses, abunds, subClass2GroupIndex, subclass2Class);
+ if (sig) { significantOtuLabels[i] = it->second; }
+
+ }//bins flagged from kw
+ }//for bins
+
+ return significantOtuLabels;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "runWilcoxon");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+//lefse.py - test_rep_wilcoxon_r function
+bool LefseCommand::testOTUWilcoxon(map<string, set<string> >& class2SubClasses, vector<float> abunds, map<string, vector<int> >& subClass2GroupIndex, map<string, string> subclass2Class) {
+ try {
+ int totalOk = 0;
+ double alphaMtc = wilcoxonAlpha;
+ vector< set<string> > allDiffs;
+ LinearAlgebra linear;
+
+ //for each subclass comparision
+ map<string, set<string> >::iterator itB;
+ for(map<string, set<string> >::iterator it=class2SubClasses.begin();it!=class2SubClasses.end();it++){
+ itB = it;itB++;
+ for(itB;itB!=class2SubClasses.end();itB++){
+ if (m->control_pressed) { return false; }
+ bool first = true;
+ int dirCmp = 0; // not set?? dir_cmp = "not_set" # 0=notset or none, 1=true, 2=false.
+ int curv_sign = 0;
+ int ok = 0;
+ int count = 0;
+ for (set<string>::iterator itClass1 = (it->second).begin(); itClass1 != (it->second).end(); itClass1++) {
+ bool br = false;
+ for (set<string>::iterator itClass2 = (itB->second).begin(); itClass2 != (itB->second).end(); itClass2++) {
+ string subclass1 = *itClass1;
+ string subclass2 = *itClass2;
+ count++;
+
+ if (m->debug) { m->mothurOut( "[DEBUG comparing " + it->first + "-" + *itClass1 + " to " + itB->first + "-" + *itClass2 + "\n"); }
+
+ string treatment1 = subclass2Class[subclass1];
+ string treatment2 = subclass2Class[subclass2];
+ int numSubs1 = class2SubClasses[treatment1].size();
+ int numSubs2 = class2SubClasses[treatment2].size();
+
+ //if mul_cor != 0: alpha_mtc = th*l_subcl1*l_subcl2 if mul_cor == 2 else 1.0-math.pow(1.0-th,l_subcl1*l_subcl2)
+ if (strict != 0) { alphaMtc = wilcoxonAlpha * numSubs1 * numSubs2 ; }
+ if (strict == 2) {}else{ alphaMtc = 1.0-pow((1.0-wilcoxonAlpha),(double)(numSubs1 * numSubs2)); }
+
+ //fill x and y with this comparisons data
+ vector<double> x; vector<double> y;
+
+ //fill x and y
+ vector<int> xIndexes = subClass2GroupIndex[subclass1]; //indexes in lookup for this subclass
+ vector<int> yIndexes = subClass2GroupIndex[subclass2]; //indexes in lookup for this subclass
+ for (int k = 0; k < yIndexes.size(); k++) { y.push_back(abunds[yIndexes[k]]); }
+ for (int k = 0; k < xIndexes.size(); k++) { x.push_back(abunds[xIndexes[k]]); }
+
+ // med_comp = False
+ //if len(cl1) < min_c or len(cl2) < min_c:
+ //med_comp = True
+ bool medComp = false; // are there enough samples per subclass
+ if ((xIndexes.size() < minC) || (yIndexes.size() < minC)) { medComp = true; }
+
+ double sx = m->median(x);
+ double sy = m->median(y);
+
+ //if cl1[0] == cl2[0] and len(set(cl1)) == 1 and len(set(cl2)) == 1:
+ //tres, first = False, False
+ double pValue = 0.0;
+ double H = 0.0;
+ bool tres = true; //don't think this is set in the python source. Not sure how that is handled, but setting it here.
+ if ((x[0] == y[0]) && (x.size() == 1) && (y.size() == 1)) { tres = false; first = false; }
+ else if (!medComp) {
+ H = linear.calcWilcoxon(x, y, pValue);
+ if (pValue < (alphaMtc*2.0)) { tres = true; }
+ else { tres = false; }
+ }
+ /*if first:
+ first = False
+ if not curv and ( med_comp or tres ):
+ dir_cmp = sx < sy
+ if sx == sy: br = True
+ elif curv:
+ dir_cmp = None
+ if med_comp or tres:
+ curv_sign += 1
+ dir_cmp = sx < sy
+ else: br = True
+ elif not curv and med_comp:
+ if ((sx < sy) != dir_cmp or sx == sy): br = True
+ elif curv:
+ if tres and dir_cmp == None:
+ curv_sign += 1
+ dir_cmp = sx < sy
+ if tres and dir_cmp != (sx < sy):
+ br = True
+ curv_sign = -1
+ elif not tres or (sx < sy) != dir_cmp or sx == sy: br = True
+ */
+ int sxSy = 2; //false
+ if (sx<sy) { sxSy = 1; } //true
+
+ if (first) {
+ first = false;
+ if ((!curv) && (medComp || tres)) {
+ dirCmp = 2; if (sx<sy) { dirCmp = 1; } //dir_cmp = sx < sy
+ if (sx == sy) { br = true; }
+ }else if (curv) {
+ dirCmp = 0;
+ if (medComp || tres) {
+ curv_sign++;
+ dirCmp = 2; if (sx<sy) { dirCmp = 1; } //dir_cmp = sx < sy
+ }
+ }else { br = true; }
+ }else if (!curv && medComp) {
+ if (sxSy != dirCmp || sx == sy) { br = true; }
+ }else if (curv) {
+ if (tres && dirCmp == 0) { curv_sign++; }
+ dirCmp = 2; if (sx<sy) { dirCmp = 1; } //dir_cmp = sx < sy
+ if (tres && dirCmp != sxSy) { //if tres and dir_cmp != (sx < sy):
+ br = true;
+ curv_sign = -1;
+ }
+ }else if (!tres || sxSy != dirCmp || sx == sy) { br = true; } //elif not tres or (sx < sy) != dir_cmp or sx == sy: br = True
+ if (br) { break; }
+ ok++;
+ }//for class2 subclasses
+ if (br) { break; }
+ }//for class1 subclasses
+ bool diff = false;
+ if (curv) { diff = false; if (curv_sign > 0) { diff = true; } } //if curv: diff = curv_sign > 0
+ else { //else: diff = (ok == len(cl_hie[pair[1]])*len(cl_hie[pair[0]]))
+ diff = false;
+ if (ok == count) { diff = true; }
+ }
+ if (diff) { totalOk++; }
+ if (!diff && (multiClassStrat == "onevone")) { return false; }
+ if (diff && (multiClassStrat == "onevall")) { //all_diff.append(pair)
+ set<string> pair; pair.insert(it->first); pair.insert(itB->first);
+ allDiffs.push_back(pair);
+ }
+ }//classes
+ }//classes
+
+ if (multiClassStrat == "onevall") {
+ int tot_k = class2SubClasses.size();
+ for(map<string, set<string> >::iterator it=class2SubClasses.begin();it!=class2SubClasses.end();it++){
+ if (m->control_pressed) { return false; }
+ int nk = 0;
+ //is this class okay in all comparisons
+ for (int h = 0; h < allDiffs.size(); h++) {
+ if (allDiffs[h].count(it->first) != 0) { nk++; }
+ }
+ if (nk == (tot_k-1)) { return true; }//if nk == tot_k-1: return True
}
- else {
- subclassCounts[thisSub]++;
- subClass2GroupIndex[thisSub].push_back(j);
- if (it->second != treatment) {
- error = true;
- m->mothurOut("[ERROR]: subclass " + thisSub + " has members in " + it->second + " and " + treatment + ". Subclass members must be from the same class. Ignoring wilcoxon.\n");
+ return false;
+ }
+
+ return true;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "testOTUWilcoxon");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+//modelled after lefse.py test_lda_r function
+map<int, double> LefseCommand::testLDA(vector<SharedRAbundFloatVector*>& lookup, map<int, double> bins, map<string, vector<int> >& class2GroupIndex, map<string, vector<int> >& subClass2GroupIndex) {
+ try {
+ map<int, double> sigOTUS;
+ map<int, double>::iterator it;
+ LinearAlgebra linear;
+
+ int numBins = lookup[0]->getNumBins();
+ vector< vector<double> > adjustedLookup;
+
+ for (int i = 0; i < numBins; i++) {
+ if (m->control_pressed) { break; }
+
+ if (m->debug) { m->mothurOut("[DEBUG]: bin = " + toString(i) + "\n."); }
+
+ it = bins.find(i);
+ if (it != bins.end()) { //flagged in Kruskal Wallis and Wilcoxon(if we ran it)
+
+ if (m->debug) { m->mothurOut("[DEBUG]:flagged bin = " + toString(i) + "\n."); }
+
+ //fill x with this OTUs abundances
+ vector<double> x;
+ for (int j = 0; j < lookup.size(); j++) { x.push_back(lookup[j]->getAbundance(i)); }
+
+ //go through classes
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it != class2GroupIndex.end(); it++) {
+
+ if (m->debug) { m->mothurOut("[DEBUG]: class = " + it->first + "\n."); }
+
+ //max(float(feats['class'].count(c))*0.5,4)
+ //max(numGroups in this class*0.5, 4.0)
+ double necessaryNum = ((double)((it->second).size())*0.5);
+ if (4.0 > necessaryNum) { necessaryNum = 4.0; }
+
+ set<double> uniques;
+ for (int j = 0; j < (it->second).size(); j++) { uniques.insert(x[(it->second)[j]]); }
+
+ //if len(set([float(v[1]) for v in ff if v[0] == c])) > max(float(feats['class'].count(c))*0.5,4): continue
+ if ((double)(uniques.size()) > necessaryNum) { }
+ else {
+ //feats[k][i] = math.fabs(feats[k][i] + lrand.normalvariate(0.0,max(feats[k][i]*0.05,0.01)))
+ for (int j = 0; j < (it->second).size(); j++) { //(it->second) contains indexes of abundance for this class
+ double sigma = max((x[(it->second)[j]]*0.05), 0.01);
+ x[(it->second)[j]] = abs(x[(it->second)[j]] + linear.normalvariate(0.0, sigma));
+ }
+ }
}
+ adjustedLookup.push_back(x);
+ }
+ }
+
+ //go through classes
+ int minCl = 1e6;
+ map<int, string> indexToClass;
+ vector<string> classes;
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it != class2GroupIndex.end(); it++) {
+ //class with minimum number of groups
+ if ((it->second).size() < minCl) { minCl = (it->second).size(); }
+ for (int i = 0; i < (it->second).size(); i++) { indexToClass[(it->second)[i]] = it->first; }
+ classes.push_back(it->first);
+ }
+
+ int numGroups = lookup.size(); //lfk
+ int fractionNumGroups = numGroups * fBoots; //rfk
+ minCl = (int)((float)(minCl*fBoots*fBoots*0.05));
+ minCl = max(minCl, 1);
+
+ if (m->debug) { m->mothurOut("[DEBUG]: about to start iters. \n."); }
+
+ vector< vector< vector<double> > > results;//[iters][numComparison][numOTUs]
+ for (int j = 0; j < iters; j++) {
+ if (m->control_pressed) { return sigOTUS; }
+
+ if (m->debug) { m->mothurOut("[DEBUG]: iter = " + toString(j) + "\n."); }
+
+ //find "good" random vector
+ vector<int> rand_s;
+ for (int h = 0; h < 1000; h++) { //generate a vector of length fractionNumGroups with range 0 to numGroups-1
+ rand_s.clear();
+ for (int k = 0; k < fractionNumGroups; k++) { rand_s.push_back(m->getRandomIndex(numGroups-1)); }
+ if (!contastWithinClassesOrFewPerClass(adjustedLookup, rand_s, minCl, class2GroupIndex, indexToClass)) { h+=1000; } //break out of loop
}
+ if (m->control_pressed) { return sigOTUS; }
+
+ //print data in R input format for testing
+ if (false) {
+ vector<string> groups; for (int h = 0; h < rand_s.size(); h++) { groups.push_back(lookup[rand_s[h]]->getGroup()); }
+ printToCoutForRTesting(adjustedLookup, rand_s, class2GroupIndex, bins, subClass2GroupIndex, groups);
+ }
+
+ //for each pair of classes
+ vector< vector<double> > temp = lda(adjustedLookup, rand_s, indexToClass, classes); //[numComparison][numOTUs]
+ if (temp.size() != 0) { results.push_back(temp); }
}
- if (error) { return significantOtuLabels; }
+ if (m->control_pressed) { return sigOTUS; }
+
+ //m = max([numpy.mean([means[k][kk][p] for kk in range(boots)]) for p in range(len(pairs))])
+ int k = 0;
+ for (it = bins.begin(); it != bins.end(); it++) { //[numOTUs] - need to go through bins so we can tie adjustedLookup back to the binNumber. adjustedLookup[0] ->bins entry[0].
+ vector<double> averageForEachComparison; averageForEachComparison.resize(results[0].size(), 0.0);
+ double maxM = 0.0; //max of averages for each comparison
+ for (int j = 0; j < results[0].size(); j++) { //numComparisons
+ for (int i = 0; i < results.size(); i++) { //iters
+ averageForEachComparison[j]+= results[i][j][k];
+ }
+ averageForEachComparison[j] /= (double) results.size();
+ if (averageForEachComparison[j] > maxM) { maxM = averageForEachComparison[j]; }
+ }
+ //res[k] = math.copysign(1.0,m)*math.log(1.0+math.fabs(m),10)
+ double multiple = 1.0; if (maxM < 0.0) { multiple = -1.0; }
+ double resK = multiple * log10(1.0+abs(maxM));
+ if (resK > ldaThreshold) { sigOTUS[it->first] = resK; }
+ k++;
+ }
+ return sigOTUS;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "testLDA");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+vector< vector<double> > LefseCommand::getMeans(vector<SharedRAbundFloatVector*>& lookup, map<string, vector<int> >& class2GroupIndex) {
+ try {
int numBins = lookup[0]->getNumBins();
- vector<compGroup> comp;
- //find comparisons and fill comp
- map<string, int>::iterator itB;
- for(map<string, int>::iterator it=subclassCounts.begin();it!=subclassCounts.end();it++){
- itB = it;itB++;
- for(itB;itB!=subclassCounts.end();itB++){
- compGroup temp(it->first,itB->first);
- comp.push_back(temp);
- }
+ int numClasses = class2GroupIndex.size();
+ vector< vector<double> > means; //[numOTUS][classes]
+ means.resize(numBins);
+ for (int i = 0; i < means.size(); i++) { means[i].resize(numClasses, 0.0); }
+
+ map<int, string> indexToClass;
+ int count = 0;
+ //shortcut for vectors below
+ map<string, int> quickIndex;
+ vector<int> classCounts;
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it != class2GroupIndex.end(); it++) {
+ for (int i = 0; i < (it->second).size(); i++) { indexToClass[(it->second)[i]] = it->first; }
+ quickIndex[it->first] = count; count++;
+ classCounts.push_back((it->second).size());
+ }
+
+ for (int i = 0; i < numBins; i++) {
+ for (int j = 0; j < lookup.size(); j++) {
+ if (m->control_pressed) { return means; }
+ means[i][quickIndex[indexToClass[j]]] += lookup[j]->getAbundance(i);
+ }
+ }
+
+ for (int i = 0; i < numBins; i++) {
+ for (int j = 0; j < numClasses; j++) { means[i][j] /= (double) classCounts[j]; }
+ }
+
+ return means;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "getMeans");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+vector< vector<double> > LefseCommand::lda(vector< vector<double> >& adjustedLookup, vector<int> rand_s, map<int, string>& indexToClass, vector<string> classes) {
+ try {
+ //shortcut for vectors below
+ map<string, int> quickIndex;
+ for (int i = 0; i < classes.size(); i++) { quickIndex[classes[i]] = i; }
+
+ vector<string> randClass; //classes for rand sample
+ vector<int> counts; counts.resize(classes.size(), 0);
+ for (int i = 0; i < rand_s.size(); i++) {
+ string thisClass = indexToClass[rand_s[i]];
+ randClass.push_back(thisClass);
+ counts[quickIndex[thisClass]]++;
}
- int numComp = comp.size();
- if (numComp < 2) { m->mothurOut("[ERROR]: Need at least 2 subclasses, Ignoring Wilcoxon.\n");
- return significantOtuLabels; }
+ vector< vector<double> > a; //[numOTUs][numSampled]
+ for (int i = 0; i < adjustedLookup.size(); i++) {
+ vector<double> temp;
+ for (int j = 0; j < rand_s.size(); j++) {
+ temp.push_back(adjustedLookup[i][rand_s[j]]);
+ }
+ a.push_back(temp);
+ }
+
+ LinearAlgebra linear;
+ vector< vector<double> > means; bool ignore;
+ vector< vector<double> > scaling = linear.lda(a, randClass, means, ignore); //means are returned sorted, quickIndex sorts as well since it uses a map. means[class][otu] =
+ if (ignore) { scaling.clear(); return scaling; }
+ if (m->control_pressed) { return scaling; }
+
+ vector< vector<double> > w; w.resize(a.size()); //w.unit <- w/sqrt(sum(w^2))
+ double denom = 0.0;
+ for (int i = 0; i < scaling.size(); i++) { w[i].push_back(scaling[i][0]); denom += (w[i][0]*w[i][0]); }
+ denom = sqrt(denom);
+ for (int i = 0; i < w.size(); i++) { w[i][0] /= denom; } //[numOTUs][1] - w.unit
+
+ //robjects.r('LD <- xy.matrix%*%w.unit') [numSampled][numOtus] * [numOTUs][1]
+ vector< vector<double> > LD = linear.matrix_mult(linear.transpose(a), w);
+
+ //find means for each groups LDs
+ vector<double> LDMeans; LDMeans.resize(classes.size(), 0.0); //means[0] -> average for [group0].
+ for (int i = 0; i < LD.size(); i++) { LDMeans[quickIndex[randClass[i]]] += LD[i][0]; }
+ for (int i = 0; i < LDMeans.size(); i++) { LDMeans[i] /= (double) counts[i]; }
+
+ //calculate for each comparisons i.e. with groups A,B,C = AB, AC, BC = 3;
+ vector< vector<double> > results;// [numComparison][numOTUs]
+ for (int i = 0; i < LDMeans.size(); i++) {
+ for (int l = 0; l < i; l++) {
+ if (m->control_pressed) { return scaling; }
+ //robjects.r('effect.size <- abs(mean(LD[sub_d[,"class"]=="'+p[0]+'"]) - mean(LD[sub_d[,"class"]=="'+p[1]+'"]))')
+ double effectSize = abs(LDMeans[i] - LDMeans[l]);
+ //scal = robjects.r('wfinal <- w.unit * effect.size')
+ vector<double> compResults;
+ for (int j = 0; j < w.size(); j++) { //[numOTUs][1]
+ //coeff = [abs(float(v)) if not math.isnan(float(v)) else 0.0 for v in scal]
+ double coeff = abs(w[j][0]*effectSize); if (isnan(coeff) || isinf(coeff)) { coeff = 0.0; }
+ //gm = abs(res[p[0]][j] - res[p[1]][j]) - res is the means for each group for each otu
+ double gm = abs(means[i][j] - means[l][j]);
+ //means[k][i].append((gm+coeff[j])*0.5)
+ compResults.push_back((gm+coeff)*0.5);
+ }
+ results.push_back(compResults);
+ }
+ }
+
+ return results;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "lda");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+//modelled after lefse.py contast_within_classes_or_few_per_class function
+bool LefseCommand::contastWithinClassesOrFewPerClass(vector< vector<double> >& lookup, vector<int> rands, int minCl, map<string, vector<int> > class2GroupIndex, map<int, string> indexToClass) {
+ try {
+ set<string> cls;
+ int countFound = 0;
+
+ for (int i = 0; i < rands.size(); i++) { //fill cls with the classes represented in the random selection
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it != class2GroupIndex.end(); it++) {
+ if (m->inUsersGroups(rands[i], (it->second))) {
+ cls.insert(it->first);
+ countFound++;
+ }
+ }
+ }
+
+ //sanity check
+ if (rands.size() != countFound) { m->mothurOut("oops, should never get here, missing something.\n"); }
+
+ if (cls.size() < class2GroupIndex.size()) { return true; } //some classes are not present in sampling
+
+ for (set<string>::iterator it = cls.begin(); it != cls.end(); it++) {
+ if (cls.count(*it) < minCl) { return true; } //this sampling has class count below minimum
+ }
+ //for this otu
+ int numBins = lookup.size();
+ for (int i = 0; i < numBins; i++) {
+ if (m->control_pressed) { break; }
+
+ //break up random sampling by class
+ map<string, set<double> > class2Values; //maps class name -> set of abunds present in random sampling. F003Early -> 0.001, 0.003...
+ for (int j = 0; j < rands.size(); j++) {
+ class2Values[indexToClass[rands[j]]].insert(lookup[i][rands[j]]);
+ //rands[j] = index of randomly selected group in lookup, randIndex2Class[rands[j]] = class this group belongs to. lookup[rands[j]]->getAbundance(i) = abundance of this group for this OTU.
+ }
+ //are the unique values less than we want
+ //if (len(set(col)) <= min_cl and min_cl > 1) or (min_cl == 1 and len(set(col)) <= 1):
+ for (map<string, set<double> >::iterator it = class2Values.begin(); it != class2Values.end(); it++) {
+ if (((it->second).size() <= minCl && minCl > 1) || (minCl == 1 && (it->second).size() <= 1)) { return true; }
+ }
+ }
+
+ return false;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "contastWithinClassesOrFewPerClass");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+int LefseCommand::printResults(vector< vector<double> > means, map<int, double> sigKW, map<int, double> sigLDA, string label, vector<string> classes) {
+ try {
map<string, string> variables;
variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile));
- variables["[distance]"] = lookup[0]->getLabel();
- string outputFileName = getOutputFileName("wilcoxon",variables);
+ variables["[distance]"] = label;
+ string outputFileName = getOutputFileName("summary",variables);
+ ofstream out;
+ m->openOutputFile(outputFileName, out);
+ outputNames.push_back(outputFileName); outputTypes["summary"].push_back(outputFileName);
- ofstream out;
- m->openOutputFile(outputFileName, out);
- outputNames.push_back(outputFileName); outputTypes["wilcoxon"].push_back(outputFileName);
- out << "OTULabel\tComparision\tWilcoxon\tPvalue\n";
+ //output headers
+ out << "OTU\tLogMaxMean\tClass\tLDA\tpValue\n";
- for (int i = 0; i < numBins; i++) {
+ string temp = "";
+ for (int i = 0; i < means.size(); i++) { //[numOTUs][classes]
+ //find max mean of classes
+ double maxMean = -1.0; string maxClass = "none";
+ for (int j = 0; j < means[i].size(); j++) { if (means[i][j] > maxMean) { maxMean = means[i][j]; maxClass = classes[j]; } }
+
+ //str(math.log(max(max(v),1.0),10.0))
+ double logMaxMean = 1.0;
+ if (maxMean > logMaxMean) { logMaxMean = maxMean; }
+ logMaxMean = log10(logMaxMean);
+
+ out << m->currentBinLabels[i] << '\t' << logMaxMean << '\t';
+ if (m->debug) { temp = m->currentBinLabels[i] + '\t' + toString(logMaxMean) + '\t'; }
+
+ map<int, double>::iterator it = sigLDA.find(i);
+ if (it != sigLDA.end()) {
+ out << maxClass << '\t' << it->second << '\t' << sigKW[i] << endl; //sigLDA is a subset of sigKW so no need to look
+ if (m->debug) { temp += maxClass + '\t' + toString(it->second) + '\t' + toString(sigKW[i]) + '\n'; m->mothurOut(temp); temp = ""; }
+ }else { out << '-' << endl; }
+ }
+
+ out.close();
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "printResults");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+//printToCoutForRTesting(adjustedLookup, rand_s, class2GroupIndex, numBins);
+bool LefseCommand::printToCoutForRTesting(vector< vector<double> >& adjustedLookup, vector<int> rand_s, map<string, vector<int> >& class2GroupIndex, map<int, double> bins, map<string, vector<int> >& subClass2GroupIndex, vector<string> groups) {
+ try {
+ cout << "rand_s = ";
+ for (int h = 0; h < rand_s.size(); h++) { cout << rand_s[h] << '\t'; } cout << endl;
+
+ //print otu data
+ int count = 0;
+ for (map<int, double>::iterator it = bins.begin(); it != bins.end(); it++) {
+ if (m->control_pressed) { break; }
+
+ cout << m->currentBinLabels[it->first] << " <- c(";
+ for (int h = 0; h < rand_s.size()-1; h++) { cout << (adjustedLookup[count][rand_s[h]]) << ", "; }
+ cout << (adjustedLookup[count][rand_s[rand_s.size()-1]]) << ")\n";
+ count++;
+ }
+ /*
+ string tempOutput = "";
+ for (int h = 0; h < rand_s.size(); h++) {
+ //find class this index is in
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it!= class2GroupIndex.end(); it++) {
+ if (m->inUsersGroups(rand_s[h], (it->second)) ) { cout << (h+1) << " <- c(\"" +it->first + "\")\n" ; }
+ }
+ }*/
+
+
+ string tempOutput = "treatments <- c(";
+ for (int h = 0; h < rand_s.size(); h++) {
+ //find class this index is in
+ for (map<string, vector<int> >::iterator it = class2GroupIndex.begin(); it!= class2GroupIndex.end(); it++) {
+ if (m->inUsersGroups(rand_s[h], (it->second)) ) { tempOutput += "\"" +it->first + "\"" + ","; } //"\"" +it->first + "\""
+ }
+ }
+ tempOutput = tempOutput.substr(0, tempOutput.length()-1);
+ tempOutput += ")\n";
+ cout << tempOutput;
+
+ /*
+ if (subclass != "") {
+ string tempOutput = "sub <- c(";
+ for (int h = 0; h < rand_s.size(); h++) {
+ //find class this index is in
+ for (map<string, vector<int> >::iterator it = subClass2GroupIndex.begin(); it!= subClass2GroupIndex.end(); it++) {
+ if (m->inUsersGroups(rand_s[h], (it->second)) ) { tempOutput += "\"" +it->first + "\"" + ','; }
+ }
+ }
+ tempOutput = tempOutput.substr(0, tempOutput.length()-1);
+ tempOutput += ")\n";
+ cout << tempOutput;
+ }
+
+ if (subject) {
+ string tempOutput = "group <- c(";
+ for (int h = 0; h < groups.size(); h++) {
+ tempOutput += "\"" +groups[h] + "\"" + ',';
+ }
+ tempOutput = tempOutput.substr(0, tempOutput.length()-1);
+ tempOutput += ")\n";
+ cout << tempOutput;
+ }*/
+
+
+ //print data frame
+ tempOutput = "dat <- data.frame(";
+ for (map<int, double>::iterator it = bins.begin(); it != bins.end(); it++) {
if (m->control_pressed) { break; }
- if (m->inUsersGroups(i, bins)) { //flagged in Kruskal Wallis
+ tempOutput += "\"" + m->currentBinLabels[it->first] + "\"=" + m->currentBinLabels[it->first] + ",";
+ }
+ //tempOutput = tempOutput.substr(0, tempOutput.length()-1);
+ tempOutput += " class=treatments";
+ //if (subclass != "") { tempOutput += ", subclass=sub"; }
+ //if (subject) { tempOutput += ", subject=group"; }
+ tempOutput += ")\n";
+ cout << tempOutput;
- bool sig = false;
- //for each subclass comparision
- for (int j = 0; j < numComp; j++) {
- //fill x and y with this comparisons data
- vector<double> x; vector<double> y;
-
- cout << m->currentBinLabels[i] << '\t' << comp[j].getCombo() << " x <- (";
- //fill x and y
- vector<int> xIndexes = subClass2GroupIndex[comp[j].group1]; //indexes in lookup for this subclass
- for (int k = 0; k < xIndexes.size(); k++) { x.push_back(lookup[xIndexes[k]]->getAbundance(i)); cout << lookup[xIndexes[k]]->getAbundance(i) << ", "; }
- cout << ")\n";
-
- cout << m->currentBinLabels[i] << '\t' << comp[j].getCombo() << " y <- (";
-
- vector<int> yIndexes = subClass2GroupIndex[comp[j].group2]; //indexes in lookup for this subclass
- for (int k = 0; k < yIndexes.size(); k++) { y.push_back(lookup[yIndexes[k]]->getAbundance(i)); cout << lookup[yIndexes[k]]->getAbundance(i) << ", ";}
- cout << ")\n";
-
- double pValue = 0.0;
- double H = linear.calcWilcoxon(x, y, pValue);
+ tempOutput = "z <- suppressWarnings(mylda(as.formula(class ~ ";
+ for (map<int, double>::iterator it = bins.begin(); it != bins.end(); it++) {
+ if (m->control_pressed) { break; }
- //output H and signifigance
- if (!isnan(pValue)) { out << m->currentBinLabels[i] << '\t' << comp[j].getCombo() << '\t' << H << '\t' << pValue << endl; }
- else { out << m->currentBinLabels[i] << '\t' << comp[j].getCombo() << '\t' << H << '\t' << "NA" << endl; }
-
- //set sig - not sure how yet
- }
- if (sig) { significantOtuLabels.push_back(i); }
+ tempOutput += m->currentBinLabels[it->first] + "+";
+ }
+ tempOutput = tempOutput.substr(0, tempOutput.length()-1); //rip off extra plus sign
+ tempOutput += "), data = dat, tol = 1e-10))";
+ cout << tempOutput + "\nz\n";
+ cout << "w <- z$scaling[,1]\n"; //robjects.r('w <- z$scaling[,1]')
+ cout << "w.unit <- w/sqrt(sum(w^2))\n"; //robjects.r('w.unit <- w/sqrt(sum(w^2))')
+ cout << "ss <- dat[,-match(\"class\",colnames(dat))]\n"; //robjects.r('ss <- sub_d[,-match("class",colnames(sub_d))]')
+ //if (subclass != "") { cout << "ss <- ss[,-match(\"subclass\",colnames(ss))]\n"; }//robjects.r('ss <- ss[,-match("subclass",colnames(ss))]')
+ //if (subject) { cout << "ss <- ss[,-match(\"subject\",colnames(ss))]\n"; }//robjects.r('ss <- ss[,-match("subject",colnames(ss))]')
+ cout << "xy.matrix <- as.matrix(ss)\n"; //robjects.r('xy.matrix <- as.matrix(ss)')
+ cout << "LD <- xy.matrix%*%w.unit\n"; //robjects.r('LD <- xy.matrix%*%w.unit')
+ cout << "effect.size <- abs(mean(LD[dat[,\"class\"]==\"'+p[0]+'\"]) - mean(LD[dat[,\"class\"]==\"'+p[1]+'\"]))\n"; //robjects.r('effect.size <- abs(mean(LD[sub_d[,"class"]=="'+p[0]+'"]) - mean(LD[sub_d[,"class"]=="'+p[1]+'"]))')
+ cout << "wfinal <- w.unit * effect.size\n"; //scal = robjects.r('wfinal <- w.unit * effect.size')
+ cout << "mm <- z$means\n"; //rres = robjects.r('mm <- z$means')
+
+ return true;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "LefseCommand", "printToCoutForRTesting");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
+int LefseCommand::makeShared(int numDesignLines) {
+ try {
+ ifstream in;
+ m->openInputFile(sharedfile, in);
+ vector< vector<string> > lines;
+ for(int i = 0; i < numDesignLines; i++) {
+ if (m->control_pressed) { return 0; }
+
+ string line = m->getline(in);
+ cout << line << endl;
+ vector<string> pieces = m->splitWhiteSpace(line);
+ lines.push_back(pieces);
+ }
+
+ ofstream out;
+ m->openOutputFile(sharedfile+".design", out); out << "group" << '\t';
+ for (int j = 0; j < lines.size(); j++) { out << lines[j][0] << '\t'; } out << endl;
+ for (int j = 1; j < lines[0].size(); j++) {
+ out <<(j-1) << '\t';
+ for (int i = 0; i < lines.size(); i++) {
+ out << lines[i][j] << '\t';
}
+ out << endl;
}
out.close();
+ DesignMap design(sharedfile+".design");
- return significantOtuLabels;
+ vector<SharedRAbundFloatVector*> lookup;
+ for (int k = 0; k < lines[0].size()-1; k++) {
+ SharedRAbundFloatVector* temp = new SharedRAbundFloatVector();
+ temp->setLabel("0.03");
+ temp->setGroup(toString(k));
+ lookup.push_back(temp);
+ }
+
+ m->currentBinLabels.clear();
+ int count = 0;
+ while (!in.eof()) {
+ if (m->control_pressed) { return 0; }
+
+ string line = m->getline(in);
+ vector<string> pieces = m->splitWhiteSpace(line);
+
+ float sum = 0.0;
+ for (int i = 1; i < pieces.size(); i++) {
+ float value; m->mothurConvert(pieces[i], value);
+ sum += value;
+ }
+
+ if (sum != 0.0) {
+ //cout << count << '\t';
+ for (int i = 1; i < pieces.size(); i++) {
+ float value; m->mothurConvert(pieces[i], value);
+ lookup[i-1]->push_back(value, toString(i-1));
+ //cout << pieces[i] << '\t';
+ }
+ m->currentBinLabels.push_back(toString(count));
+ //m->currentBinLabels.push_back(pieces[0]);
+ //cout << line<< endl;
+ //cout << endl;
+ }
+ count++;
+ }
+ in.close();
+
+ for (int k = 0; k < lookup.size(); k++) {
+ //cout << "0.03" << '\t' << toString(k) << endl; lookup[k]->print(cout);
+ }
+
+ process(lookup, design);
+
+ return 0;
}
catch(exception& e) {
- m->errorOut(e, "LefseCommand", "runWilcoxon");
+ m->errorOut(e, "LefseCommand", "printToCoutForRTesting");
exit(1);
}
}
projects / mothur.git / blobdiff