5 * Created by westcott on 9/23/10.
6 * Copyright 2010 Schloss Lab. All rights reserved.
10 #include "clusterfragmentscommand.h"
11 #include "needlemanoverlap.hpp"
13 //**********************************************************************************************************************
15 inline bool comparePriority(seqRNode first, seqRNode second) {
18 if (first.length > second.length) {
20 }else if (first.length == second.length) {
21 if (first.numIdentical > second.numIdentical) {
28 //**********************************************************************************************************************
29 vector<string> ClusterFragmentsCommand::setParameters(){
31 CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta);
32 CommandParameter pname("name", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pname);
33 CommandParameter pdiffs("diffs", "Number", "", "0", "", "", "",false,false); parameters.push_back(pdiffs);
34 CommandParameter ppercent("percent", "Number", "", "0", "", "", "",false,false); parameters.push_back(ppercent);
35 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
36 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
38 vector<string> myArray;
39 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
43 m->errorOut(e, "ClusterFragmentsCommand", "setParameters");
47 //**********************************************************************************************************************
48 string ClusterFragmentsCommand::getHelpString(){
50 string helpString = "";
51 helpString += "The cluster.fragments command groups sequences that are part of a larger sequence.\n";
52 helpString += "The cluster.fragments command outputs a new fasta and name file.\n";
53 helpString += "The cluster.fragments command parameters are fasta, name, diffs and percent. The fasta parameter is required, unless you have a valid current file. \n";
54 helpString += "The names parameter allows you to give a list of seqs that are identical. This file is 2 columns, first column is name or representative sequence, second column is a list of its identical sequences separated by commas.\n";
55 helpString += "The diffs parameter allows you to set the number of differences allowed, default=0. \n";
56 helpString += "The percent parameter allows you to set percentage of differences allowed, default=0. percent=2 means if the number of difference is less than or equal to two percent of the length of the fragment, then cluster.\n";
57 helpString += "You may use diffs and percent at the same time to say something like: If the number or differences is greater than 1 or more than 2% of the fragment length, don't merge. \n";
58 helpString += "The cluster.fragments command should be in the following format: \n";
59 helpString += "cluster.fragments(fasta=yourFastaFile, names=yourNamesFile) \n";
60 helpString += "Example cluster.fragments(fasta=amazon.fasta).\n";
61 helpString += "Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n";
65 m->errorOut(e, "ClusterFragmentsCommand", "getHelpString");
69 //**********************************************************************************************************************
70 ClusterFragmentsCommand::ClusterFragmentsCommand(){
72 abort = true; calledHelp = true;
74 vector<string> tempOutNames;
75 outputTypes["fasta"] = tempOutNames;
76 outputTypes["name"] = tempOutNames;
79 m->errorOut(e, "ClusterFragmentsCommand", "ClusterFragmentsCommand");
83 //**********************************************************************************************************************
84 ClusterFragmentsCommand::ClusterFragmentsCommand(string option) {
86 abort = false; calledHelp = false;
88 //allow user to run help
89 if(option == "help") { help(); abort = true; calledHelp = true; }
90 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
93 vector<string> myArray = setParameters();
95 OptionParser parser(option);
96 map<string, string> parameters = parser.getParameters();
98 ValidParameters validParameter;
99 map<string, string>::iterator it;
101 //check to make sure all parameters are valid for command
102 for (map<string, string>::iterator it2 = parameters.begin(); it2 != parameters.end(); it2++) {
103 if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; }
106 //initialize outputTypes
107 vector<string> tempOutNames;
108 outputTypes["fasta"] = tempOutNames;
109 outputTypes["name"] = tempOutNames;
111 //if the user changes the input directory command factory will send this info to us in the output parameter
112 string inputDir = validParameter.validFile(parameters, "inputdir", false);
113 if (inputDir == "not found"){ inputDir = ""; }
116 it = parameters.find("fasta");
117 //user has given a template file
118 if(it != parameters.end()){
119 path = m->hasPath(it->second);
120 //if the user has not given a path then, add inputdir. else leave path alone.
121 if (path == "") { parameters["fasta"] = inputDir + it->second; }
124 it = parameters.find("name");
125 //user has given a template file
126 if(it != parameters.end()){
127 path = m->hasPath(it->second);
128 //if the user has not given a path then, add inputdir. else leave path alone.
129 if (path == "") { parameters["name"] = inputDir + it->second; }
133 //check for required parameters
134 fastafile = validParameter.validFile(parameters, "fasta", true);
135 if (fastafile == "not found") {
136 fastafile = m->getFastaFile();
137 if (fastafile != "") { m->mothurOut("Using " + fastafile + " as input file for the fasta parameter."); m->mothurOutEndLine(); }
138 else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); m->mothurOutEndLine(); abort = true; }
140 else if (fastafile == "not open") { fastafile = ""; abort = true; }
142 //if the user changes the output directory command factory will send this info to us in the output parameter
143 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = m->hasPath(fastafile); }
145 //check for optional parameter and set defaults
146 // ...at some point should added some additional type checking...
147 namefile = validParameter.validFile(parameters, "name", true);
148 if (namefile == "not found") { namefile = ""; }
149 else if (namefile == "not open") { abort = true; }
150 else { readNameFile(); }
153 temp = validParameter.validFile(parameters, "diffs", false); if (temp == "not found"){ temp = "0"; }
154 convert(temp, diffs);
156 temp = validParameter.validFile(parameters, "percent", false); if (temp == "not found"){ temp = "0"; }
157 convert(temp, percent);
162 catch(exception& e) {
163 m->errorOut(e, "ClusterFragmentsCommand", "ClusterFragmentsCommand");
167 //**********************************************************************************************************************
168 int ClusterFragmentsCommand::execute(){
171 if (abort == true) { if (calledHelp) { return 0; } return 2; }
173 int start = time(NULL);
175 //reads fasta file and return number of seqs
176 int numSeqs = readFASTA(); //fills alignSeqs and makes all seqs active
178 if (m->control_pressed) { return 0; }
180 if (numSeqs == 0) { m->mothurOut("Error reading fasta file...please correct."); m->mothurOutEndLine(); return 0; }
182 //sort seqs by length of unaligned sequence
183 sort(alignSeqs.begin(), alignSeqs.end(), comparePriority);
187 //think about running through twice...
188 for (int i = 0; i < numSeqs; i++) {
190 if (alignSeqs[i].active) { //this sequence has not been merged yet
192 string iBases = alignSeqs[i].seq.getUnaligned();
194 //try to merge it with all smaller seqs
195 for (int j = i+1; j < numSeqs; j++) {
197 if (m->control_pressed) { return 0; }
199 if (alignSeqs[j].active) { //this sequence has not been merged yet
201 string jBases = alignSeqs[j].seq.getUnaligned();
203 if (isFragment(iBases, jBases)) {
205 alignSeqs[i].names += ',' + alignSeqs[j].names;
206 alignSeqs[i].numIdentical += alignSeqs[j].numIdentical;
208 alignSeqs[j].active = 0;
209 alignSeqs[j].numIdentical = 0;
215 //remove from active list
216 alignSeqs[i].active = 0;
219 if(i % 100 == 0) { m->mothurOut(toString(i) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); }
222 if(numSeqs % 100 != 0) { m->mothurOut(toString(numSeqs) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); }
225 string fileroot = outputDir + m->getRootName(m->getSimpleName(fastafile));
227 string newFastaFile = fileroot + "fragclust.fasta";
228 string newNamesFile = fileroot + "names";
230 if (m->control_pressed) { return 0; }
232 m->mothurOutEndLine();
233 m->mothurOut("Total number of sequences before cluster.fragments was " + toString(alignSeqs.size()) + "."); m->mothurOutEndLine();
234 m->mothurOut("cluster.fragments removed " + toString(count) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
236 printData(newFastaFile, newNamesFile);
238 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to cluster " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
240 if (m->control_pressed) { remove(newFastaFile.c_str()); remove(newNamesFile.c_str()); return 0; }
242 m->mothurOutEndLine();
243 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
244 m->mothurOut(newFastaFile); m->mothurOutEndLine();
245 m->mothurOut(newNamesFile); m->mothurOutEndLine();
246 outputNames.push_back(newFastaFile); outputNames.push_back(newNamesFile); outputTypes["fasta"].push_back(newFastaFile); outputTypes["name"].push_back(newNamesFile);
247 m->mothurOutEndLine();
249 //set fasta file as new current fastafile
251 itTypes = outputTypes.find("fasta");
252 if (itTypes != outputTypes.end()) {
253 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setFastaFile(current); }
256 itTypes = outputTypes.find("name");
257 if (itTypes != outputTypes.end()) {
258 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setNameFile(current); }
264 catch(exception& e) {
265 m->errorOut(e, "ClusterFragmentsCommand", "execute");
269 //***************************************************************************************************************
270 bool ClusterFragmentsCommand::isFragment(string seq1, string seq2){
272 bool fragment = false;
275 int pos = seq1.find(seq2);
276 if (pos != string::npos) { return true; }
277 //no match, no diffs wanted
278 else if ((diffs == 0) && (percent == 0)) { return false; }
279 else { //try aligning and see if you can find it
281 //find number of acceptable differences for this sequence fragment
283 if (diffs == 0) { //you didnt set diffs you want a percentage
284 totalDiffs = floor((seq2.length() * (percent / 100.0)));
285 }else if (percent == 0) { //you didn't set percent you want diffs
287 }else if ((percent != 0) && (diffs != 0)) { //you want both, set total diffs to smaller of 2
289 int percentDiff = floor((seq2.length() * (percent / 100.0)));
290 if (percentDiff < totalDiffs) { totalDiffs = percentDiff; }
293 Alignment* alignment = new NeedlemanOverlap(-1.0, 1.0, -1.0, (seq1.length()+totalDiffs+1));
295 //use needleman to align
296 alignment->align(seq2, seq1);
297 string tempSeq2 = alignment->getSeqAAln();
298 string temp = alignment->getSeqBAln();
304 int endPos = tempSeq2.length()-1;
305 for (int i = 0; i < tempSeq2.length(); i++) { if (isalpha(tempSeq2[i])) { startPos = i; break; } }
306 for (int i = tempSeq2.length()-1; i >= 0; i--) { if (isalpha(tempSeq2[i])) { endPos = i; break; } }
308 //count number of diffs
310 for (int i = startPos; i <= endPos; i++) {
311 if (tempSeq2[i] != temp[i]) { numDiffs++; }
314 if (numDiffs <= totalDiffs) { fragment = true; }
321 catch(exception& e) {
322 m->errorOut(e, "ClusterFragmentsCommand", "isFragment");
326 /**************************************************************************************************/
327 int ClusterFragmentsCommand::readFASTA(){
331 m->openInputFile(fastafile, inFasta);
333 while (!inFasta.eof()) {
335 if (m->control_pressed) { inFasta.close(); return 0; }
337 Sequence seq(inFasta); m->gobble(inFasta);
339 if (seq.getName() != "") { //can get "" if commented line is at end of fasta file
340 if (namefile != "") {
341 itSize = sizes.find(seq.getName());
343 if (itSize == sizes.end()) { m->mothurOut(seq.getName() + " is not in your names file, please correct."); m->mothurOutEndLine(); exit(1); }
345 seqRNode tempNode(itSize->second, seq, names[seq.getName()], seq.getUnaligned().length());
346 alignSeqs.push_back(tempNode);
348 }else { //no names file, you are identical to yourself
349 seqRNode tempNode(1, seq, seq.getName(), seq.getUnaligned().length());
350 alignSeqs.push_back(tempNode);
356 return alignSeqs.size();
359 catch(exception& e) {
360 m->errorOut(e, "ClusterFragmentsCommand", "readFASTA");
364 /**************************************************************************************************/
365 void ClusterFragmentsCommand::printData(string newfasta, string newname){
370 m->openOutputFile(newfasta, outFasta);
371 m->openOutputFile(newname, outNames);
373 for (int i = 0; i < alignSeqs.size(); i++) {
374 if (alignSeqs[i].numIdentical != 0) {
375 alignSeqs[i].seq.printSequence(outFasta);
376 outNames << alignSeqs[i].seq.getName() << '\t' << alignSeqs[i].names << endl;
383 catch(exception& e) {
384 m->errorOut(e, "ClusterFragmentsCommand", "printData");
388 /**************************************************************************************************/
390 void ClusterFragmentsCommand::readNameFile(){
393 m->openInputFile(namefile, in);
394 string firstCol, secondCol;
397 in >> firstCol >> secondCol; m->gobble(in);
398 names[firstCol] = secondCol;
401 for(int i=0;i<secondCol.size();i++){
402 if(secondCol[i] == ','){ size++; }
404 sizes[firstCol] = size;
408 catch(exception& e) {
409 m->errorOut(e, "ClusterFragmentsCommand", "readNameFile");
414 /**************************************************************************************************/