5 * Created by Sarah Westcott on 5/15/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
8 * This version of nast does everything I think that the greengenes nast server does and then some. I have added the
9 * feature of allowing users to define their database, kmer size for searching, alignment penalty values and alignment
10 * method. This latter feature is perhaps most significant. nastPlus enables a user to use either a Needleman-Wunsch
11 * (non-affine gap penalty) or Gotoh (affine gap penalty) pairwise alignment algorithm. This is significant because it
12 * allows for a global alignment and not the local alignment provided by bLAst. Furthermore, it has the potential to
13 * provide a better alignment because of the banding method employed by blast (I'm not sure about this).
17 #include "aligncommand.h"
18 #include "sequence.hpp"
20 #include "gotohoverlap.hpp"
21 #include "needlemanoverlap.hpp"
22 #include "blastalign.hpp"
23 #include "noalign.hpp"
26 #include "nastreport.hpp"
29 //**********************************************************************************************************************
31 AlignCommand::AlignCommand(string option){
33 // globaldata = GlobalData::getInstance();
36 //allow user to run help
37 if(option == "help") { help(); abort = true; }
41 //valid paramters for this command
42 string AlignArray[] = {"template","candidate","search","ksize","align","match","mismatch","gapopen","gapextend", "processors","flip","threshold"};
43 vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
45 OptionParser parser(option);
46 map<string, string> parameters = parser.getParameters();
48 ValidParameters validParameter;
50 //check to make sure all parameters are valid for command
51 for (map<string, string>::iterator it = parameters.begin(); it != parameters.end(); it++) {
52 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
55 //check for required parameters
56 templateFileName = validParameter.validFile(parameters, "template", true);
57 if (templateFileName == "not found") {
58 mothurOut("template is a required parameter for the align.seqs command.");
62 else if (templateFileName == "not open") { abort = true; }
64 candidateFileName = validParameter.validFile(parameters, "candidate", true);
65 if (candidateFileName == "not found") {
66 mothurOut("candidate is a required parameter for the align.seqs command.");
70 else if (candidateFileName == "not open") { abort = true; }
72 //check for optional parameter and set defaults
73 // ...at some point should added some additional type checking...
75 temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; }
76 convert(temp, kmerSize);
78 temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; }
81 temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; }
82 convert(temp, misMatch);
84 temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; }
85 convert(temp, gapOpen);
87 temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; }
88 convert(temp, gapExtend);
90 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
91 convert(temp, processors);
93 temp = validParameter.validFile(parameters, "flip", false); if (temp == "not found"){ temp = "f"; }
96 temp = validParameter.validFile(parameters, "threshold", false); if (temp == "not found"){ temp = "0.80"; }
97 convert(temp, threshold);
99 search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; }
101 align = validParameter.validFile(parameters, "align", false); if (align == "not found"){ align = "needleman"; }
105 catch(exception& e) {
106 errorOut(e, "AlignCommand", "AlignCommand");
111 //**********************************************************************************************************************
113 AlignCommand::~AlignCommand(){
115 if (abort == false) {
116 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
122 //**********************************************************************************************************************
124 void AlignCommand::help(){
126 mothurOut("The align.seqs command reads a file containing sequences and creates an alignment file and a report file.\n");
127 mothurOut("The align.seqs command parameters are template, candidate, search, ksize, align, match, mismatch, gapopen and gapextend.\n");
128 mothurOut("The template and candidate parameters are required.\n");
129 mothurOut("The search parameter allows you to specify the method to find most similar template. Your options are: suffix, kmer and blast. The default is kmer.\n");
130 mothurOut("The align parameter allows you to specify the alignment method to use. Your options are: gotoh, needleman, blast and noalign. The default is needleman.\n");
131 mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n");
132 mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n");
133 mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n");
134 mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -1.0.\n");
135 mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -2.0.\n");
136 mothurOut("The flip parameter is used to specify whether or not you want mothur to try the reverse compement if a sequence falls below the threshold. The default is false.\n");
137 mothurOut("The threshold is used to specify a cutoff at which an alignment is deemed 'bad' and the reverse complement may be tried. \n");
138 mothurOut("If the flip parameter is set to true the reverse complement of the sequence is aligned and the better alignment is reported.\n");
139 mothurOut("The default for the threshold parameter is 0.80, meaning at least 80% of the bases must remain or the sequence is reported as potentially reversed.\n");
140 mothurOut("The align.seqs command should be in the following format: \n");
141 mothurOut("align.seqs(template=yourTemplateFile, candidate=yourCandidateFile, align=yourAlignmentMethod, search=yourSearchmethod, ksize=yourKmerSize, match=yourMatchBonus, mismatch=yourMismatchpenalty, gapopen=yourGapopenPenalty, gapextend=yourGapExtendPenalty) \n");
142 mothurOut("Example align.seqs(candidate=candidate.fasta, template=core.filtered, align=kmer, search=gotoh, ksize=8, match=2.0, mismatch=3.0, gapopen=-2.0, gapextend=-1.0)\n");
143 mothurOut("Note: No spaces between parameter labels (i.e. candidate), '=' and parameters (i.e.yourFastaFile).\n\n");
145 catch(exception& e) {
146 errorOut(e, "AlignCommand", "help");
152 //**********************************************************************************************************************
154 int AlignCommand::execute(){
156 if (abort == true) { return 0; }
158 templateDB = new AlignmentDB(templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch);
159 int longestBase = templateDB->getLongestBase();
161 if(align == "gotoh") { alignment = new GotohOverlap(gapOpen, gapExtend, match, misMatch, longestBase); }
162 else if(align == "needleman") { alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); }
163 else if(align == "blast") { alignment = new BlastAlignment(gapOpen, gapExtend, match, misMatch); }
164 else if(align == "noalign") { alignment = new NoAlign(); }
166 mothurOut(align + " is not a valid alignment option. I will run the command using needleman.");
168 alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase);
170 mothurOut("Aligning sequences...");
173 string alignFileName = candidateFileName.substr(0,candidateFileName.find_last_of(".")+1) + "align";
174 string reportFileName = candidateFileName.substr(0,candidateFileName.find_last_of(".")+1) + "align.report";
175 string accnosFileName = candidateFileName.substr(0,candidateFileName.find_last_of(".")+1) + "flip.accnos";
177 int numFastaSeqs = 0;
178 int start = time(NULL);
180 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
183 openInputFile(candidateFileName, inFASTA);
184 numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
187 lines.push_back(new linePair(0, numFastaSeqs));
189 driver(lines[0], alignFileName, reportFileName, accnosFileName);
191 //delete accnos file if its blank else report to user
192 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); }
194 mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + ".");
196 mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well.");
197 }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); }
202 vector<int> positions;
203 processIDS.resize(0);
206 openInputFile(candidateFileName, inFASTA);
209 while(!inFASTA.eof()){
210 input = getline(inFASTA);
211 if (input.length() != 0) {
212 if(input[0] == '>'){ int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
217 numFastaSeqs = positions.size();
219 int numSeqsPerProcessor = numFastaSeqs / processors;
221 for (int i = 0; i < processors; i++) {
222 int startPos = positions[ i * numSeqsPerProcessor ];
223 if(i == processors - 1){
224 numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor;
226 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
228 createProcesses(alignFileName, reportFileName, accnosFileName);
230 rename((alignFileName + toString(processIDS[0]) + ".temp").c_str(), alignFileName.c_str());
231 rename((reportFileName + toString(processIDS[0]) + ".temp").c_str(), reportFileName.c_str());
233 vector<string> nonBlankAccnosFiles;
234 //delete blank accnos files generated with multiple processes
235 for(int i=0;i<processors;i++){
236 if (!(isBlank(accnosFileName + toString(processIDS[i]) + ".temp"))) {
237 nonBlankAccnosFiles.push_back(accnosFileName + toString(processIDS[i]) + ".temp");
238 }else { remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str()); }
241 //append accnos files
242 if (nonBlankAccnosFiles.size() != 0) {
243 rename(nonBlankAccnosFiles[0].c_str(), accnosFileName.c_str());
245 for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
246 appendAlignFiles(nonBlankAccnosFiles[h], accnosFileName);
247 remove(nonBlankAccnosFiles[h].c_str());
249 mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + ".");
251 mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well.");
252 }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); }
257 for(int i=1;i<processors;i++){
258 appendAlignFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName);
259 remove((alignFileName + toString(processIDS[i]) + ".temp").c_str());
261 appendReportFiles((reportFileName + toString(processIDS[i]) + ".temp"), reportFileName);
262 remove((reportFileName + toString(processIDS[i]) + ".temp").c_str());
268 openInputFile(candidateFileName, inFASTA);
269 numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
272 lines.push_back(new linePair(0, numFastaSeqs));
274 driver(lines[0], alignFileName, reportFileName, accnosFileName);
276 //delete accnos file if its blank else report to user
277 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); }
279 mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + ".");
281 mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well.");
282 }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); }
290 mothurOut("It took " + toString(time(NULL) - start) + " secs to align " + toString(numFastaSeqs) + " sequences.");
296 catch(exception& e) {
297 errorOut(e, "AlignCommand", "execute");
302 //**********************************************************************************************************************
304 int AlignCommand::driver(linePair* line, string alignFName, string reportFName, string accnosFName){
306 ofstream alignmentFile;
307 openOutputFile(alignFName, alignmentFile);
310 openOutputFile(accnosFName, accnosFile);
312 NastReport report(reportFName);
315 openInputFile(candidateFileName, inFASTA);
317 inFASTA.seekg(line->start);
319 for(int i=0;i<line->numSeqs;i++){
321 Sequence* candidateSeq = new Sequence(inFASTA);
322 int origNumBases = candidateSeq->getNumBases();
323 string originalUnaligned = candidateSeq->getUnaligned();
324 int numBasesNeeded = origNumBases * threshold;
326 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
327 if (candidateSeq->getUnaligned().length() > alignment->getnRows()) {
328 alignment->resize(candidateSeq->getUnaligned().length()+1);
331 Sequence temp = templateDB->findClosestSequence(candidateSeq);
332 Sequence* templateSeq = &temp;
334 float searchScore = templateDB->getSearchScore();
336 Nast* nast = new Nast(alignment, candidateSeq, templateSeq);
340 bool needToDeleteCopy = false; //this is needed in case you have you enter the ifs below
341 //since nast does not make a copy of hte sequence passed, and it is used by the reporter below
342 //you can't delete the copy sequence til after you report, but you may choose not to create it in the first place
343 //so this bool tells you if you need to delete it
345 //if there is a possibility that this sequence should be reversed
346 if (candidateSeq->getNumBases() < numBasesNeeded) {
348 string wasBetter = "";
349 //if the user wants you to try the reverse
351 //get reverse compliment
352 copy = new Sequence(candidateSeq->getName(), originalUnaligned);
353 copy->reverseComplement();
356 Sequence temp2 = templateDB->findClosestSequence(copy);
357 Sequence* templateSeq2 = &temp2;
359 searchScore = templateDB->getSearchScore();
361 nast2 = new Nast(alignment, copy, templateSeq2);
363 //check if any better
364 if (copy->getNumBases() > candidateSeq->getNumBases()) {
365 candidateSeq->setAligned(copy->getAligned()); //use reverse compliments alignment since its better
366 templateSeq = templateSeq2;
369 needToDeleteCopy = true;
371 wasBetter = "\treverse complement did NOT produce a better alignment, please check sequence.";
377 //create accnos file with names
378 accnosFile << candidateSeq->getName() << wasBetter << endl;
381 report.setCandidate(candidateSeq);
382 report.setTemplate(templateSeq);
383 report.setSearchParameters(search, searchScore);
384 report.setAlignmentParameters(align, alignment);
385 report.setNastParameters(*nast);
387 alignmentFile << '>' << candidateSeq->getName() << '\n' << candidateSeq->getAligned() << endl;
391 if (needToDeleteCopy) { delete copy; }
396 alignmentFile.close();
402 catch(exception& e) {
403 errorOut(e, "AlignCommand", "driver");
408 /**************************************************************************************************/
410 void AlignCommand::createProcesses(string alignFileName, string reportFileName, string accnosFName) {
412 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
414 // processIDS.resize(0);
416 //loop through and create all the processes you want
417 while (process != processors) {
421 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
424 driver(lines[process], alignFileName + toString(getpid()) + ".temp", reportFileName + toString(getpid()) + ".temp", accnosFName + toString(getpid()) + ".temp");
426 }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); }
429 //force parent to wait until all the processes are done
430 for (int i=0;i<processors;i++) {
431 int temp = processIDS[i];
436 catch(exception& e) {
437 errorOut(e, "AlignCommand", "createProcesses");
442 /**************************************************************************************************/
444 void AlignCommand::appendAlignFiles(string temp, string filename) {
449 openOutputFileAppend(filename, output);
450 openInputFile(temp, input);
452 while(char c = input.get()){
453 if(input.eof()) { break; }
454 else { output << c; }
460 catch(exception& e) {
461 errorOut(e, "AlignCommand", "appendAlignFiles");
465 //**********************************************************************************************************************
467 void AlignCommand::appendReportFiles(string temp, string filename) {
472 openOutputFileAppend(filename, output);
473 openInputFile(temp, input);
475 while (!input.eof()) { char c = input.get(); if (c == 10 || c == 13){ break; } } // get header line
477 while(char c = input.get()){
478 if(input.eof()) { break; }
479 else { output << c; }
485 catch(exception& e) {
486 errorOut(e, "AlignCommand", "appendReportFiles");
491 //**********************************************************************************************************************