2 * chimeraslayercommand.cpp
5 * Created by westcott on 3/31/10.
6 * Copyright 2010 Schloss Lab. All rights reserved.
10 #include "chimeraslayercommand.h"
11 #include "bellerophon.h"
14 #include "chimeracheckrdp.h"
15 #include "chimeraslayer.h"
18 //***************************************************************************************************************
20 ChimeraSlayerCommand::ChimeraSlayerCommand(string option) {
24 //allow user to run help
25 if(option == "help") { help(); abort = true; }
28 //valid paramters for this command
29 string Array[] = {"fasta", "processors", "window", "template","numwanted", "ksize", "match","mismatch",
30 "divergence", "minsim","mincov","minbs", "minsnp","parents", "iters","outputdir","inputdir", "search","realign" };
31 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
33 OptionParser parser(option);
34 map<string,string> parameters = parser.getParameters();
36 ValidParameters validParameter;
37 map<string,string>::iterator it;
39 //check to make sure all parameters are valid for command
40 for (it = parameters.begin(); it != parameters.end(); it++) {
41 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
44 //if the user changes the input directory command factory will send this info to us in the output parameter
45 string inputDir = validParameter.validFile(parameters, "inputdir", false);
46 if (inputDir == "not found"){ inputDir = ""; }
49 it = parameters.find("fasta");
50 //user has given a template file
51 if(it != parameters.end()){
52 path = hasPath(it->second);
53 //if the user has not given a path then, add inputdir. else leave path alone.
54 if (path == "") { parameters["fasta"] = inputDir + it->second; }
57 it = parameters.find("template");
58 //user has given a template file
59 if(it != parameters.end()){
60 path = hasPath(it->second);
61 //if the user has not given a path then, add inputdir. else leave path alone.
62 if (path == "") { parameters["template"] = inputDir + it->second; }
67 //check for required parameters
68 fastafile = validParameter.validFile(parameters, "fasta", true);
69 if (fastafile == "not open") { abort = true; }
70 else if (fastafile == "not found") { fastafile = ""; m->mothurOut("fasta is a required parameter for the chimera.slayer command."); m->mothurOutEndLine(); abort = true; }
72 //if the user changes the output directory command factory will send this info to us in the output parameter
73 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
75 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
78 templatefile = validParameter.validFile(parameters, "template", true);
79 if (templatefile == "not open") { abort = true; }
80 else if (templatefile == "not found") { templatefile = ""; m->mothurOut("template is a required parameter for the chimera.slayer command."); m->mothurOutEndLine(); abort = true; }
82 string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found") { temp = "1"; }
83 convert(temp, processors);
85 temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found") { temp = "7"; }
88 temp = validParameter.validFile(parameters, "window", false); if (temp == "not found") { temp = "50"; }
89 convert(temp, window);
91 temp = validParameter.validFile(parameters, "match", false); if (temp == "not found") { temp = "5"; }
94 temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found") { temp = "-4"; }
95 convert(temp, mismatch);
97 temp = validParameter.validFile(parameters, "divergence", false); if (temp == "not found") { temp = "1.007"; }
100 temp = validParameter.validFile(parameters, "minsim", false); if (temp == "not found") { temp = "90"; }
101 convert(temp, minSimilarity);
103 temp = validParameter.validFile(parameters, "mincov", false); if (temp == "not found") { temp = "70"; }
104 convert(temp, minCoverage);
106 temp = validParameter.validFile(parameters, "minbs", false); if (temp == "not found") { temp = "90"; }
107 convert(temp, minBS);
109 temp = validParameter.validFile(parameters, "minsnp", false); if (temp == "not found") { temp = "10"; }
110 convert(temp, minSNP);
112 temp = validParameter.validFile(parameters, "parents", false); if (temp == "not found") { temp = "3"; }
113 convert(temp, parents);
115 temp = validParameter.validFile(parameters, "realign", false); if (temp == "not found") { temp = "f"; }
116 realign = isTrue(temp);
118 search = validParameter.validFile(parameters, "search", false); if (search == "not found") { search = "distance"; }
120 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "100"; }
121 convert(temp, iters);
123 temp = validParameter.validFile(parameters, "increment", false); if (temp == "not found") { temp = "5"; }
124 convert(temp, increment);
126 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found") { temp = "15"; }
127 convert(temp, numwanted);
129 if ((search != "distance") && (search != "blast") && (search != "kmer")) { m->mothurOut(search + " is not a valid search."); m->mothurOutEndLine(); abort = true; }
132 catch(exception& e) {
133 m->errorOut(e, "ChimeraSlayerCommand", "ChimeraSlayerCommand");
137 //**********************************************************************************************************************
139 void ChimeraSlayerCommand::help(){
142 m->mothurOut("The chimera.slayer command reads a fastafile and templatefile and outputs potentially chimeric sequences.\n");
143 m->mothurOut("This command was modeled after the chimeraSlayer written by the Broad Institute.\n");
144 m->mothurOut("The chimera.slayer command parameters are fasta, template, filter, mask, processors, ksize, window, match, mismatch, divergence. minsim, mincov, minbs, minsnp, parents, search, iters, increment and numwanted.\n"); //realign,
145 m->mothurOut("The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required. \n");
146 m->mothurOut("The template parameter allows you to enter a template file containing known non-chimeric sequences, and is required. \n");
147 m->mothurOut("The processors parameter allows you to specify how many processors you would like to use. The default is 1. \n");
149 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
151 m->mothurOut("The mask parameter allows you to specify a file containing one sequence you wish to use as a mask for the your sequences. \n");
152 m->mothurOut("The window parameter allows you to specify the window size for searching for chimeras, default=50. \n");
153 m->mothurOut("The increment parameter allows you to specify how far you move each window while finding chimeric sequences, default=5.\n");
154 m->mothurOut("The numwanted parameter allows you to specify how many sequences you would each query sequence compared with, default=15.\n");
155 m->mothurOut("The ksize parameter allows you to input kmersize, default is 7, used if search is kmer. \n");
156 m->mothurOut("The match parameter allows you to reward matched bases in blast search, default is 5. \n");
157 m->mothurOut("The parents parameter allows you to select the number of potential parents to investigate from the numwanted best matches after rating them, default is 3. \n");
158 m->mothurOut("The mismatch parameter allows you to penalize mismatched bases in blast search, default is -4. \n");
159 m->mothurOut("The divergence parameter allows you to set a cutoff for chimera determination, default is 1.007. \n");
160 m->mothurOut("The iters parameter allows you to specify the number of bootstrap iters to do with the chimeraslayer method, default=100.\n");
161 m->mothurOut("The minsim parameter allows you to specify a minimum similarity with the parent fragments, default=90. \n");
162 m->mothurOut("The mincov parameter allows you to specify minimum coverage by closest matches found in template. Default is 70, meaning 70%. \n");
163 m->mothurOut("The minbs parameter allows you to specify minimum bootstrap support for calling a sequence chimeric. Default is 90, meaning 90%. \n");
164 m->mothurOut("The minsnp parameter allows you to specify percent of SNPs to sample on each side of breakpoint for computing bootstrap support (default: 10) \n");
165 m->mothurOut("The search parameter allows you to specify search method for finding the closest parent. Choices are distance, blast, and kmer, default distance. \n");
166 //m->mothurOut("The realign parameter allows you to realign the query to the potential parents. Choices are true or false, default false. Found to make results worse. \n");
167 m->mothurOut("NOT ALL PARAMETERS ARE USED BY ALL METHODS. Please look below for method specifics.\n\n");
168 m->mothurOut("The chimera.slayer command should be in the following format: \n");
169 m->mothurOut("chimera.slayer(fasta=yourFastaFile, template=yourTemplate, search=yourSearch) \n");
170 m->mothurOut("Example: chimera.slayer(fasta=AD.align, template=core_set_aligned.imputed.fasta, search=kmer) \n");
171 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
173 catch(exception& e) {
174 m->errorOut(e, "ChimeraSlayerCommand", "help");
179 //***************************************************************************************************************
181 ChimeraSlayerCommand::~ChimeraSlayerCommand(){ /* do nothing */ }
183 //***************************************************************************************************************
185 int ChimeraSlayerCommand::execute(){
188 if (abort == true) { return 0; }
190 int start = time(NULL);
192 chimera = new ChimeraSlayer(fastafile, templatefile, search, ksize, match, mismatch, window, divR, minSimilarity, minCoverage, minBS, minSNP, parents, iters, increment, numwanted, realign);
194 string outputFileName = outputDir + getRootName(getSimpleName(fastafile)) + "slayer.chimeras";
195 string accnosFileName = outputDir + getRootName(getSimpleName(fastafile)) + "slayer.accnos";
196 bool hasAccnos = true;
198 if (m->control_pressed) { delete chimera; return 0; }
200 if (chimera->getUnaligned()) {
201 m->mothurOut("Your template sequences are different lengths, please correct."); m->mothurOutEndLine();
205 templateSeqsLength = chimera->getLength();
208 int pid, end, numSeqsPerProcessor;
211 MPIWroteAccnos = false;
214 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
215 MPI_Comm_size(MPI_COMM_WORLD, &processors);
219 MPI_File outMPIAccnos;
221 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
222 int inMode=MPI_MODE_RDONLY;
224 char outFilename[outputFileName.length()];
225 strcpy(outFilename, outputFileName.c_str());
227 char outAccnosFilename[accnosFileName.length()];
228 strcpy(outAccnosFilename, accnosFileName.c_str());
230 char inFileName[fastafile.length()];
231 strcpy(inFileName, fastafile.c_str());
233 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
234 MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI);
235 MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos);
237 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); delete chimera; return 0; }
240 if (pid == 0) { //you are the root process
241 m->mothurOutEndLine();
242 m->mothurOut("Only reporting sequence supported by " + toString(minBS) + "% of bootstrapped results.");
243 m->mothurOutEndLine();
245 string outTemp = "Name\tLeftParent\tRightParent\tDivQLAQRB\tPerIDQLAQRB\tBootStrapA\tDivQLBQRA\tPerIDQLBQRA\tBootStrapB\tFlag\tLeftWindow\tRightWindow\n";
248 int length = outTemp.length();
250 strcpy(buf2, outTemp.c_str());
251 MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &status);
253 MPIPos = setFilePosFasta(fastafile, numSeqs); //fills MPIPos, returns numSeqs
255 //send file positions to all processes
256 MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs
257 MPI_Bcast(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos
259 //figure out how many sequences you have to align
260 numSeqsPerProcessor = numSeqs / processors;
261 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
262 int startIndex = pid * numSeqsPerProcessor;
265 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
267 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); remove(outputFileName.c_str()); remove(accnosFileName.c_str()); delete chimera; return 0; }
269 for (int i = 1; i < processors; i++) {
271 MPI_Recv(&tempResult, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
272 if (tempResult != 0) { MPIWroteAccnos = true; }
274 }else{ //you are a child process
275 MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs
276 MPIPos.resize(numSeqs+1);
277 MPI_Bcast(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions
279 //figure out how many sequences you have to align
280 numSeqsPerProcessor = numSeqs / processors;
281 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
282 int startIndex = pid * numSeqsPerProcessor;
285 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
287 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); delete chimera; return 0; }
289 MPI_Send(&MPIWroteAccnos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
293 MPI_File_close(&inMPI);
294 MPI_File_close(&outMPI);
295 MPI_File_close(&outMPIAccnos);
297 //delete accnos file if blank
299 if (!MPIWroteAccnos) {
301 //MPI_File_delete(outAccnosFilename, info);
303 remove(accnosFileName.c_str());
309 string tempHeader = outputDir + getRootName(getSimpleName(fastafile)) + "slayer.chimeras.tempHeader";
310 openOutputFile(tempHeader, outHeader);
312 chimera->printHeader(outHeader);
316 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
319 openInputFile(fastafile, inFASTA);
320 numSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
323 lines.push_back(new linePair(0, numSeqs));
325 driver(lines[0], outputFileName, fastafile, accnosFileName);
327 if (m->control_pressed) {
328 remove(outputFileName.c_str());
329 remove(tempHeader.c_str());
330 remove(accnosFileName.c_str());
331 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
336 //delete accnos file if its blank
337 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
340 vector<int> positions;
341 processIDS.resize(0);
344 openInputFile(fastafile, inFASTA);
347 while(!inFASTA.eof()){
348 input = getline(inFASTA);
349 if (input.length() != 0) {
350 if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
355 numSeqs = positions.size();
357 int numSeqsPerProcessor = numSeqs / processors;
359 for (int i = 0; i < processors; i++) {
360 long int startPos = positions[ i * numSeqsPerProcessor ];
361 if(i == processors - 1){
362 numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor;
364 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
368 createProcesses(outputFileName, fastafile, accnosFileName);
370 rename((outputFileName + toString(processIDS[0]) + ".temp").c_str(), outputFileName.c_str());
372 //append output files
373 for(int i=1;i<processors;i++){
374 appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
375 remove((outputFileName + toString(processIDS[i]) + ".temp").c_str());
378 vector<string> nonBlankAccnosFiles;
379 //delete blank accnos files generated with multiple processes
380 for(int i=0;i<processors;i++){
381 if (!(isBlank(accnosFileName + toString(processIDS[i]) + ".temp"))) {
382 nonBlankAccnosFiles.push_back(accnosFileName + toString(processIDS[i]) + ".temp");
383 }else { remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str()); }
386 //append accnos files
387 if (nonBlankAccnosFiles.size() != 0) {
388 rename(nonBlankAccnosFiles[0].c_str(), accnosFileName.c_str());
390 for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
391 appendFiles(nonBlankAccnosFiles[h], accnosFileName);
392 remove(nonBlankAccnosFiles[h].c_str());
394 }else{ hasAccnos = false; }
396 if (m->control_pressed) {
397 remove(outputFileName.c_str());
398 remove(accnosFileName.c_str());
399 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
408 openInputFile(fastafile, inFASTA);
409 numSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
411 lines.push_back(new linePair(0, numSeqs));
413 driver(lines[0], outputFileName, fastafile, accnosFileName);
415 if (m->control_pressed) {
416 remove(outputFileName.c_str());
417 remove(tempHeader.c_str());
418 remove(accnosFileName.c_str());
419 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
424 //delete accnos file if its blank
425 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
428 appendFiles(tempHeader, outputFileName);
430 remove(outputFileName.c_str());
431 rename(tempHeader.c_str(), outputFileName.c_str());
436 m->mothurOutEndLine();
437 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
438 m->mothurOut(outputFileName); m->mothurOutEndLine();
439 if (hasAccnos) { m->mothurOut(accnosFileName); m->mothurOutEndLine(); }
440 m->mothurOutEndLine();
442 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
444 m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
449 catch(exception& e) {
450 m->errorOut(e, "ChimeraSlayerCommand", "execute");
454 //**********************************************************************************************************************
456 int ChimeraSlayerCommand::driver(linePair* line, string outputFName, string filename, string accnos){
459 openOutputFile(outputFName, out);
462 openOutputFile(accnos, out2);
465 openInputFile(filename, inFASTA);
467 inFASTA.seekg(line->start);
469 for(int i=0;i<line->numSeqs;i++){
471 if (m->control_pressed) { return 1; }
473 Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA);
475 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
477 if (candidateSeq->getAligned().length() != templateSeqsLength) {
478 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
481 chimera->getChimeras(candidateSeq);
483 if (m->control_pressed) { delete candidateSeq; return 1; }
486 chimera->print(out, out2);
492 if((i+1) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(i+1)); m->mothurOutEndLine(); }
495 if((line->numSeqs) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(line->numSeqs)); m->mothurOutEndLine(); }
503 catch(exception& e) {
504 m->errorOut(e, "ChimeraSlayerCommand", "driver");
508 //**********************************************************************************************************************
510 int ChimeraSlayerCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& outMPI, MPI_File& outAccMPI, vector<long>& MPIPos){
515 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
517 for(int i=0;i<num;i++){
519 if (m->control_pressed) { return 1; }
522 int length = MPIPos[start+i+1] - MPIPos[start+i];
525 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
527 string tempBuf = buf4;
528 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
529 istringstream iss (tempBuf,istringstream::in);
531 Sequence* candidateSeq = new Sequence(iss); gobble(iss);
533 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
535 if (candidateSeq->getAligned().length() != templateSeqsLength) {
536 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
539 chimera->getChimeras(candidateSeq);
541 if (m->control_pressed) { delete candidateSeq; return 1; }
544 bool isChimeric = chimera->print(outMPI, outAccMPI);
545 if (isChimeric) { MPIWroteAccnos = true; }
551 if((i+1) % 100 == 0){ cout << "Processing sequence: " << (i+1) << endl; m->mothurOutJustToLog("Processing sequence: " + toString(i+1) + "\n"); }
554 if(num % 100 != 0){ cout << "Processing sequence: " << num << endl; m->mothurOutJustToLog("Processing sequence: " + toString(num) + "\n"); }
559 catch(exception& e) {
560 m->errorOut(e, "ChimeraSlayerCommand", "driverMPI");
566 /**************************************************************************************************/
568 int ChimeraSlayerCommand::createProcesses(string outputFileName, string filename, string accnos) {
570 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
572 // processIDS.resize(0);
574 //loop through and create all the processes you want
575 while (process != processors) {
579 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
582 driver(lines[process], outputFileName + toString(getpid()) + ".temp", filename, accnos + toString(getpid()) + ".temp");
584 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
587 //force parent to wait until all the processes are done
588 for (int i=0;i<processors;i++) {
589 int temp = processIDS[i];
596 catch(exception& e) {
597 m->errorOut(e, "ChimeraSlayerCommand", "createProcesses");
602 /**************************************************************************************************/