X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=blastdb.cpp;h=2eced7e14dc6fa42e0147747fc622bfa088250f7;hp=14a1fa3a67368f5207a1c0df7a08ca5bab7fade9;hb=615301e57c25e241356a9c2380648d117709458d;hpb=5f07b42548368731cedeb92cbf3e79f630b4e20f diff --git a/blastdb.cpp b/blastdb.cpp index 14a1fa3..2eced7e 100644 --- a/blastdb.cpp +++ b/blastdb.cpp @@ -14,27 +14,172 @@ /**************************************************************************************************/ -BlastDB::BlastDB(float gO, float gE, float m, float mM) : Database(), -gapOpen(gO), gapExtend(gE), match(m), misMatch(mM) { - - count = 0; - - int randNumber = rand(); - dbFileName = toString(randNumber) + ".template.unaligned.fasta"; - queryFileName = toString(randNumber) + ".candidate.unaligned.fasta"; - blastFileName = toString(randNumber) + ".blast"; +BlastDB::BlastDB(string tag, float gO, float gE, float mm, float mM, string b, int tid) : Database(), +gapOpen(gO), gapExtend(gE), match(mm), misMatch(mM) { + try { + count = 0; + path = b; + threadID = tid; + int randNumber = rand(); + //int randNumber = 12345; + string pid = ""; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + pid += getpid(); +#else + pid += toString(threadID); +#endif + + + dbFileName = tag + pid + toString(randNumber) + ".template.unaligned.fasta"; + queryFileName = tag + pid + toString(randNumber) + ".candidate.unaligned.fasta"; + blastFileName = tag + pid + toString(randNumber) + ".blast"; + + //make sure blast exists in the write place + if (path == "") { + path = m->argv; + string tempPath = path; + for (int i = 0; i < path.length(); i++) { tempPath[i] = tolower(path[i]); } + path = path.substr(0, (tempPath.find_last_of('m'))); + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + path += "blast/bin/"; +#else + path += "blast\\bin\\"; +#endif + } + + + string formatdbCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + formatdbCommand = path + "formatdb"; // format the database, -o option gives us the ability +#else + formatdbCommand = path + "formatdb.exe"; +#endif + + //test to make sure formatdb exists + ifstream in; + formatdbCommand = m->getFullPathName(formatdbCommand); + int ableToOpen = m->openInputFile(formatdbCommand, in, "no error"); in.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + formatdbCommand + " file does not exist. mothur requires formatdb.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + string blastCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + blastCommand = path + "blastall"; // format the database, -o option gives us the ability +#else + blastCommand = path + "blastall.exe"; + //wrap entire string in "" + //blastCommand = "\"" + blastCommand + "\""; +#endif + + //test to make sure formatdb exists + ifstream in2; + blastCommand = m->getFullPathName(blastCommand); + ableToOpen = m->openInputFile(blastCommand, in2, "no error"); in2.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + blastCommand + " file does not exist. mothur requires blastall.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + + string megablastCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + megablastCommand = path + "megablast"; // format the database, -o option gives us the ability +#else + megablastCommand = path + "megablast.exe"; +#endif + + //test to make sure formatdb exists + ifstream in3; + megablastCommand = m->getFullPathName(megablastCommand); + ableToOpen = m->openInputFile(megablastCommand, in3, "no error"); in3.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + megablastCommand + " file does not exist. mothur requires megablast.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + } + catch(exception& e) { + m->errorOut(e, "BlastDB", "BlastDB"); + exit(1); + } } /**************************************************************************************************/ -BlastDB::BlastDB() : Database() { +BlastDB::BlastDB(string b, int tid) : Database() { try { count = 0; - + + path = b; + threadID = tid; + + //make sure blast exists in the write place + if (path == "") { + path = m->argv; + string tempPath = path; + for (int i = 0; i < path.length(); i++) { tempPath[i] = tolower(path[i]); } + path = path.substr(0, (tempPath.find_last_of('m'))); + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + path += "blast/bin/"; +#else + path += "blast\\bin\\"; +#endif + } + int randNumber = rand(); - dbFileName = toString(randNumber) + ".template.unaligned.fasta"; - queryFileName = toString(randNumber) + ".candidate.unaligned.fasta"; - blastFileName = toString(randNumber) + ".blast"; + string pid = ""; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + pid += getpid(); +#else + pid += toString(threadID); +#endif + + dbFileName = pid + toString(randNumber) + ".template.unaligned.fasta"; + queryFileName = pid + toString(randNumber) + ".candidate.unaligned.fasta"; + blastFileName = pid + toString(randNumber) + ".blast"; + + string formatdbCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + formatdbCommand = path + "formatdb"; // format the database, -o option gives us the ability +#else + formatdbCommand = path + "formatdb.exe"; + //wrap entire string in "" + //formatdbCommand = "\"" + formatdbCommand + "\""; +#endif + + //test to make sure formatdb exists + ifstream in; + formatdbCommand = m->getFullPathName(formatdbCommand); + int ableToOpen = m->openInputFile(formatdbCommand, in, "no error"); in.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + formatdbCommand + " file does not exist. mothur requires formatdb.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + string blastCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + blastCommand = path + "blastall"; // format the database, -o option gives us the ability +#else + blastCommand = path + "blastall.exe"; + //wrap entire string in "" + //blastCommand = "\"" + blastCommand + "\""; +#endif + + //test to make sure formatdb exists + ifstream in2; + blastCommand = m->getFullPathName(blastCommand); + ableToOpen = m->openInputFile(blastCommand, in2, "no error"); in2.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + blastCommand + " file does not exist. mothur requires blastall.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + + string megablastCommand; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + megablastCommand = path + "megablast"; // format the database, -o option gives us the ability +#else + megablastCommand = path + "megablast.exe"; + //wrap entire string in "" + //megablastCommand = "\"" + megablastCommand + "\""; +#endif + + //test to make sure formatdb exists + ifstream in3; + megablastCommand = m->getFullPathName(megablastCommand); + ableToOpen = m->openInputFile(megablastCommand, in3, "no error"); in3.close(); + if(ableToOpen == 1) { m->mothurOut("[ERROR]: " + megablastCommand + " file does not exist. mothur requires megablast.exe."); m->mothurOutEndLine(); m->control_pressed = true; } + + } catch(exception& e) { m->errorOut(e, "BlastDB", "BlastDB"); @@ -46,14 +191,14 @@ BlastDB::BlastDB() : Database() { BlastDB::~BlastDB(){ try{ - remove(queryFileName.c_str()); // let's clean stuff up and remove the temp files - remove(dbFileName.c_str()); // let's clean stuff up and remove the temp files - remove((dbFileName+".nsq").c_str()); // let's clean stuff up and remove the temp files - remove((dbFileName+".nsi").c_str()); // let's clean stuff up and remove the temp files - remove((dbFileName+".nsd").c_str()); // let's clean stuff up and remove the temp files - remove((dbFileName+".nin").c_str()); // let's clean stuff up and remove the temp files - remove((dbFileName+".nhr").c_str()); // let's clean stuff up and remove the temp files - remove(blastFileName.c_str()); // let's clean stuff up and remove the temp files + m->mothurRemove(queryFileName); // let's clean stuff up and remove the temp files + m->mothurRemove(dbFileName); // let's clean stuff up and remove the temp files + m->mothurRemove((dbFileName+".nsq")); // let's clean stuff up and remove the temp files + m->mothurRemove((dbFileName+".nsi")); // let's clean stuff up and remove the temp files + m->mothurRemove((dbFileName+".nsd")); // let's clean stuff up and remove the temp files + m->mothurRemove((dbFileName+".nin")); // let's clean stuff up and remove the temp files + m->mothurRemove((dbFileName+".nhr")); // let's clean stuff up and remove the temp files + m->mothurRemove(blastFileName.c_str()); // let's clean stuff up and remove the temp files } catch(exception& e) { m->errorOut(e, "BlastDB", "~BlastDB"); @@ -67,7 +212,10 @@ vector BlastDB::findClosestSequences(Sequence* seq, int n) { vector topMatches; ofstream queryFile; - m->openOutputFile((queryFileName+seq->getName()), queryFile); + int randNumber = rand(); + string pid = scrubName(seq->getName()); + + m->openOutputFile((queryFileName+pid+toString(randNumber)), queryFile); queryFile << '>' << seq->getName() << endl; queryFile << seq->getUnaligned() << endl; queryFile.close(); @@ -77,12 +225,22 @@ vector BlastDB::findClosestSequences(Sequence* seq, int n) { // wordsize used in megablast. I'm sure we're sacrificing accuracy for speed, but anyother way would take way too // long. With this setting, it seems comparable in speed to the suffix tree approach. - string blastCommand = path + "blast/bin/blastall -p blastn -d " + dbFileName + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n);; - blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName()); + string blastCommand; + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + + blastCommand = path + "blastall -p blastn -d " + dbFileName + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n); + blastCommand += (" -i " + (queryFileName+pid+toString(randNumber)) + " -o " + blastFileName+pid+toString(randNumber)); + #else + blastCommand = "\"" + path + "blastall\" -p blastn -d " + "\"" + dbFileName + "\"" + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n); + blastCommand += (" -i " + (queryFileName+pid+toString(randNumber)) + " -o " + blastFileName+pid+toString(randNumber)); + //wrap entire string in "" + blastCommand = "\"" + blastCommand + "\""; + #endif + system(blastCommand.c_str()); ifstream m8FileHandle; - m->openInputFile(blastFileName+seq->getName(), m8FileHandle, "no error"); + m->openInputFile(blastFileName+pid+toString(randNumber), m8FileHandle, "no error"); string dummy; int templateAccession; @@ -98,8 +256,8 @@ vector BlastDB::findClosestSequences(Sequence* seq, int n) { topMatches.push_back(templateAccession); } m8FileHandle.close(); - remove((queryFileName+seq->getName()).c_str()); - remove((blastFileName+seq->getName()).c_str()); + m->mothurRemove((queryFileName+pid+toString(randNumber))); + m->mothurRemove((blastFileName+pid+toString(randNumber))); return topMatches; } @@ -111,46 +269,65 @@ vector BlastDB::findClosestSequences(Sequence* seq, int n) { } /**************************************************************************************************/ //assumes you have added all the template sequences using the addSequence function and run generateDB. -vector BlastDB::findClosestMegaBlast(Sequence* seq, int n) { +vector BlastDB::findClosestMegaBlast(Sequence* seq, int n, int minPerID) { try{ vector topMatches; + float numBases, mismatch, gap, startQuery, endQuery, startRef, endRef, score; + Scores.clear(); ofstream queryFile; - - m->openOutputFile((queryFileName+seq->getName()), queryFile); + int randNumber = rand(); + string pid = scrubName(seq->getName()); + + m->openOutputFile((queryFileName+pid+toString(randNumber)), queryFile); queryFile << '>' << seq->getName() << endl; queryFile << seq->getUnaligned() << endl; queryFile.close(); - +// cout << seq->getUnaligned() << endl; // the goal here is to quickly survey the database to find the closest match. To do this we are using the default // wordsize used in megablast. I'm sure we're sacrificing accuracy for speed, but anyother way would take way too // long. With this setting, it seems comparable in speed to the suffix tree approach. - - string blastCommand = path + "blast/bin/megablast -e 1e-10 -d " + dbFileName + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn - blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName()); +//7000004128189528left 0 100 66 0 0 1 66 61 126 1e-31 131 + string blastCommand; + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + blastCommand = path + "megablast -e 1e-10 -d " + dbFileName + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn + blastCommand += (" -i " + (queryFileName+pid+toString(randNumber)) + " -o " + blastFileName+pid+toString(randNumber)); + #else + //blastCommand = path + "blast\\bin\\megablast -e 1e-10 -d " + dbFileName + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn + //blastCommand += (" -i " + (queryFileName+toString(randNumber)) + " -o " + blastFileName+toString(randNumber)); + + blastCommand = "\"" + path + "megablast\" -e 1e-10 -d " + "\"" + dbFileName + "\"" + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn + blastCommand += (" -i " + (queryFileName+pid+toString(randNumber)) + " -o " + blastFileName+pid+toString(randNumber)); + //wrap entire string in "" + blastCommand = "\"" + blastCommand + "\""; + + #endif system(blastCommand.c_str()); ifstream m8FileHandle; - m->openInputFile(blastFileName+seq->getName(), m8FileHandle, "no error"); + m->openInputFile(blastFileName+pid+toString(randNumber), m8FileHandle, "no error"); - string dummy; + string dummy, eScore; int templateAccession; m->gobble(m8FileHandle); while(!m8FileHandle.eof()){ - m8FileHandle >> dummy >> templateAccession >> searchScore; - //cout << templateAccession << '\t' << searchScore << endl; + m8FileHandle >> dummy >> templateAccession >> searchScore >> numBases >> mismatch >> gap >> startQuery >> endQuery >> startRef >> endRef >> eScore >> score; +// cout << dummy << '\t' << templateAccession << '\t' << searchScore << '\t' << numBases << '\t' << mismatch << '\t' << gap << '\t' << startQuery << '\t' << endQuery << '\t' << startRef << '\t' << endRef << '\t' << eScore << '\t' << score << endl; //get rest of junk in line - while (!m8FileHandle.eof()) { char c = m8FileHandle.get(); if (c == 10 || c == 13){ break; } } - + //while (!m8FileHandle.eof()) { char c = m8FileHandle.get(); if (c == 10 || c == 13){ break; }else{ cout << c; } } // + //cout << endl; m->gobble(m8FileHandle); - topMatches.push_back(templateAccession); + if (searchScore >= minPerID) { + topMatches.push_back(templateAccession); + Scores.push_back(searchScore); + } //cout << templateAccession << endl; } m8FileHandle.close(); - remove((queryFileName+seq->getName()).c_str()); - remove((blastFileName+seq->getName()).c_str()); + m->mothurRemove((queryFileName+pid+toString(randNumber))); + m->mothurRemove((blastFileName+pid+toString(randNumber))); //cout << "\n" ; return topMatches; } @@ -183,11 +360,19 @@ void BlastDB::generateDB() { try { //m->mothurOut("Generating the temporary BLAST database...\t"); cout.flush(); + + string formatdbCommand; - path = m->argv; - path = path.substr(0, (path.find_last_of('m'))); - - string formatdbCommand = path + "blast/bin/formatdb -p F -o T -i " + dbFileName; // format the database, -o option gives us the ability + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + formatdbCommand = path + "formatdb -p F -o T -i " + dbFileName; // format the database, -o option gives us the ability + #else + //formatdbCommand = path + "blast\\bin\\formatdb -p F -o T -i " + dbFileName; // format the database, -o option gives us the ability + + formatdbCommand = "\"" + path + "formatdb\" -p F -o T -i " + "\"" + dbFileName + "\""; + //wrap entire string in "" + formatdbCommand = "\"" + formatdbCommand + "\""; + #endif + //cout << formatdbCommand << endl; system(formatdbCommand.c_str()); // to get the right sequence names, i think. -p F // option tells formatdb that seqs are DNA, not prot //m->mothurOut("DONE."); m->mothurOutEndLine(); m->mothurOutEndLine(); cout.flush(); @@ -198,6 +383,24 @@ void BlastDB::generateDB() { } } /**************************************************************************************************/ +string BlastDB::scrubName(string seqName) { + try { + + string cleanName = ""; + + for (int i = 0; i < seqName.length(); i++) { + if (isalnum(seqName[i])) { cleanName += seqName[i]; } + else { cleanName += "_"; } + } + + return cleanName; + } + catch(exception& e) { + m->errorOut(e, "BlastDB", "scrubName"); + exit(1); + } +} +/**************************************************************************************************/ /**************************************************************************************************/