1 // This file is part of the Mussa source distribution.
2 // http://mussa.caltech.edu/
3 // Contact author: Tristan De Buysscher, tristan@caltech.edu
5 // This program and all associated source code files are Copyright (C) 2005
6 // the California Institute of Technology, Pasadena, CA, 91125 USA. It is
7 // under the GNU Public License; please see the included LICENSE.txt
8 // file for more information, or contact Tristan directly.
11 // ----------------------------------------
12 // ---------- mussa_class.cc -----------
13 // ----------------------------------------
19 #include "mussa_exceptions.hpp"
20 #include "alg/mussa.hpp"
21 #include "alg/flp.hpp"
30 Mussa::Mussa(const Mussa& m)
31 : analysis_name(m.analysis_name),
33 threshold(m.threshold),
34 soft_thres(m.soft_thres),
36 win_append(m.win_append),
37 thres_append(m.thres_append),
38 motif_sequences(m.motif_sequences),
39 color_mapper(m.color_mapper)
43 // set all parameters to null state
48 ana_mode = TransitiveNway;
54 motif_sequences.clear();
58 // these 5 simple methods manually set the parameters for doing an analysis
59 // used so that the gui can take input from user and setup the analysis
60 // note - still need a set_append(bool, bool) method...
62 Mussa::set_name(string a_name)
64 analysis_name = a_name;
67 string Mussa::get_name()
75 if (the_seqs.size() > 0)
76 return the_seqs.size();
82 Mussa::set_window(int a_window)
87 int Mussa::get_window() const
93 Mussa::set_threshold(int a_threshold)
95 threshold = a_threshold;
96 //soft_thres = a_threshold;
99 int Mussa::get_threshold() const
105 Mussa::set_soft_thres(int sft_thres)
107 soft_thres = sft_thres;
110 int Mussa::get_soft_thres() const
116 Mussa::set_analysis_mode(enum analysis_modes new_ana_mode)
118 ana_mode = new_ana_mode;
121 enum Mussa::analysis_modes Mussa::get_analysis_mode() const
126 string Mussa::get_analysis_mode_name() const
131 return string("Transitive");
134 return string("Radial");
137 return string("Entropy");
140 return string("[deprecated] Recursive");
143 throw runtime_error("invalid analysis mode type");
148 const NwayPaths& Mussa::paths() const
153 // takes a string and sets it as the next seq
155 Mussa::add_a_seq(string a_seq)
160 the_seqs.push_back(aSeq);
163 const vector<Sequence>&
164 Mussa::sequences() const
169 void Mussa::load_sequence(string seq_file, string annot_file, int fasta_index,
170 int sub_seq_start, int sub_seq_end)
173 aseq.load_fasta(seq_file, fasta_index, sub_seq_start, sub_seq_end);
174 if (annot_file.size() > 0) {
175 aseq.load_annot(annot_file, sub_seq_start, sub_seq_end);
177 the_seqs.push_back(aseq);
181 Mussa::load_mupa_file(string para_file_path)
183 string file_path_base;
185 string file_data_line;
186 string param, value, annot_file;
187 int split_index, fasta_index;
188 int sub_seq_start, sub_seq_end;
189 bool seq_params, did_seq;
192 string::size_type new_index, dir_index;
197 para_file.open(para_file_path.c_str(), ios::in);
199 // if file was opened, read the parameter values
202 // need to find the path to the .mupa file
207 new_index = (para_file_path.substr(dir_index)).find("/");
208 if (new_index != string::npos)
209 dir_index += new_index + 1;
211 parsing_path = false;
214 file_path_base = para_file_path.substr(0,dir_index);
216 // setup loop by getting file's first line
217 getline(para_file,file_data_line);
218 split_index = file_data_line.find(" ");
219 param = file_data_line.substr(0,split_index);
220 value = file_data_line.substr(split_index+1);
222 while (!para_file.eof())
225 if (param == "ANA_NAME")
226 analysis_name = value;
227 else if (param == "APPEND_WIN")
229 else if (param == "APPEND_THRES")
231 else if (param == "SEQUENCE_NUM")
232 ; // ignore sequence_num now
233 else if (param == "WINDOW")
234 window = atoi(value.c_str());
235 else if (param == "THRESHOLD")
236 threshold = atoi(value.c_str());
237 else if (param == "SEQUENCE")
239 string seq_file = file_path_base + value;
240 //cout << "seq_file_name " << seq_files.back() << endl;
247 while ((!para_file.eof()) && seq_params)
249 getline(para_file,file_data_line);
250 split_index = file_data_line.find(" ");
251 param = file_data_line.substr(0,split_index);
252 value = file_data_line.substr(split_index+1);
254 if (param == "FASTA_INDEX")
255 fasta_index = atoi(value.c_str());
256 else if (param == "ANNOTATION")
257 annot_file = file_path_base + value;
258 else if (param == "SEQ_START")
259 sub_seq_start = atoi(value.c_str());
260 else if (param == "SEQ_END")
262 sub_seq_end = atoi(value.c_str());
264 //ignore empty lines or that start with '#'
265 else if ((param == "") || (param == "#")) {}
266 else seq_params = false;
268 load_sequence(seq_file, annot_file, fasta_index, sub_seq_start,
272 //ignore empty lines or that start with '#'
273 else if ((param == "") || (param == "#")) {}
276 clog << "Illegal/misplaced mussa parameter in file\n";
277 clog << param << "\n";
282 getline(para_file,file_data_line);
283 split_index = file_data_line.find(" ");
284 param = file_data_line.substr(0,split_index);
285 value = file_data_line.substr(split_index+1);
292 soft_thres = threshold;
293 //cout << "nway mupa: analysis_name = " << analysis_name
294 // << " window = " << window
295 // << " threshold = " << threshold << endl;
297 // no file was loaded, signal error
300 throw mussa_load_error("Config File: " + para_file_path + " not found");
306 Mussa::analyze(int w, int t, enum Mussa::analysis_modes the_ana_mode, double new_ent_thres)
308 time_t t1, t2, begin, end;
309 double seqloadtime, seqcomptime, nwaytime, savetime, totaltime;
313 ana_mode = the_ana_mode;
314 ent_thres = new_ent_thres;
325 if (the_seqs.size() < 2) {
326 throw mussa_analysis_error("you need to have at least 2 sequences to "
329 //cout << "nway ana: seq_num = " << the_seqs.size() << endl;
332 seqloadtime = difftime(t2, t1);
337 seqcomptime = difftime(t2, t1);
341 the_paths.setup(window, threshold);
344 nwaytime = difftime(t2, t1);
349 savetime = difftime(t2, t1);
352 totaltime = difftime(end, begin);
355 //cout << "seqload\tseqcomp\tnway\tsave\ttotal\n";
356 //cout << seqloadtime << "\t";
357 //cout << seqcomptime << "\t";
358 //cout << nwaytime << "\t";
359 //cout << savetime << "\t";
360 //cout << totaltime << "\n";
366 vector<int> seq_lens;
367 vector<FLPs> empty_FLP_vector;
369 string save_file_string;
371 empty_FLP_vector.clear();
372 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
374 all_comps.push_back(empty_FLP_vector);
375 for(vector<Sequence>::size_type i2 = 0; i2 < the_seqs.size(); i2++)
376 all_comps[i].push_back(dummy_comp);
378 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
379 seq_lens.push_back(the_seqs[i].size());
381 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
382 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
384 //cout << "seqcomping: " << i << " v. " << i2 << endl;
385 all_comps[i][i2].setup(window, threshold);
386 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(), the_seqs[i2].get_seq(), false);
387 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(),the_seqs[i2].rev_comp(),true);
394 vector<string> some_Seqs;
396 the_paths.set_soft_thres(soft_thres);
398 if (ana_mode == TransitiveNway) {
399 the_paths.trans_path_search(all_comps);
401 else if (ana_mode == RadialNway) {
402 the_paths.radiate_path_search(all_comps);
404 else if (ana_mode == EntropyNway)
406 //unlike other methods, entropy needs to look at the sequence at this stage
408 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
410 some_Seqs.push_back(the_seqs[i].get_seq());
413 the_paths.setup_ent(ent_thres, some_Seqs); // ent analysis extra setup
414 the_paths.entropy_path_search(all_comps);
417 // old recursive transitive analysis function
418 else if (ana_mode == RecursiveNway)
419 the_paths.find_paths_r(all_comps);
421 the_paths.simple_refine();
427 string save_name, save_path, create_dir_cmd, flp_filepath;
429 ostringstream append_info;
430 int dir_create_status;
433 // not sure why, but gotta close file each time since can't pass file streams
435 save_name = analysis_name;
437 // gotta do bit with adding win & thres if to be appended
441 append_info << "_w" << window;
442 save_name += append_info.str();
448 append_info << "_t" << threshold;
449 save_name += append_info.str();
452 //#include <stdlib.h>
453 // ******* use appropriate for os ------- 1 of 4
454 // the additions for osX make it more sane where it saves the analysis
455 // will come up with a cleaner sol'n later...
456 create_dir_cmd = "mkdir " + save_name; //linux
457 //create_dir_cmd = "mkdir " + file_path_base + save_name; //osX
459 dir_create_status = system( (const char*) create_dir_cmd.c_str());
460 //cout << "action: " << dir_create_status << endl;
462 // save sequence and annots to a special mussa file
464 // ******** use appropriate for OS ---------- 2 of 4
465 save_path = save_name + "/" + save_name + ".museq"; //linux
466 //save_path = file_path_base + save_name + "/" + save_name + ".museq"; //osX
468 save_file.open(save_path.c_str(), ios::out);
469 save_file << "<Mussa_Sequence>" << endl;
472 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
474 the_seqs[i].save(save_file);
477 //save_file.open(save_path.c_str(), ios::app);
478 save_file << "</Mussa_Sequence>" << endl;
481 // save nway paths to its mussa save file
483 // ******** use appropriate for OS -------- 3 of 4
484 save_path = save_name + "/" + save_name + ".muway"; //linux
485 //save_path = file_path_base + save_name + "/" + save_name + ".muway"; //os X
486 the_paths.save(save_path);
488 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
489 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
492 append_info << "_sp_" << i << "v" << i2;
493 // ******** use appropriate for OS --------- 4 of 4
495 save_path = save_name + "/" + save_name + append_info.str() + ".flp";
497 //save_path = file_path_base + save_name + "/" + save_name + append_info.str() + ".flp";
498 all_comps[i][i2].save(save_path);
503 Mussa::save_muway(string save_path)
505 the_paths.save(save_path);
509 Mussa::load(string ana_file)
512 string::size_type start_index, end_index;
513 string file_path_base, a_file_path, ana_path;
517 ostringstream append_info;
518 vector<FLPs> empty_FLP_vector;
521 //cout << "ana_file name " << ana_file << endl;
524 end_index = ana_path.size()-1;
525 if (ana_path[end_index] == '/') {
528 start_index = ana_path.rfind('/', end_index);
529 if (start_index == string::npos) {
533 // skip the / we found
536 analysis_name = ana_path.substr(start_index, end_index-start_index+1);
537 //cout << " ana_name " << analysis_name << endl;
538 file_path_base = ana_path.substr(0, start_index) + analysis_name
539 + "/" + analysis_name;
540 a_file_path = file_path_base + ".muway";
541 //cout << " loading museq: " << a_file_path << endl;
542 the_paths.load(a_file_path);
544 int seq_num = the_paths.sequence_count();
546 a_file_path = file_path_base + ".museq";
548 // this is a bit of a hack due to C++ not acting like it should with files
549 for (i = 1; i <= seq_num; i++)
552 //cout << "mussa_class: loading museq frag... " << a_file_path << endl;
553 tmp_seq.load_museq(a_file_path, i);
554 the_seqs.push_back(tmp_seq);
557 empty_FLP_vector.clear();
558 for(i = 0; i < seq_num; i++)
560 all_comps.push_back(empty_FLP_vector);
561 for(i2 = 0; i2 < seq_num; i2++)
562 all_comps[i].push_back(dummy_comp);
565 for(i = 0; i < seq_num; i++)
567 for(i2 = i+1; i2 < seq_num; i2++)
570 append_info << "_sp_" << i << "v" << i2;
571 //cout << append_info.str() << endl;
572 a_file_path = file_path_base + append_info.str() + ".flp";
573 all_comps[i][i2].load(a_file_path);
574 //cout << "real size = " << all_comps[i][i2].size() << endl;
585 save_file.open(analysis_name.c_str(), ios::out);
587 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
588 save_file << the_seqs[i].get_seq() << endl;
590 save_file << window << endl;
592 //note more complex eventually since analysis_name may need to have
593 //window size, threshold and other stuff to modify it...
594 the_paths.save_old(analysis_name);
599 Mussa::load_old(char * load_file_path, int s_num)
602 string file_data_line;
603 int i, space_split_i, comma_split_i;
604 vector<int> loaded_path;
605 string node_pair, node;
609 the_paths.setup(0, 0);
610 save_file.open(load_file_path, ios::in);
612 // currently loads old mussa format
615 for(i = 0; i < seq_num; i++)
617 getline(save_file, file_data_line);
618 a_seq.set_seq(file_data_line);
619 the_seqs.push_back(a_seq);
623 getline(save_file, file_data_line);
624 window = atoi(file_data_line.c_str());
627 while (!save_file.eof())
630 getline(save_file, file_data_line);
631 if (file_data_line != "")
632 for(i = 0; i < seq_num; i++)
634 space_split_i = file_data_line.find(" ");
635 node_pair = file_data_line.substr(0,space_split_i);
636 //cout << "np= " << node_pair;
637 comma_split_i = node_pair.find(",");
638 node = node_pair.substr(comma_split_i+1);
639 //cout << "n= " << node << " ";
640 loaded_path.push_back(atoi (node.c_str()));
641 file_data_line = file_data_line.substr(space_split_i+1);
644 // FIXME: do we have any information about what the threshold should be?
645 the_paths.add_path(0, loaded_path);
649 //the_paths.save("tmp.save");
652 void Mussa::add_motifs(const vector<string>& motifs,
653 const vector<Color>& colors)
655 if (motifs.size() != colors.size()) {
656 throw mussa_error("motif and color vectors must be the same size");
659 for(size_t i = 0; i != motifs.size(); ++i)
661 motif_sequences.insert(motifs[i]);
662 color_mapper.appendInstanceColor("motif", motifs[i], colors[i]);
664 update_sequences_motifs();
667 // I mostly split the ifstream out so I can use a stringstream to test it.
668 void Mussa::load_motifs(std::istream &in)
677 in >> seq >> red >> green >> blue;
678 // if we couldn't read this line 'cause we're like at the end of the file
679 // try to exit the loop
683 seq = Sequence::motif_normalize(seq);
684 } catch(motif_normalize_error e) {
685 clog << "unable to parse " << seq << " skipping" << endl;
686 clog << e.what() << endl;
689 if (red < 0.0 or red > 1.0) {
690 clog << "invalid red value " << red << ". must be in range [0..1]"
694 if (green < 0.0 or green > 1.0) {
695 clog << "invalid green value " << green << ". must be in range [0..1]"
699 if (blue < 0.0 or blue > 1.0) {
700 clog << "invalid blue value " << blue << ". must be in range [0..1]"
704 if (motif_sequences.find(seq) == motif_sequences.end()) {
705 // sequence wasn't found
706 motif_sequences.insert(seq);
707 Color c(red, green, blue);
708 color_mapper.appendInstanceColor("motif", seq, c);
710 clog << "sequence " << seq << " was already defined skipping"
715 update_sequences_motifs();
718 void Mussa::load_motifs(string filename)
721 f.open(filename.c_str(), ifstream::in);
725 void Mussa::update_sequences_motifs()
727 // once we've loaded all the motifs from the file,
728 // lets attach them to the sequences
729 for(vector<Sequence>::iterator seq_i = the_seqs.begin();
730 seq_i != the_seqs.end();
733 // clear out old motifs
734 seq_i->clear_motifs();
735 // for all the motifs in our set, attach them to the current sequence
736 for(set<string>::iterator motif_i = motif_sequences.begin();
737 motif_i != motif_sequences.end();
740 seq_i->add_motif(*motif_i);
745 const set<string>& Mussa::motifs() const
747 return motif_sequences;
750 AnnotationColors& Mussa::colorMapper()