vidjil.cpp 60.1 KB
Newer Older
Mikaël Salson's avatar
Mikaël Salson committed
1
/*
2 3
  This file is part of Vidjil-algo <http://www.vidjil.org>
  Copyright (C) 2011-2018 by Bonsai bioinformatics
4 5 6 7 8
  at CRIStAL (UMR CNRS 9189, Université Lille) and Inria Lille
  Contributors: 
      Mathieu Giraud <mathieu.giraud@vidjil.org>
      Mikaël Salson <mikael.salson@vidjil.org>
      Marc Duez <marc.duez@vidjil.org>
Mikaël Salson's avatar
Mikaël Salson committed
9

10
  "Vidjil-algo" is free software: you can redistribute it and/or modify
Mikaël Salson's avatar
Mikaël Salson committed
11 12 13 14
  it under the terms of the GNU General Public License as published by
  the Free Software Foundation, either version 3 of the License, or
  (at your option) any later version.

15
  "Vidjil-algo" is distributed in the hope that it will be useful,
Mikaël Salson's avatar
Mikaël Salson committed
16 17 18 19 20
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  GNU General Public License for more details.

  You should have received a copy of the GNU General Public License
21
  along with "Vidjil-algo". If not, see <http://www.gnu.org/licenses/>
Mikaël Salson's avatar
Mikaël Salson committed
22 23
*/

Mathieu Giraud's avatar
Mathieu Giraud committed
24
//$$ #include
Mikaël Salson's avatar
Mikaël Salson committed
25 26 27 28 29 30 31 32 33 34 35 36 37

#include<algorithm>
#include<utility>
#include <cstdlib>
#include <fstream>
#include <iostream>
#include <string>
#include <cstring>
#include <time.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>

38
#include "core/check-compiler.h"
Mikaël Salson's avatar
Mikaël Salson committed
39
#include "core/tools.h"
40
#include "core/json.h"
41
#include "core/germline.h"
Mikaël Salson's avatar
Mikaël Salson committed
42 43
#include "core/kmerstore.h"
#include "core/fasta.h"
44
#include "core/bioreader.hpp"
Mikaël Salson's avatar
Mikaël Salson committed
45
#include "core/segment.h"
Mathieu Giraud's avatar
Mathieu Giraud committed
46
#include "core/windows.h"
Mikaël Salson's avatar
Mikaël Salson committed
47 48 49 50 51 52 53
#include "core/cluster-junctions.h"
#include "core/dynprog.h"
#include "core/read_score.h"
#include "core/read_chooser.h"
#include "core/compare-all.h"
#include "core/mkdir.h"
#include "core/labels.h"
Mikaël Salson's avatar
Mikaël Salson committed
54
#include "core/list_utils.h"
Mathieu Giraud's avatar
Mathieu Giraud committed
55
#include "core/windowExtractor.h"
Mikaël Salson's avatar
Mikaël Salson committed
56

57
#include "lib/CLI11.hpp"
58 59
#include "lib/json.hpp"

Mikaël Salson's avatar
Mikaël Salson committed
60 61 62 63 64 65 66
#include "vidjil.h"

// RELEASE_TAG may be defined in the "release.h" file.
// If RELEASE_TAG is undefined, the version will be the git hash.
// #define RELEASE_TAG  "2013.04"
#include "release.h"

67 68 69
// GIT_VERSION should be defined in "git-version.h", created by "create-git-version-h.sh", to be used outside of releases
#include "git-version.h"

70
#define PROGNAME "vidjil-algo"
71
#define VIDJIL_JSON_VERSION "2016b"
Mikaël Salson's avatar
Mikaël Salson committed
72

Mathieu Giraud's avatar
Mathieu Giraud committed
73 74
//$$ #define (mainly default options)

75
#define DEFAULT_MULTI_GERMLINE_PATH "germline/"
76
#define DEFAULT_MULTI_GERMLINE_FILE "homo-sapiens.g"
Mikaël Salson's avatar
Mikaël Salson committed
77

78
#define DEFAULT_READ_HEADER_SEPARATOR " "
79
#define DEFAULT_READS  "./demo/Stanford_S22.fasta"
80
#define DEFAULT_MIN_READS_CLONE 5
81
#define DEFAULT_MAX_REPRESENTATIVES 100
82
#define DEFAULT_MAX_CLONES 100
83
#define DEFAULT_RATIO_READS_CLONE 0.0
84
#define NO_LIMIT "all"
Mikaël Salson's avatar
Mikaël Salson committed
85

Mathieu Giraud's avatar
Mathieu Giraud committed
86
#define COMMAND_WINDOWS "windows"
87
#define COMMAND_CLONES "clones"
Mikaël Salson's avatar
Mikaël Salson committed
88
#define COMMAND_SEGMENT "segment"
89
#define COMMAND_GERMLINES "germlines"
Mikaël Salson's avatar
Mikaël Salson committed
90
 
91
enum { CMD_WINDOWS, CMD_CLONES, CMD_SEGMENT, CMD_GERMLINES } ;
Mikaël Salson's avatar
Mikaël Salson committed
92

93 94 95
#define DEFAULT_OUT_DIR "./out/" 

// Fixed filenames/suffixes
96
#define CLONES_FILENAME ".vdj.fa"
Mikaël Salson's avatar
Mikaël Salson committed
97
#define CLONE_FILENAME "clone.fa-"
98 99
#define WINDOWS_FILENAME ".windows.fa"
#define SEGMENTED_FILENAME ".segmented.vdj.fa"
100
#define UNSEGMENTED_FILENAME ".unsegmented.vdj.fa"
101
#define UNSEGMENTED_DETAIL_FILENAME ".fa"
102
#define AFFECTS_FILENAME ".affects"
103 104 105
#define EDGES_FILENAME ".edges"
#define COMP_FILENAME "comp.vidjil"
#define JSON_SUFFIX ".vidjil"
Mikaël Salson's avatar
Mikaël Salson committed
106

107
#define DEFAULT_K      0
108
#define DEFAULT_W      50
Mikaël Salson's avatar
Mikaël Salson committed
109

Mikaël Salson's avatar
Mikaël Salson committed
110
#define DEFAULT_MAX_AUDITIONED 2000
Mathieu Giraud's avatar
Mathieu Giraud committed
111 112
#define DEFAULT_RATIO_REPRESENTATIVE 0.5

113
#define DEFAULT_KMER_THRESHOLD 3
Mathieu Giraud's avatar
Mathieu Giraud committed
114

Mikaël Salson's avatar
Mikaël Salson committed
115 116 117 118 119 120
#define DEFAULT_EPSILON  0
#define DEFAULT_MINPTS   10

#define DEFAULT_CLUSTER_COST  Cluster
#define DEFAULT_SEGMENT_COST   VDJ

121
#define DEFAULT_TRIM 0
Mikaël Salson's avatar
Mikaël Salson committed
122

123
#define MAX_CLONES_FOR_SIMILARITY 20
124

125
// warn
126
#define WARN_MAX_CLONES 5000
127
#define WARN_PERCENT_SEGMENTED 40
128
#define WARN_COVERAGE 0.6
129
#define WARN_NUM_CLONES_SIMILAR 10
130

Mikaël Salson's avatar
Mikaël Salson committed
131 132 133
// display
#define WIDTH_NB_READS 7
#define WIDTH_NB_CLONES 3
134

Mikaël Salson's avatar
Mikaël Salson committed
135 136

using namespace std ;
137
using json = nlohmann::json;
Mikaël Salson's avatar
Mikaël Salson committed
138

Mathieu Giraud's avatar
Mathieu Giraud committed
139
//$$ options: usage
Mikaël Salson's avatar
Mikaël Salson committed
140

Mathieu Giraud's avatar
Mathieu Giraud committed
141
extern char *optarg;
Mikaël Salson's avatar
Mikaël Salson committed
142 143 144

extern int optind, optopt, opterr;

145
string usage_examples(char *progname)
Mikaël Salson's avatar
Mikaël Salson committed
146
{
147 148
  stringstream ss;
  ss
149
       << "Examples (see doc/algo.org)" << endl
150 151 152
       << "  " << progname << " -c clones   -g germline/homo-sapiens.g   -2 -3 -r 1  demo/Demo-X5.fa           # (basic usage, detect the locus for each read," << endl
       << "                                                                                               #  cluster reads and report clones starting from the first read (-r 1)," << endl
       << "                                                                                               #  including unexpected recombinations (-2), assign V(D)J genes and try to detect the CDR3s (-3))" << endl
153
       << "  " << progname << " -c clones   -g germline/homo-sapiens.g:IGH    -3     demo/Stanford_S22.fasta   # (restrict to complete recombinations on the IGH locus)" << endl
154 155 156
       << "  " << progname << " -c clones   -g germline/homo-sapiens.g   -2 -3 -z 20 demo/LIL-L4.fastq.gz      # (basic usage, output detailed V(D)J analysis on the first 20 clones)" << endl
       << "  " << progname << " -c windows  -g germline/homo-sapiens.g   -y 0 -uu -U demo/LIL-L4.fastq.gz      # (splits all the reads into (large) files depending on the detection of V(D)J recombinations)" << endl
       << "  " << progname << " -c segment  -g germline/homo-sapiens.g   -2 -3 -X 50 demo/Stanford_S22.fasta   # (full analysis of each read, only for debug/testing, here on 50 sampled reads)" << endl
157
       << "  " << progname << " -c germlines -g germline/homo-sapiens.g              demo/Stanford_S22.fasta   # (statistics on the k-mers)" << endl
Mikaël Salson's avatar
Mikaël Salson committed
158
    ;
159 160

  return ss.str();
Mikaël Salson's avatar
Mikaël Salson committed
161 162
}

163

164
int atoi_NO_LIMIT(const char *optarg)
165
{
166 167
  return strcmp(NO_LIMIT, optarg) ? atoi(optarg) : NO_LIMIT_VALUE ;
}
168
double atof_NO_LIMIT(const char *optarg)
169 170
{
  return strcmp(NO_LIMIT, optarg) ? atof(optarg) : NO_LIMIT_VALUE ;
171 172
}

173 174 175 176 177 178 179 180 181 182
string string_NO_LIMIT(string s)
{
  if (!strcmp(NO_LIMIT, s.c_str()))
    return NO_LIMIT_VALUE_STRING ;

  return s;
}



Mikaël Salson's avatar
Mikaël Salson committed
183 184
int main (int argc, char **argv)
{
185
  cout << "# " << PROGNAME << " -- V(D)J recombinations analysis <http://www.vidjil.org/>" << endl
186
       << "# Copyright (C) 2011-2018 by the Vidjil team" << endl
187
       << "# Bonsai bioinformatics at CRIStAL (UMR CNRS 9189, Université Lille) and Inria Lille" << endl 
Mathieu Giraud's avatar
Mathieu Giraud committed
188
       << endl
189
       << "# " << PROGNAME << " is free software, and you are welcome to redistribute it" << endl
Mathieu Giraud's avatar
Mathieu Giraud committed
190 191 192
       << "# under certain conditions -- see http://git.vidjil.org/blob/master/doc/LICENSE" << endl
       << "# No lymphocyte was harmed in the making of this software," << endl
       << "# however this software is for research use only and comes with no warranty." << endl
193
       << endl
194
       << "# Please cite http://biomedcentral.com/1471-2164/15/409 if you use " << PROGNAME << "." << endl
Mikaël Salson's avatar
Mikaël Salson committed
195 196
       << endl ;

197 198 199
  //////////////////////////////////
  // Display version information or git log

200 201
  string soft_version = PROGNAME ;
  soft_version += " " ;
202
#ifdef RELEASE_TAG
203
  cout << "# version: " PROGNAME << " " << RELEASE_TAG << endl ;
204 205 206 207 208 209 210 211 212 213
  soft_version.append(RELEASE_TAG);
#else
  cout << "# development version" << endl ;
#ifdef GIT_VERSION
  cout << "# git: " << GIT_VERSION << endl ;
  soft_version.append("dev ");
  soft_version.append(GIT_VERSION);
#endif
#endif

214
  CLI::App app{"# vidjil-algo -- V(D)J recombinations analysis", argv[0]};
215

Mathieu Giraud's avatar
Mathieu Giraud committed
216
  //$$ options: defaults
Mathieu Giraud's avatar
Mathieu Giraud committed
217
  float ratio_representative = DEFAULT_RATIO_REPRESENTATIVE;
Mikaël Salson's avatar
Mikaël Salson committed
218
  unsigned int max_auditionned = DEFAULT_MAX_AUDITIONED;
219
  // int average_deletion = 4;     // Average number of deletion in V or J
Mathieu Giraud's avatar
Mathieu Giraud committed
220

221
  //$$ options: definition with CLI11
222
  string group = "";
Mathieu Giraud's avatar
Mathieu Giraud committed
223

224
  // ----------------------------------------------------------------------------------------------------------------------
225
  string f_reads = DEFAULT_READS ;
226 227 228 229 230 231 232
  app.add_option("reads_file", f_reads, R"Z(reads file, in one of the following formats:
                                  - FASTA (.fa/.fasta, .fa.gz/.fasta.gz)
                                  - FASTQ (.fq/.fastq, .fq.gz/.fastq.gz)
                                  - BAM (.bam)
                              Paired-end reads should be merged before given as an input to vidjil-algo.
                 )Z")
    -> required() -> set_type_name("");
233

234

235
  // ----------------------------------------------------------------------------------------------------------------------
236
  group = "Command selection";
237

Mathieu Giraud's avatar
Mathieu Giraud committed
238 239 240 241 242 243 244
  string cmd = COMMAND_CLONES;
  app.add_option("-c", cmd, "command"
                 "\n  \t\t" COMMAND_CLONES    "  \t locus detection, window extraction, clone clustering (default command, most efficient, all outputs)"
                 "\n  \t\t" COMMAND_WINDOWS   "  \t locus detection, window extraction"
                 "\n  \t\t" COMMAND_SEGMENT   "  \t detailed V(D)J designation (not recommended)"
                 "\n  \t\t" COMMAND_GERMLINES "  \t statistics on k-mers in different germlines")
    -> group(group) -> set_type_name("COMMAND");
Mikaël Salson's avatar
Mikaël Salson committed
245

246
  // ----------------------------------------------------------------------------------------------------------------------
247
  group = "Input" ;
248 249

  string read_header_separator = DEFAULT_READ_HEADER_SEPARATOR ;
250 251
  app.add_option("--header-sep", read_header_separator, "separator for headers in the reads file", false)
    -> group(group) -> level() -> set_type_name("CHAR='" DEFAULT_READ_HEADER_SEPARATOR "'");
Mikaël Salson's avatar
Mikaël Salson committed
252

253 254

  // ----------------------------------------------------------------------------------------------------------------------
255
  group = "Germline presets (at least one -g or -V/(-D)/-J option must be given)";
256

257
  vector <string> multi_germlines ;
Mathieu Giraud's avatar
Mathieu Giraud committed
258
  app.add_option("-g", multi_germlines, R"Z(
259
         -g <.g FILE>(:FILTER)
Mathieu Giraud's avatar
Mathieu Giraud committed
260 261 262
                    multiple locus/germlines, with tuned parameters.
                    Common values are '-g germline/homo-sapiens.g' or '-g germline/mus-musculus.g'
                    The list of locus/recombinations can be restricted, such as in '-g germline/homo-sapiens.g:IGH,IGK,IGL'
263 264
         -g PATH
                    multiple locus/germlines, shortcut for '-g PATH/)Z" DEFAULT_MULTI_GERMLINE_FILE R"Z(',
265
                    processes human TRA, TRB, TRG, TRD, IGH, IGK and IGL locus, possibly with some incomplete/unusal recombinations)Z")
266 267
    -> group(group) -> set_type_name("GERMLINES");

268 269 270 271
  vector <string> v_reps_V ;
  vector <string> v_reps_D ;
  vector <string> v_reps_J ;
   
272 273 274 275
  app.add_option("-V", v_reps_V,
                 "custom V germline multi-fasta file(s)")
    -> group(group) -> set_type_name("FILE");

276

277
  app.add_option("-D", v_reps_D,
278
                 "custom D germline multi-fasta file(s), segment into V(D)J components")
279
    -> group(group) -> set_type_name("FILE");
280

281 282 283
  app.add_option("-J", v_reps_J,
                 "custom V germline multi-fasta file(s)")
    -> group(group) -> set_type_name("FILE");
284

285 286

  // ----------------------------------------------------------------------------------------------------------------------
287
  group = "Locus/recombinations";
288 289 290 291

  bool several_D = false;
  bool multi_germline_unexpected_recombinations_12 = false;

292
  app.add_flag("-d", several_D, "try to detect several D (experimental)") -> group(group);
293
  app.add_flag("-2", multi_germline_unexpected_recombinations_12, "try to detect unexpected recombinations (must be used with -g)") -> group(group);
294

295

296 297
  // ----------------------------------------------------------------------------------------------------------------------
  group = "Recombination detection (\"window\" prediction, first pass)";
298 299 300
  group += "\n    (use either -s or -k option, but not both)";
  group += "\n    (using -k option is equivalent to set with -s a contiguous seed with only '#' characters)" ;
  group += "\n    (all these options, except -w, are overriden when using -g)";
301 302 303 304

  int options_s_k = 0 ;

  IndexTypes indexType = KMER_INDEX;
305
  app.add_flag_function("-q",
306
                        [&](size_t n) { UNUSED(n); indexType = AC_AUTOMATON; },
307 308
                        "use Aho-Corasick-like automaton (experimental)")
    -> group(group) -> level();
309

310 311
  string seed = DEFAULT_SEED ;
  bool seed_changed = false;
312 313 314 315 316 317 318 319 320 321 322
  app.add_option("-k",
                 [&](CLI::results_t res) {
                   int kmer_size ;
                   bool worked = CLI::detail::lexical_cast(res[0], kmer_size);
                   if (worked) {
                     seed = seed_contiguous(kmer_size);
                     seed_changed = true;
                     options_s_k++ ;
                   }
                   return worked;
                 },
323
                 "k-mer size used for the V/J affectation (default: 10, 12, 13, depends on germline)")
324
    -> group(group) -> level() -> set_type_name("INT");
325

326
  int wmer_size = DEFAULT_W ;
327
  app.add_option("-w", wmer_size,
328
                 "w-mer size used for the length of the extracted window ('" NO_LIMIT "': use all the read, no window clustering)")
329
    -> group(group) -> level() -> transform(string_NO_LIMIT);
330

331 332

  double expected_value = THRESHOLD_NB_EXPECTED;
333
  app.add_option("-e", expected_value,
334
                 "maximal e-value for determining if a V-J segmentation can be trusted", true)
335
    -> group(group) -> level() -> transform(string_NO_LIMIT);
336

337 338
  int trim_sequences = DEFAULT_TRIM;
  bool trim_sequences_changed = false;
339 340
  app.add_option("-t",
                 [&](CLI::results_t res) {
341
                   CLI::detail::lexical_cast(res[0], trim_sequences);
342 343 344 345
                   trim_sequences_changed = true;
                   return true;
                 },
                 // trim_sequences,
346
                 "trim V and J genes (resp. 5' and 3' regions) to keep at most <INT> nt  (0: no trim)")
347
    -> group(group) -> level() ->  set_type_name("INT");
Mikaël Salson's avatar
Mikaël Salson committed
348

349 350 351 352 353 354 355
  app.add_option("-s",
                 [&](CLI::results_t res) {
                   seed = res[0] ;
                   options_s_k++ ;
                   seed_changed = true;
                   return true;
                 },
356
                 "seed, possibly spaced, used for the V/J affectation (default: depends on germline), given either explicitely or by an alias"
357 358 359
                 "\n                             " + string_of_map(seedMap, " ")
                 )
    -> group(group) -> level() -> set_type_name("SEED=" DEFAULT_SEED);
360

Mikaël Salson's avatar
Mikaël Salson committed
361

362 363
  // ----------------------------------------------------------------------------------------------------------------------
  group = "Recombination detection, experimental options (do not use)";
364

365 366 367 368
  bool multi_germline_mark = false;
  bool multi_germline_one_unique_index = false;
  bool multi_germline_unexpected_recombinations_1U = false;

369 370 371 372
  app.add_flag("-I", multi_germline_mark,
               "ignore k-mers common to different germline systems (experimental, must be used with -g, do not use)")
    -> group(group) -> level();

373
  app.add_flag("-1", multi_germline_one_unique_index,
374 375 376 377 378 379 380
               "use a unique index for all germline systems (experimental, must be used with -g, do not use)")
    -> group(group) -> level();

  app.add_flag("-4", multi_germline_unexpected_recombinations_1U,
               "try to detect unexpected recombinations with translocations (experimental, must be used with -g, do not use)")
    -> group(group) -> level();

381
  bool keep_unsegmented_as_clone = false;
382 383 384
  app.add_flag("--keep", keep_unsegmented_as_clone,
               "keep unsegmented reads as clones, taking for junction the complete sequence, to be used on very small datasets (for example --keep -AX 20)")
    -> group(group) -> level();
385

386

387
  // ----------------------------------------------------------------------------------------------------------------------
388
  group = "Labeled sequences (windows related to these sequences will be kept even if -r/--ratio thresholds are not reached)";
389 390 391 392

  vector <string> windows_labels_explicit ;
  string windows_labels_file = "" ;

393 394
  app.add_option("-W", windows_labels_explicit, "label the given sequence(s)") -> group(group) -> level() -> set_type_name("SEQUENCE");
  app.add_option("-l", windows_labels_file, "label a set of sequences given in <file>") -> group(group) -> level() -> set_type_name("FILE");
395 396

  bool only_labeled_windows = false ;
397
  app.add_flag("-F", only_labeled_windows, "filter -- keep only the windows related to the labeled sequences") -> group(group) -> level();
Mikaël Salson's avatar
Mikaël Salson committed
398 399


400 401
  // ----------------------------------------------------------------------------------------------------------------------
  group = "Limits to report a clone (or a window)";
402
  int max_clones_id = NO_LIMIT_VALUE ;
403 404 405
  int min_reads_clone = DEFAULT_MIN_READS_CLONE ;
  float ratio_reads_clone = DEFAULT_RATIO_READS_CLONE;

406
  app.add_option("--max-clones", max_clones_id, "maximal number of output clones ('" NO_LIMIT "': no maximum, default)", false) -> group(group);
407 408 409 410 411 412
  app.add_option("-r", min_reads_clone, "minimal number of reads supporting a clone", true) -> group(group);
  app.add_option("--ratio", ratio_reads_clone, "minimal percentage of reads supporting a clone", true) -> group(group);


  // ----------------------------------------------------------------------------------------------------------------------
  group = "Limits to further analyze some clones (second pass)";
413

414 415 416
  int max_clones = DEFAULT_MAX_CLONES ;
  int max_representatives = DEFAULT_MAX_REPRESENTATIVES ;

417
  app.add_option("-y", max_representatives,
418 419
                 "maximal number of clones computed with a consensus sequence ('" NO_LIMIT "': no limit)", true)
    -> group(group) -> transform(string_NO_LIMIT);
420 421 422 423 424 425 426 427 428

  app.add_option("-z",
                 [&max_clones, &max_representatives](CLI::results_t res) {
                   max_clones = atoi_NO_LIMIT(res[0].c_str());
                   if ((max_representatives < max_clones) && (max_representatives != NO_LIMIT_VALUE))
                     max_representatives = max_clones ;
                   return true;
                   // TODO: return false on bad input
                 },
429 430
                 "maximal number of clones to be analyzed with a full V(D)J designation ('" NO_LIMIT "': no limit, do not use)")
    -> group(group) -> set_type_name("INT=" + string_of_int(max_clones));
431 432

  app.add_flag_function("-A", [&](size_t n) {
433
      UNUSED(n);
434 435 436 437 438
      ratio_reads_clone = 0 ;
      min_reads_clone = 1 ;
      max_representatives = NO_LIMIT_VALUE ;
      max_clones = NO_LIMIT_VALUE ;
    },
439 440
    "reports and segments all clones (-r 0 --ratio 0 -y " NO_LIMIT " -z " NO_LIMIT "), to be used only on very small datasets (for example -AX 20)")
    -> group(group);
441

442 443 444
  int max_reads_processed = NO_LIMIT_VALUE;
  int max_reads_processed_sample = NO_LIMIT_VALUE;

445
  app.add_option("-x", max_reads_processed,
446 447 448
                 "maximal number of reads to process ('" NO_LIMIT "': no limit, default), only first reads")
    -> group(group) -> transform(string_NO_LIMIT);

449
  app.add_option("-X", max_reads_processed_sample,
450 451
                 "maximal number of reads to process ('" NO_LIMIT "': no limit, default), sampled reads")
    -> group(group) -> transform(string_NO_LIMIT);
452 453 454 455


  // ----------------------------------------------------------------------------------------------------------------------
  group = "Clone analysis (second pass)";
456

457
  Cost segment_cost = DEFAULT_SEGMENT_COST ;
458 459 460 461 462
  app.add_option("-f",
                 [&segment_cost](CLI::results_t res) {
                   segment_cost = strToCost(res[0].c_str(), VDJ); 
                   return true;
                 },
463 464
                 "use custom Cost for fine segmenter : format \"match, subst, indels, del_end, homo\" (default " + string_of_cost(DEFAULT_SEGMENT_COST) + ")")
    -> group(group) -> level() -> set_type_name("COST");
Mathieu Giraud's avatar
Mathieu Giraud committed
465

466
  double expected_value_D = THRESHOLD_NB_EXPECTED_D;
467 468 469
  app.add_option("-E", expected_value_D,
                 "maximal e-value for determining if a D segment can be trusted", true)
    -> group(group) -> level();
470

471
  int kmer_threshold = DEFAULT_KMER_THRESHOLD;
472
  app.add_option("-Z", kmer_threshold,
473
                 "typical number of V genes, selected by k-mer comparison, to compare to the read ('" NO_LIMIT "': all genes)", true)
474
    -> group(group) -> transform(string_NO_LIMIT) -> level();
475

476
  bool detect_CDR3 = false;
477 478
  app.add_flag("-3,--cdr3", detect_CDR3, "CDR3/JUNCTION detection (requires gapped V/J germlines)")
    -> group(group);
479

480

481 482
  // ----------------------------------------------------------------------------------------------------------------------
  group = "Additional clustering (third pass, experimental)" ;
483 484 485

  int epsilon = DEFAULT_EPSILON ;
  int minPts = DEFAULT_MINPTS ;
486
  app.add_option("-n", epsilon, "minimum required neighbors for automatic clustering. No automatic clusterisation if =0.", true) -> group(group) -> level();
487
  app.add_option("-N", minPts, "minimum required neighbors for automatic clustering", true) -> group(group) -> level();
488 489 490

  bool save_comp = false;
  bool load_comp = false;
491 492
  app.add_flag("-S", save_comp, "generate and save comparative matrix for clustering") -> group(group) -> level();
  app.add_flag("-L", load_comp, "load comparative matrix for clustering") -> group(group) -> level();
493 494

  string forced_edges = "" ;
495
  app.add_option("--forced-edges", forced_edges, "manual clustering -- a file used to force some specific edges") -> group(group) -> level() -> set_type_name("FILE");
496

497
  Cost cluster_cost = DEFAULT_CLUSTER_COST ;
498 499 500 501 502
  app.add_option("-C",
                 [&cluster_cost](CLI::results_t res) {
                   cluster_cost = strToCost(res[0].c_str(), Cluster);
                   return true;
                 },
503 504
                 "use custom Cost for automatic clustering : format \"match, subst, indels, del_end, homo\" (default " + string_of_cost(DEFAULT_CLUSTER_COST) + ")")
    -> group(group) -> level() -> set_type_name("COST");
505 506

  
507
  // ----------------------------------------------------------------------------------------------------------------------
508
  group = "Detailed output per read (generally not recommended, large files, but may be used for filtering, as in -uu -X 1000)";
509 510

  bool output_segmented = false;
511 512 513 514
  app.add_flag("-U", output_segmented,
               "output segmented reads (in " SEGMENTED_FILENAME " file)")
    -> group(group);

515 516 517 518
  bool output_unsegmented = false;
  bool output_unsegmented_detail = false;
  bool output_unsegmented_detail_full = false;

Mathieu Giraud's avatar
Mathieu Giraud committed
519
  app.add_flag_function("-u", [&](size_t n) {
520 521 522
      output_unsegmented = (n >= 3);             // -uuu
      output_unsegmented_detail_full = (n >= 2); // -uu
      output_unsegmented_detail = (n >= 1);      // -u
Mathieu Giraud's avatar
Mathieu Giraud committed
523 524 525
    }, R"Z(
        -u          output unsegmented reads, gathered by unsegmentation cause, except for very short and 'too few V/J' reads (in *)Z" UNSEGMENTED_DETAIL_FILENAME R"Z( files)
        -uu         output unsegmented reads, gathered by unsegmentation cause, all reads (in *)Z" UNSEGMENTED_DETAIL_FILENAME R"Z( files) (use only for debug)
526 527
        -uuu        output unsegmented reads, all reads, including a )Z" UNSEGMENTED_FILENAME R"Z( file (use only for debug))Z")
    -> group(group);
Mathieu Giraud's avatar
Mathieu Giraud committed
528

529
  bool output_affects = false;
530 531 532
  app.add_flag("-K", output_affects,
               "output detailed k-mer affectation on all reads (in " AFFECTS_FILENAME " file) (use only for debug, for example -KX 100)")
    -> group(group);
533

534

535
  // ----------------------------------------------------------------------------------------------------------------------
536
  group = "Output";
537 538 539 540

  string out_dir = DEFAULT_OUT_DIR;
  string f_basename = "";

541
  app.add_option("-o", out_dir, "output directory", true) -> group(group) -> set_type_name("PATH");
542
  app.add_option("-b", f_basename, "output basename (by default basename of the input file)") -> group(group) -> set_type_name("STRING");
543 544

  bool output_sequences_by_cluster = false;
545
  app.add_flag("-a", output_sequences_by_cluster, "output all sequences by cluster (" CLONE_FILENAME "*), to be used only on small datasets") -> group(group);
546 547

  int verbose = 0 ;
Mathieu Giraud's avatar
Mathieu Giraud committed
548
  app.add_flag_function("-v", [&](size_t n) { verbose += n ; }, "verbose mode") -> group(group);
549

550

551 552 553 554 555
  // ----------------------------------------------------------------------------------------------------------------------
  group = "Help";
  app.set_help_flag("-h", "help")
    -> group(group);

556
  app.add_flag_function("-H", [&](size_t n) { UNUSED(n); throw CLI::CallForAdvancedHelp() ; },
557 558
                        "help, including advanced and experimental options"
                        "\n                              "
559 560
                        "The full help is available in the doc/algo.org file.")
    -> group(group);
Mikaël Salson's avatar
Mikaël Salson committed
561 562


563
  // ----------------------------------------------------------------------------------------------------------------------
564 565
  app.set_footer(usage_examples(argv[0]));

566 567
  //$$ options: parsing
  CLI11_PARSE(app, argc, argv);
Mathieu Giraud's avatar
Mathieu Giraud committed
568

569
  //$$ options: post-processing+display
Mikaël Salson's avatar
Mikaël Salson committed
570

571
  int command = CMD_CLONES;
Mathieu Giraud's avatar
Mathieu Giraud committed
572 573 574 575 576 577 578 579 580 581
  if (cmd == COMMAND_CLONES)
    command = CMD_CLONES;
  else if (cmd == COMMAND_SEGMENT)
    command = CMD_SEGMENT;
  else if (cmd == COMMAND_WINDOWS)
    command = CMD_WINDOWS;
  else if (cmd == COMMAND_GERMLINES)
    command = CMD_GERMLINES;
  else {
    cerr << "Unknwown command " << optarg << endl;
582
    throw CLI::CallForHelp();
Mathieu Giraud's avatar
Mathieu Giraud committed
583
  }
584

585 586 587
  list <string> f_reps_V(v_reps_V.begin(), v_reps_V.end());
  list <string> f_reps_D(v_reps_D.begin(), v_reps_D.end());
  list <string> f_reps_J(v_reps_J.begin(), v_reps_J.end());
588

589

590 591 592
  list <pair <string, string>> multi_germline_paths_and_files ;
  bool multi_germline = false;

Mathieu Giraud's avatar
Mathieu Giraud committed
593 594 595 596 597 598 599 600 601 602 603
  for (string arg: multi_germlines)
    {
      multi_germline = true;

      struct stat buffer;
      if (stat(arg.c_str(), &buffer) == 0)
        {
          if( buffer.st_mode & S_IFDIR )
            {
              // argument is a directory
              multi_germline_paths_and_files.push_back(make_pair(arg, DEFAULT_MULTI_GERMLINE_FILE)) ;
604 605
              continue ;
            }
Mathieu Giraud's avatar
Mathieu Giraud committed
606 607 608 609 610 611 612
        }

      // argument is not a directory (and basename can include ':' with a filter)
      multi_germline_paths_and_files.push_back(make_pair(extract_dirname(arg), extract_basename(arg, false)));
    }


613
  if (!multi_germline && (!f_reps_V.size() || !f_reps_J.size()))
614
    {
615
      cerr << ERROR_STRING << "At least one germline must be given with -g or -V/(-D)/-J." << endl ;
616
      return 1;
617 618
    }

619 620
  if (options_s_k > 1)
    {
621
      cerr << ERROR_STRING << "Use at most one -s or -k option." << endl ;
622
      return 1;
623 624
    }

625 626
  map <string, string> windows_labels ;

627 628 629
  for(string lab : windows_labels_explicit)
    windows_labels[lab] = string("-W");
  
Mikaël Salson's avatar
Mikaël Salson committed
630 631
  string out_seqdir = out_dir + "/seq/" ;

Mikaël Salson's avatar
Mikaël Salson committed
632 633 634
  if (verbose)
    cout << "# verbose " << verbose << endl ;

635
  if (f_reads == DEFAULT_READS)
Mikaël Salson's avatar
Mikaël Salson committed
636 637 638
    {
      cout << "# using default sequence file: " << f_reads << endl ;
    }
Mathieu Giraud's avatar
Mathieu Giraud committed
639

640 641 642 643 644
  //  else
  //  {
  //    cerr << ERROR_STRING << "Wrong number of arguments." << endl ;
  //    return 1;
  //  }
Mathieu Giraud's avatar
Mathieu Giraud committed
645

646
  size_t min_cover_representative = (size_t) (min_reads_clone < (int) max_auditionned ? min_reads_clone : max_auditionned) ;
647

Mathieu Giraud's avatar
Mathieu Giraud committed
648 649 650
  // Check seed buffer  
  if (seed.size() >= MAX_SEED_SIZE)
    {
651
      cerr << ERROR_STRING << "Seed size is too large (MAX_SEED_SIZE)." << endl ;
652
      return 1;
Mathieu Giraud's avatar
Mathieu Giraud committed
653
    }
654

Mathieu Giraud's avatar
Mathieu Giraud committed
655
  if ((wmer_size< 0) && (wmer_size!= NO_LIMIT_VALUE))
656
    {
657
      cerr << ERROR_STRING << "Too small -w. The window size should be positive" << endl;
658
      return 1;
659 660
    }

Mikaël Salson's avatar
Mikaël Salson committed
661 662 663 664
  // Check that out_dir is an existing directory or creates it
  const char *out_cstr = out_dir.c_str();

  if (mkpath(out_cstr, 0755) == -1) {
665
    cerr << ERROR_STRING << "Directory creation: " << out_dir << endl; perror("");
666
    return 2;
Mikaël Salson's avatar
Mikaël Salson committed
667 668
  }

Mikaël Salson's avatar
Mikaël Salson committed
669 670
  const char *outseq_cstr = out_seqdir.c_str();
  if (mkpath(outseq_cstr, 0755) == -1) {
671
    cerr << ERROR_STRING << "Directory creation: " << out_seqdir << endl; perror("");
672
    return 2;
Mikaël Salson's avatar
Mikaël Salson committed
673 674
  }

675 676 677 678 679
  // Compute basename if not given as an option
  if (f_basename == "") {
    f_basename = extract_basename(f_reads);
  }

Mikaël Salson's avatar
Mikaël Salson committed
680 681 682
  out_dir += "/" ;

  /// Load labels ;
683
  load_into_map(windows_labels, windows_labels_file, "-l");
Mikaël Salson's avatar
Mikaël Salson committed
684 685

  switch(command) {
Mathieu Giraud's avatar
Mathieu Giraud committed
686
  case CMD_WINDOWS: cout << "Extracting windows" << endl; 
Mikaël Salson's avatar
Mikaël Salson committed
687
    break;
688
  case CMD_CLONES: cout << "Analysing clones" << endl; 
Mikaël Salson's avatar
Mikaël Salson committed
689 690 691
    break;
  case CMD_SEGMENT: cout << "Segmenting V(D)J" << endl;
    break;
692 693
  case CMD_GERMLINES: cout << "Discovering germlines" << endl;
    break;
Mikaël Salson's avatar
Mikaël Salson committed
694 695
  }

Mathieu Giraud's avatar
Mathieu Giraud committed
696
  cout << "Command line: ";
Mikaël Salson's avatar
Mikaël Salson committed
697
  for (int i=0; i < argc; i++) {
Mathieu Giraud's avatar
Mathieu Giraud committed
698
    cout << argv[i] << " ";
Mikaël Salson's avatar
Mikaël Salson committed
699
  }
Mathieu Giraud's avatar
Mathieu Giraud committed
700
  cout << endl;
Mikaël Salson's avatar
Mikaël Salson committed
701 702 703 704 705 706 707 708 709 710 711 712

  //////////////////////////////////
  // Display time and date
  time_t rawtime;
  struct tm *timeinfo;
  char time_buffer[80];

  time (&rawtime );
  timeinfo = localtime (&rawtime);

  strftime (time_buffer, 80,"%F %T", timeinfo);

Mathieu Giraud's avatar
Mathieu Giraud committed
713
  cout << "# " << time_buffer << endl ;
Mikaël Salson's avatar
Mikaël Salson committed
714 715


716 717
  //////////////////////////////////
  // Warning for non-optimal use
718

719
  if (max_clones == NO_LIMIT_VALUE || max_clones > WARN_MAX_CLONES)
720 721
    {
      cout << endl
722
	   << "* WARNING: " << PROGNAME << " was run with '-A' option or with a large '-z' option" << endl ;
723 724 725 726 727
    }
  
  if (command == CMD_SEGMENT)
    {
      cout << endl
728
	   << "* WARNING: " << PROGNAME << " was run with '-c segment' option" << endl ;
729 730
    }
  
731
  if (max_clones == NO_LIMIT_VALUE || max_clones > WARN_MAX_CLONES || command == CMD_SEGMENT)
732
    {
733
      cout << "* " << PROGNAME << " efficientl extracts windows overlapping the CDR3" << endl
734
           << "* to cluster reads into clones ('-c clones')." << endl
735 736
           << "* Computing accurate V(D)J designations for many sequences ('-c segment' or large '-z' values)" << endl
           << "* is slow and should be done only on small datasets or for testing purposes." << endl
737
	   << "* More information is provided in the 'doc/vidjil-algo.md' file." << endl 
738 739
	   << endl ;
    }
740

741

742 743 744 745 746 747 748
  //

  //json which contains the Levenshtein distances
  json jsonLevenshtein;
  bool jsonLevenshteinComputed = false ;


749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
  /////////////////////////////////////////
  //            JSON OUTPUT              //
  /////////////////////////////////////////

  string f_json = out_dir + f_basename + JSON_SUFFIX ;

  ostringstream stream_cmdline;
  for (int i=0; i < argc; i++) stream_cmdline << argv[i] << " ";

  json j = {
    {"vidjil_json_version", VIDJIL_JSON_VERSION},
    {"samples", {
        {"number", 1},
        {"original_names", {f_reads}},
        {"run_timestamp", {time_buffer}},
        {"producer", {soft_version}},
        {"commandline", {stream_cmdline.str()}}
      }}
  };


770 771 772
  /////////////////////////////////////////
  //            LOAD GERMLINES           //
  /////////////////////////////////////////
773

774 775 776
  if (command == CMD_GERMLINES)
    {
      multi_germline = true ;
777
      multi_germline_one_unique_index = true ;
778 779
    }

780
  MultiGermline *multigermline = new MultiGermline(indexType, !multi_germline_one_unique_index);
781

782
    {
783
      cout << "Load germlines and build Kmer indexes" << endl ;
784 785 786
    
      if (multi_germline)
	{
787
          for (pair <string, string> path_file: multi_germline_paths_and_files)
788 789
            {
              try {
790
                multigermline->build_from_json(path_file.first, path_file.second, GERMLINES_REGULAR,
791
                                               FIRST_IF_UNCHANGED("", seed, seed_changed),
792
                                               FIRST_IF_UNCHANGED(0, trim_sequences, trim_sequences_changed), (kmer_threshold != NO_LIMIT_VALUE));
793
              } catch (std::exception& e) {
794
                cerr << ERROR_STRING << PROGNAME << " cannot properly read " << path_file.first << "/" << path_file.second << ": " << e.what() << endl;
795
                delete multigermline;
796
                return 1;
797 798
              }
            }
799 800 801 802 803
	}
      else
	{
	  // Custom germline
	  Germline *germline;
804
	  germline = new Germline("custom", 'X',
805 806
                                  f_reps_V, f_reps_D, f_reps_J,
                                  seed, trim_sequences, (kmer_threshold != NO_LIMIT_VALUE));
807

808
          germline->new_index(indexType);
809

810 811 812
	  multigermline->insert(germline);
	}
    }
813

814
    cout << endl ;
815

816
    if (multi_germline_one_unique_index) {
817
      multigermline->build_with_one_index(seed, true);
818
    }
819

820
      if (multi_germline_unexpected_recombinations_12 || multi_germline_unexpected_recombinations_1U) {
821 822 823
        if (!multigermline->index) {
          multigermline->build_with_one_index(seed, false);
        }
824
      }
825

826
      if (multi_germline_unexpected_recombinations_12) {
827
        Germline *pseudo = new Germline(PSEUDO_UNEXPECTED, PSEUDO_UNEXPECTED_CODE, "", trim_sequences, (kmer_threshold != NO_LIMIT_VALUE));
828
        pseudo->seg_method = SEG_METHOD_MAX12 ;
829
        pseudo->set_index(multigermline->index);
830
        multigermline->germlines.push_back(pseudo);
831 832 833
      }

      if (multi_germline_unexpected_recombinations_1U) {
834
        Germline *pseudo_u = new Germline(PSEUDO_UNEXPECTED, PSEUDO_UNEXPECTED_CODE, "", trim_sequences, (kmer_threshold != NO_LIMIT_VALUE));
835
        pseudo_u->seg_method = SEG_METHOD_MAX1U ;
836
        // TODO: there should be more up/downstream regions for the PSEUDO_UNEXPECTED germline. And/or smaller seeds ?
837
        pseudo_u->set_index(multigermline->index);
838
        multigermline->germlines.push_back(pseudo_u);
839 840
    }

841
      // Should come after the initialization of regular (and possibly pseudo) germlines
842
    {
843
      for (pair <string, string> path_file: multi_germline_paths_and_files)
844
        multigermline->build_from_json(path_file.first, path_file.second, GERMLINES_INCOMPLETE,
845
                                       FIRST_IF_UNCHANGED("", seed, seed_changed),
846
                                       FIRST_IF_UNCHANGED(0, trim_sequences, trim_sequences_changed), (kmer_threshold != NO_LIMIT_VALUE));
847
      if ((! multigermline->one_index_per_germline) && (command != CMD_GERMLINES)) {
848 849
        multigermline->insert_in_one_index(multigermline->index, true);
      }
850 851
    }

852 853
    if (multi_germline_mark)
      multigermline->mark_cross_germlines_as_ambiguous();
854 855

    multigermline->finish();
856
    cout << "Germlines loaded: " ;
857 858
    cout << *multigermline ;
    cout << endl ;
859 860

    // Number of reads for e-value computation
861
    unsigned long long nb_reads_for_evalue = (expected_value > NO_LIMIT_VALUE) ? nb_sequences_in_file(f_reads, true) : 1 ;
862

863
    
864 865
  //////////////////////////////////
  //$$ Read sequence files
866 867 868 869

    int only_nth_read = 1 ;
    if (max_reads_processed_sample != NO_LIMIT_VALUE)
      {
870
        only_nth_read = nb_sequences_in_file(f_reads) / max_reads_processed_sample;
871 872 873
        if (only_nth_read == 0)
          only_nth_read = 1 ;

874
        max_reads_processed = max_reads_processed_sample ;
875 876 877 878 879

        if (only_nth_read > 1)
          cout << "Processing every " << only_nth_read
               << (only_nth_read == 2 ? "nd" : (only_nth_read == 3 ? "rd" : "th"))
               << " read" << endl ;
880 881
      }

882
  OnlineBioReader *reads;
883 884

  try {
885
    reads = OnlineBioReaderFactory::create(f_reads, 1, read_header_separator, max_reads_processed, only_nth_read);
886
  } catch (const invalid_argument &e) {
887
    cerr << ERROR_STRING << PROGNAME << " cannot open reads file " << f_reads << ": " << e.what() << endl;
888
    return 1;
889 890 891 892 893
  }

  out_dir += "/";


894 895 896 897 898
  //////////////////////////////://////////
  //         DISCOVER GERMLINES          //
  /////////////////////////////////////////
  if (command == CMD_GERMLINES)
    {
899
      map <char, int> stats_kmer, stats_max;
900
      IKmerStore<KmerAffect> *index = multigermline->index ;
901 902

      // Initialize statistics, with two additional categories
903 904
      index->labels.push_back(make_pair(KmerAffect::getAmbiguous(), BIOREADER_AMBIGUOUS));
      index->labels.push_back(make_pair(KmerAffect::getUnknown(), BIOREADER_UNKNOWN));
905
      
906
      for (list< pair <KmerAffect, BioReader> >::const_iterator it = index->labels.begin(); it != index->labels.end(); ++it)
907
	{
908 909 910
	  char key = affect_char(it->first.affect) ;
	  stats_kmer[key] = 0 ;
	  stats_max[key] = 0 ;
911 912
	}
      
913
      // init forbidden for .max()
914 915 916
      set<KmerAffect> forbidden;
      forbidden.insert(KmerAffect::getAmbiguous());
      forbidden.insert(KmerAffect::getUnknown());
917
      
918 919 920
      // Loop through all reads

      int nb_reads = 0 ;
921 922 923
      int total_length = 0 ;
      int s = index->getS();

924 925
      int kmer_size = seed_weight(seed);

926 927 928 929 930
      while (reads->hasNext())
	{
	  reads->next();
	  nb_reads++;
	  string seq = reads->getSequence().sequence;
931
	  total_length += seq.length() - s + 1;
932

933
	  KmerAffectAnalyser *kaa = new KmerAffectAnalyser(*index, seq);
934 935 936

	  for (int i = 0; i < kaa->count(); i++) 
	    { 
937 938
	      KmerAffect ksa = kaa->getAffectation(i);
	      stats_kmer[affect_char(ksa.affect)]++ ;
939
	    }
940

941
          delete kaa;
942

943
	  CountKmerAffectAnalyser ckaa(*index, seq);
944
	  ckaa.setAllowedOverlap(kmer_size-1);
945

946
	  stats_max[affect_char(ckaa.max(forbidden).affect)]++ ;
947

948 949
	}

950 951
      delete reads;

952 953
      // Display statistics

954 955 956 957
      cout << "  <== "
	   << nb_reads << " reads, "
	   << total_length << " kmers"
	   << endl ;
958
      cout << "\t" << " max" << "\t\t" << "        kmers" << "\n" ;
959

960
      for (list< pair <KmerAffect, BioReader> >::const_iterator it = index->labels.begin(); it != index->labels.end(); ++it)
961
	{
962 963 964
          if (it->first.getStrand() == -1)
            continue ;

965 966 967 968
	  char key = affect_char(it->first.affect) ;
	  
	  cout << setw(12) << stats_max[key] << " " ;
	  cout << setw(6) << fixed << setprecision(2) <<  (float) stats_max[key] / nb_reads * 100 << "%" ;
969

970
	  cout << "     " ;
971

972 973
	  cout << setw(12) << stats_kmer[key] << " " ;
	  cout << setw(6) << fixed << setprecision(2) <<  (float) stats_kmer[key] / total_length * 100 << "%" ;
974

975
	  cout << "     " << key << " " << it->second.name << endl ;
976
	}
977
      
978
      delete multigermline;
979 980

      return 0;
981 982 983
    }


Mikaël Salson's avatar
Mikaël Salson committed
984 985 986
  ////////////////////////////////////////
  //           CLONE ANALYSIS           //
  ////////////////////////////////////////
987
  if (command == CMD_CLONES || command == CMD_WINDOWS) {
Mikaël Salson's avatar
Mikaël Salson committed
988 989

    //////////////////////////////////
Mathieu Giraud's avatar
Mathieu Giraud committed
990
    //$$ Kmer Segmentation
Mikaël Salson's avatar
Mikaël Salson committed
991

Mathieu Giraud's avatar
Mathieu Giraud committed
992
    cout << endl;
993 994 995 996 997 998
    cout << "Loop through reads, ";

    if (wmer_size != NO_LIMIT_VALUE)
      cout << "looking for windows up to " << wmer_size << "bp" << endl;
    else
      cout << "considering all analyzed reads as windows" << endl;
Mikaël Salson's avatar
Mikaël Salson committed
999

1000
    ofstream *out_segmented = NULL;
Mathieu Giraud's avatar
Mathieu Giraud committed
1001
    ofstream *out_unsegmented = NULL;
1002
    ofstream *out_unsegmented_detail[STATS_SIZE];
1003
    ofstream *out_affects = NULL;
Mikaël Salson's avatar
Mikaël Salson committed
1004
 
1005
    WindowExtractor we(multigermline);
1006 1007
    if (! output_sequences_by_cluster)
      we.setMaximalNbReadsPerWindow(max_auditionned);
1008 1009
 
    if (output_segmented) {
1010
      string f_segmented = out_dir + f_basename + SEGMENTED_FILENAME ;
1011 1012 1013 1014 1015
      cout << "  ==> " << f_segmented << endl ;
      out_segmented = new ofstream(f_segmented.c_str());
      we.setSegmentedOutput(out_segmented);
    }

Mathieu Giraud's avatar
Mathieu Giraud committed
1016
    if (output_unsegmented) {
1017
      string f_unsegmented = out_dir + f_basename + UNSEGMENTED_FILENAME ;
Mathieu Giraud's avatar
Mathieu Giraud committed
1018 1019 1020 1021
      cout << "  ==> " << f_unsegmented << endl ;
      out_unsegmented = new ofstream(f_unsegmented.c_str());
      we.setUnsegmentedOutput(out_unsegmented);
    }
1022

1023
    if (output_unsegmented_detail) {
1024
      for (int i=STATS_FIRST_UNSEG; i<STATS_SIZE; i++)
1025 1026 1027 1028 1029 1030 1031
        {
          // Sanitize segmented_mesg[i]
          string s = segmented_mesg[i] ;
          replace(s.begin(), s.end(), '?', '_');
          replace(s.begin(), s.end(), ' ', '_');
          replace(s.begin(), s.end(), '/', '_');
          replace(s.begin(), s.end(), '<', '_');
1032
          replace(s.begin(), s.end(), '\'', '_');
1033 1034 1035 1036 1037 1038

          string f_unsegmented_detail = out_dir + f_basename + "." + s + UNSEGMENTED_DETAIL_FILENAME ;
          cout << "  ==> " << f_unsegmented_detail << endl ;
          out_unsegmented_detail[i] = new ofstream(f_unsegmented_detail.c_str());
        }

1039
      we.setUnsegmentedDetailOutput(out_unsegmented_detail, output_unsegmented_detail_full);
1040 1041 1042
    }


1043 1044 1045 1046 1047 1048 1049
    if (output_affects) {
      string f_affects = out_dir + f_basename + AFFECTS_FILENAME ;
      cout << "  ==> " << f_affects << endl ;
      out_affects = new ofstream(f_affects.c_str());
      we.setAffectsOutput(out_affects);
    }

1050
    WindowsStorage *windowsStorage = we.extract(reads, wmer_size,
1051
                                                windows_labels, only_labeled_windows,
1052
                                                keep_unsegmented_as_clone,
1053
                                                expected_value, nb_reads_for_evalue);
1054
    windowsStorage->setIdToAll();
Mathieu Giraud's avatar
Mathieu Giraud committed
1055
    size_t nb_total_reads = we.getNbReads();
Mikaël Salson's avatar
Mikaël Salson committed
1056 1057


Mathieu Giraud's avatar
Mathieu Giraud committed
1058
    //$$ Display statistics on segmentation causes
Mikaël Salson's avatar
Mikaël Salson committed
1059

1060 1061
        
    ostringstream stream_segmentation_info;
Mikaël Salson's avatar
Mikaël Salson committed
1062

Mathieu Giraud's avatar
Mathieu Giraud committed
1063
    int nb_segmented_including_too_short = we.getNbSegmented(TOTAL_SEG_AND_WINDOW) 
1064
      + we.getNbSegmented(UNSEG_TOO_SHORT_FOR_WINDOW);
Mikaël Salson's avatar
Mikaël Salson committed
1065

1066
    stream_segmentation_info << "  ==> junction detected in " << nb_segmented_including_too_short << " reads"
Mikaël Salson's avatar
Mikaël Salson committed
1067
	<< " (" << setprecision(3) << 100 * (float) nb_segmented_including_too_short / nb_total_reads << "%)" 
Mathieu Giraud's avatar
Mathieu Giraud committed
1068 1069
	<< endl ;

1070
    // nb_segmented is the main denominator for the following
Mathieu Giraud's avatar
Mathieu Giraud committed
1071
    int nb_segmented = we.getNbSegmented(TOTAL_SEG_AND_WINDOW);
1072
    float ratio_segmented = 100 * (float) nb_segmented / nb_total_reads ;