Mise à jour terminée. Pour connaître les apports de la version 13.8.4 par rapport à notre ancienne version vous pouvez lire les "Release Notes" suivantes :
https://about.gitlab.com/releases/2021/02/11/security-release-gitlab-13-8-4-released/
https://about.gitlab.com/releases/2021/02/05/gitlab-13-8-3-released/

segment.h 5.6 KB
Newer Older
Mikaël Salson's avatar
Mikaël Salson committed
1 2 3 4 5 6 7 8
#ifndef SEGMENT_H
#define SEGMENT_H

#include <string>
#include <iostream>
#include "fasta.h"
#include "dynprog.h"
#include "tools.h"
9
#include "germline.h"
Mikaël Salson's avatar
Mikaël Salson committed
10 11
#include "kmerstore.h"
#include "kmeraffect.h"
Mathieu Giraud's avatar
Mathieu Giraud committed
12 13
#include "affectanalyser.h"
#include "json.h"
Mikaël Salson's avatar
Mikaël Salson committed
14 15

#define EXTEND_D_ZONE 5
16 17 18

#define MIN_D_LENGTH 5          /* If a D-REGION is smaller than this threshold, it is not output */

Mikaël Salson's avatar
Mikaël Salson committed
19 20 21
#define RATIO_STRAND 5          /* The ratio between the affectations in one
                                   strand and the other, to safely attribute a
                                   segment to a given strand */
Mikaël Salson's avatar
Mikaël Salson committed
22

23
#define DETECT_THRESHOLD 5      /* If the number of both V and J affectations
24 25 26 27
                                   is above this threshold, then the sequence
                                   will be labeled as 'detected', and, if it
                                   not segmented, the remaining germlines will
                                   not be tested */
28

Mathieu Giraud's avatar
Mathieu Giraud committed
29 30
#define JSON_REMEMBER_BEST  4   /* The number of V/D/J predictions to keep  */

Mikaël Salson's avatar
Mikaël Salson committed
31 32 33 34
using namespace std;

enum SEGMENTED { DONT_KNOW, SEG_PLUS, SEG_MINUS, UNSEG_TOO_SHORT, UNSEG_STRAND_NOT_CONSISTENT, 
		 UNSEG_TOO_FEW_ZERO,  UNSEG_TOO_FEW_V, UNSEG_TOO_FEW_J, 
35
		 UNSEG_BAD_DELTA_MIN, UNSEG_BAD_DELTA_MAX, UNSEG_AMBIGUOUS,
Mikaël Salson's avatar
Mikaël Salson committed
36 37 38 39
		 TOTAL_SEG_AND_WINDOW, 
		 TOTAL_SEG_BUT_TOO_SHORT_FOR_THE_WINDOW,
		 STATS_SIZE } ;
const char* const segmented_mesg[] = { "?", "SEG_+", "SEG_-", "UNSEG too short", "UNSEG strand",  
40
				       "UNSEG too few (0)", "UNSEG too few V", "UNSEG too few J",
41
				       "UNSEG < delta_min", "UNSEG > delta_max", "UNSEG ambiguous",
Mikaël Salson's avatar
Mikaël Salson committed
42
                                       "= SEG, with window",
43
                                       "= SEG, no window",
Mikaël Salson's avatar
Mikaël Salson committed
44
                                      } ;
Mikaël Salson's avatar
Mikaël Salson committed
45 46 47 48 49

class Segmenter {
protected:
  string label;
  string sequence;
Mathieu Giraud's avatar
Mathieu Giraud committed
50 51
  int Vend, Jstart;
  int Dstart, Dend;
Marc Duez's avatar
Marc Duez committed
52
  int CDR3start, CDR3end;
Mikaël Salson's avatar
Mikaël Salson committed
53
  bool reversed, segmented, dSegmented;
54
  int because;
Mikaël Salson's avatar
Mikaël Salson committed
55 56 57 58 59 60

  string removeChevauchement();
  bool finishSegmentation();
  bool finishSegmentationD();

 public:
61
  Germline *segmented_germline;
Mikaël Salson's avatar
Mikaël Salson committed
62
  string code;
Mikaël Salson's avatar
Mikaël Salson committed
63
  string code_short;
Mikaël Salson's avatar
Mikaël Salson committed
64
  string code_light;
65 66
  string info;        // .vdj.fa header, fixed fields
  string info_extra;  // .vdj.fa header, other information, at the end of the header
Mikaël Salson's avatar
Mikaël Salson committed
67 68
  int best_V, best_J ;
  int del_V, del_D_left, del_D_right, del_J ;
69
  string seg_V, seg_N, seg_J, system;
70

Mikaël Salson's avatar
Mikaël Salson committed
71 72 73 74
  int best_D;
  string seg_N1, seg_D, seg_N2;
  Cost segment_cost;

Mathieu Giraud's avatar
Mathieu Giraud committed
75
  virtual ~Segmenter();
Mikaël Salson's avatar
Mikaël Salson committed
76

Mathieu Giraud's avatar
Mathieu Giraud committed
77
  /* Queries */
Mikaël Salson's avatar
Mikaël Salson committed
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128



  Sequence getSequence() const ;

  /**
   * @param l: length around the junction
   * @return the string centered on the junction (ie. at position
   *         (getLeft() + getRight())/2).
   *         The string has length l unless the original string 
   *         is not long enough.
   *         The junction is revcomp-ed if the original string comes from reverse
   *         strand.
   */
  string getJunction(int l) const;

  /**
   * @return the left position (on forward strand) of the segmentation.
   */
  int getLeft() const;
  
  /**
   * @return the right position (on forward strand) of the segmentation
   */
  int getRight() const;
  
  /**
   * @return the left position (on forward strand) of the D segmentation.
   */
  int getLeftD() const;
  
  /**
   * @return the right position (on forward strand) of the D segmentation
   */
  int getRightD() const;

  /**
   * @return true iff the string comes from reverse strand
   */
  bool isReverse() const;

  /**
   * @return true iff the sequence has been successfully segmented
   */
  bool isSegmented() const;
  
  /**
   * @return true if a D gene was found in the N region
   */
  bool isDSegmented() const;

129 130 131 132 133 134 135 136 137
  /**
   * @return the status of the segmentation. Tells if the Sequence has been segmented
   *         of if it has not, what the reason is.
   * @assert getSegmentationStatus() == SEG_PLUS || getSegmentationStatus() == SEG_MINUS
   *         <==> isSegmented()
   */
  int getSegmentationStatus() const;

  string getInfoLine() const;
Mikaël Salson's avatar
Mikaël Salson committed
138 139 140 141 142 143 144 145 146 147 148 149

  friend ostream &operator<<(ostream &out, const Segmenter &s);
};



ostream &operator<<(ostream &out, const Segmenter &s);



class KmerSegmenter : public Segmenter
{
Mathieu Giraud's avatar
Mathieu Giraud committed
150
 private:
151
  int detected;
152
  KmerAffectAnalyser *kaa;
Mikaël Salson's avatar
Mikaël Salson committed
153 154 155 156
 protected:
  string affects;

 public:
157
  bool isDetected() const;
158
  int score;
159 160

  KmerSegmenter();
Mikaël Salson's avatar
Mikaël Salson committed
161 162 163
  /**
   * Build a segmenter based on KmerSegmentation
   * @param seq: An object read from a FASTA/FASTQ file
164
   * @param germline: the germline
Mikaël Salson's avatar
Mikaël Salson committed
165
   */
166
  KmerSegmenter(Sequence seq, Germline *germline);
Mathieu Giraud's avatar
Mathieu Giraud committed
167

168 169
  KmerSegmenter(const KmerSegmenter &seg);

Mathieu Giraud's avatar
Mathieu Giraud committed
170 171 172 173 174
  ~KmerSegmenter();

  /**
   * @return the KmerAffectAnalyser of the current sequence.
   */
175
  KmerAffectAnalyser *getKmerAffectAnalyser() const;
Mathieu Giraud's avatar
Mathieu Giraud committed
176 177

 private:
178
  void computeSegmentation(int strand, KmerAffect left, KmerAffect right);
Mikaël Salson's avatar
Mikaël Salson committed
179 180
};

181 182 183 184 185 186 187 188

class KmerMultiSegmenter
{
 public:
  /**
   * @param seq: An object read from a FASTA/FASTQ file
   * @param multigermline: the multigermline
   */
189
  KmerMultiSegmenter(Sequence seq, MultiGermline *multigermline, ostream *out_unsegmented);
190 191
  ~KmerMultiSegmenter();

192
  KmerSegmenter *the_kseg;
193 194 195
};


Mikaël Salson's avatar
Mikaël Salson committed
196 197 198
class FineSegmenter : public Segmenter
{
 public:
Mikaël Salson's avatar
Mikaël Salson committed
199 200 201 202
   vector<pair<int, int> > score_V;
   vector<pair<int, int> > score_D;
   vector<pair<int, int> > score_J;
   
Mikaël Salson's avatar
Mikaël Salson committed
203 204 205
   /**
   * Build a fineSegmenter based on KmerSegmentation
   * @param seq: An object read from a FASTA/FASTQ file
206
   * @param germline: germline used
Mikaël Salson's avatar
Mikaël Salson committed
207
   */
208
  FineSegmenter(Sequence seq, Germline *germline, Cost segment_cost);
Mikaël Salson's avatar
Mikaël Salson committed
209 210 211
  
  /**
  * extend segmentation from VJ to VDJ
212
  * @param germline: germline used
Mikaël Salson's avatar
Mikaël Salson committed
213
  */
214
  void FineSegmentD(Germline *germline);
Marc Duez's avatar
Marc Duez committed
215
  void findCDR3();
Mathieu Giraud's avatar
Mathieu Giraud committed
216

217
  JsonList toJsonList(Germline *germline);
Mikaël Salson's avatar
Mikaël Salson committed
218
  
Mikaël Salson's avatar
Mikaël Salson committed
219 220 221 222 223
};



#endif