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 // ---------- flp_seqcomp.cc -----------
13 // ----------------------------------------
17 // Note one recording RC matches. This version of 'seqcomp' records RC matches
18 // as index of the window start in its nonRC form. However, past versions of
19 // the analysis recorded it as the index of the rear of the window, as the
20 // following legacy comment indicates. FR still uses this method
22 // Titus thinks the start of an RC window should be its "rear" in the
23 // normal version of the sequence. He's wrong, of course, and in the
24 // future my followers will annihilate his in the Final Battle, but for
25 // now I'll let him have his way.
27 // note seq2_i is actually the index of the leaving bp, so need to +1
30 FLPs::add(int seq1_i, int seq2_i, int a_score, int i2_offset)
35 if (a_score >= hard_threshold)
37 a_match.index = seq2_i + i2_offset;
38 a_match.score = a_score;
40 all_matches[seq1_i].push_back(a_match);
46 FLPs::seqcomp(string sseq1, string sseq2, bool is_RC)
48 int start_i, seq1_i, seq2_i, win_i; // loop variables
49 int matches; // number of matches in to a window
54 seq1 = (char *) sseq1.c_str();
55 seq2 = (char *) sseq2.c_str();
59 i2_offset = window_size - seq2_length;
64 // this does the "upper diagonals" of the search
66 // loop thru the start positions for sequence 1
67 for(start_i = 0; start_i < seq1_win_num; start_i++)
70 // compare initial window
71 for(win_i = 0; win_i < window_size; win_i++)
72 if ((seq1[start_i+win_i] == seq2[win_i]) &&
73 (seq1[start_i+win_i] != 'N')) // N's match nothing **
76 //seq2_i is always 0 for initial win
79 // call inlined function that adds match if it is above threshold
80 add(start_i, seq2_i, matches, i2_offset);
82 // run through rest of seq1 and seq2 with current seq1 offset (ie start_i)
83 // compare the bps leaving and entering the window and modify matches count
85 while( (seq1_i < seq1_win_num-1) && (seq2_i < seq2_win_num-1) )
87 // copmpare the bp leaving the window
88 if ((seq1[seq1_i] == seq2[seq2_i]) &&
89 (seq1[seq1_i] != 'N')) // N's match nothing
92 // compare the bp entering the window
93 if ((seq1[seq1_i+window_size] == seq2[seq2_i+window_size]) &&
94 (seq1[seq1_i+window_size] != 'N')) // N's match nothing
95 matches = matches + 1;
97 add(seq1_i + 1, seq2_i + 1, matches, i2_offset);
99 seq1_i = seq1_i + 1; // increment seq1_i to next window
100 seq2_i = seq2_i + 1; // increment seq2_i to next window
101 } // end of loop thru the current offset sequence
102 } // end of loop thru the start positions of seq1 sequence
105 // this does the "lower diagonals" of the search
107 // loop thru the start positions for sequence 1
108 for(start_i = 1; start_i < seq2_win_num; start_i++)
112 // compare initial window
113 for(win_i = 0; win_i < window_size; win_i++)
114 if ((seq2[start_i+win_i] == seq1[win_i]) &&
115 (seq2[start_i+win_i] != 'N')) // N's match nothing
118 //seq2_i is always start_i for initial window
121 add(0, seq2_i, matches, i2_offset);
123 // run through rest of seq1 and seq2 with current seq1 offset (ie start_i)
124 // compare the bps leaving and entering the window and modify matches count
126 while( (seq1_i < seq1_win_num-1) && (seq2_i < seq2_win_num-1) )
128 // copmpare the bp leaving the window
129 if ((seq1[seq1_i] == seq2[seq2_i]) &&
130 (seq1[seq1_i] != 'N')) // N's match nothing
131 matches = matches - 1;
133 // compare the bp entering the window
134 if ((seq1[seq1_i+window_size] == seq2[seq2_i+window_size]) &&
135 (seq1[seq1_i+window_size] != 'N')) // N's match nothing
136 matches = matches + 1;
138 add(seq1_i + 1, seq2_i + 1, matches, i2_offset);
140 seq1_i = seq1_i + 1; // increment seq1_i to next window
141 seq2_i = seq2_i + 1; // increment seq2_i to next window
142 } // end of loop thru the current offset sequence
143 } // end of loop thru the start positions of seq2 sequence