1 | #include "mltaln.h" |
---|
2 | #include "dp.h" |
---|
3 | |
---|
4 | #define MACHIGAI 0 |
---|
5 | #define OUTGAP0TRY 1 |
---|
6 | #define DEBUG 0 |
---|
7 | #define XXXXXXX 0 |
---|
8 | #define USE_PENALTY_EX 0 |
---|
9 | #define FASTMATCHCALC 1 |
---|
10 | |
---|
11 | static TLS float **impmtx = NULL; |
---|
12 | |
---|
13 | #if 0 // by D.Mathog |
---|
14 | static float countnocountxx( Gappat *pat1, float diaf1, Gappat *pat2, int offset1, int offset2 ) |
---|
15 | { |
---|
16 | // return( 0.0 ); |
---|
17 | float gclose; |
---|
18 | float gmatch; |
---|
19 | Gappat *pat1bk = pat1; |
---|
20 | Gappat *pat2bk = pat2; |
---|
21 | |
---|
22 | gmatch = 0.0; |
---|
23 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) // excl. len=0 |
---|
24 | { |
---|
25 | if( pat2->len + offset2 == offset1 ) |
---|
26 | { |
---|
27 | gmatch = diaf1 * pat2->freq; |
---|
28 | } |
---|
29 | } |
---|
30 | for( pat1=pat1bk+1; pat1->len != 0; pat1++ ) // excl. len=0 |
---|
31 | { |
---|
32 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) // excl. len=0 |
---|
33 | { |
---|
34 | if( pat1->len + offset1 == pat2->len + offset2 ) |
---|
35 | { |
---|
36 | gmatch += pat1->freq * pat2->freq; |
---|
37 | // if( r ) fprintf( stderr, "match1!!, len=%d, gmatch=%f * %f\n", pat2->len, pat1->freq, pat2->freq ); |
---|
38 | } |
---|
39 | } |
---|
40 | } |
---|
41 | return( gmatch ); |
---|
42 | } |
---|
43 | #endif |
---|
44 | |
---|
45 | static float countnocountmatchx( Gappat *pat1, Gappat *pat2, int offset1, int offset2, int r ) |
---|
46 | { |
---|
47 | Gappat *pat1bk = pat1; |
---|
48 | Gappat *pat2bk = pat2; |
---|
49 | float val = 0.0; |
---|
50 | // pat1[][0] ha total gap. |
---|
51 | for( pat1=pat1bk+1; pat1->len != 0; pat1++ ) |
---|
52 | { |
---|
53 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) |
---|
54 | { |
---|
55 | if( pat1->len + offset1 == pat2->len + offset2 ) |
---|
56 | { |
---|
57 | val += pat1->freq * pat2->freq; |
---|
58 | if( r ) fprintf( stderr, "y %d-%d, len=%d,%d, val = %f\n", (int)(pat1-pat1bk), (int)(pat2-pat2bk), pat1->len, pat2->len, val ); // 070405 |
---|
59 | // if( r ) fprintf( stderr, "y %d-%d, len=%d,%d, val = %f\n", pat1-pat1bk, pat2-pat2bk, pat1->len, pat2->len, val ); |
---|
60 | } |
---|
61 | } |
---|
62 | } |
---|
63 | if( r ) fprintf( stderr, "nocountmatch=%f\n", val ); |
---|
64 | return( val ); |
---|
65 | } |
---|
66 | |
---|
67 | #if 0 // by D.Mathog |
---|
68 | static float countnocountmatch( Gappat *pat1, Gappat *pat2, int r ) |
---|
69 | { |
---|
70 | // return( 0.0 ); |
---|
71 | Gappat *pat1bk = pat1; |
---|
72 | Gappat *pat2bk = pat2; |
---|
73 | float val = 0.0; |
---|
74 | // pat1[][0] ha total gap. |
---|
75 | for( pat1=pat1bk+1; pat1->len != 0; pat1++ ) |
---|
76 | { |
---|
77 | // if( r ) fprintf( stderr, "b %d-%d, len=%d,%d\n", pat1-pat1bk, pat2-pat2bk, pat1->len, pat2->len ); |
---|
78 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) |
---|
79 | { |
---|
80 | if( pat1->len == pat2->len ) |
---|
81 | { |
---|
82 | // if( r ) fprintf( stderr, " x%d-%d, len=%d,%d\n", pat1-pat1bk, pat2-pat2bk, pat1->len, pat2->len ); |
---|
83 | val += pat1->freq * pat2->freq; |
---|
84 | // if( r ) fprintf( stderr, "y %d-%d, val = %f\n", pat1-pat1bk, pat2-pat2bk,val ); |
---|
85 | // if( r ) fprintf( stderr, "z tsugi, %d-%d, len=%d,%d\n", pat1-pat1bk+1, pat2-pat2bk+1, (pat1+1)->len, (pat2+1)->len ); |
---|
86 | } |
---|
87 | // if( r ) fprintf( stderr, "a %d-%d, len=%d,%d\n", pat1-pat1bk, pat2-pat2bk, pat1->len, pat2->len ); |
---|
88 | } |
---|
89 | } |
---|
90 | // fprintf( stderr, "nocountmatch=%f\n", val ); |
---|
91 | return( val ); |
---|
92 | } |
---|
93 | #endif |
---|
94 | |
---|
95 | static float countnocountx( Gappat *pat1, float diaf1, Gappat *pat2, int offset1, int r ) |
---|
96 | { |
---|
97 | // return( 0.0 ); |
---|
98 | float gmatch; |
---|
99 | Gappat *pat1bk = pat1; |
---|
100 | Gappat *pat2bk = pat2; |
---|
101 | |
---|
102 | gmatch = 0.0; |
---|
103 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) // excl. len=0 |
---|
104 | { |
---|
105 | if( pat2->len == offset1 ) |
---|
106 | { |
---|
107 | gmatch = diaf1 * pat2->freq; |
---|
108 | // if( r ) fprintf( stderr, "match0!!, len=%d, gmatch=%f * %f\n", pat2->len, diaf1, pat2->freq ); |
---|
109 | } |
---|
110 | } |
---|
111 | for( pat1=pat1bk+1; pat1->len != 0; pat1++ ) // excl. len=0 |
---|
112 | { |
---|
113 | for( pat2=pat2bk+1; pat2->len != 0; pat2++ ) // excl. len=0 |
---|
114 | { |
---|
115 | if( pat1->len + offset1 == pat2->len ) |
---|
116 | { |
---|
117 | gmatch += pat1->freq * pat2->freq; |
---|
118 | // if( r ) fprintf( stderr, "match1!!, len=%d, gmatch=%f * %f\n", pat2->len, pat1->freq, pat2->freq ); |
---|
119 | } |
---|
120 | } |
---|
121 | } |
---|
122 | return( gmatch ); |
---|
123 | } |
---|
124 | |
---|
125 | #if 0 // by D.Mathog |
---|
126 | static float countnocount( Gappat *pat1, Gappat *pat2, int offset1, int offset2 ) //osoi |
---|
127 | { |
---|
128 | // return( 0.0 ); |
---|
129 | Gappat *pat1bk = pat1; |
---|
130 | Gappat *pat2bk = pat2; |
---|
131 | float val = 0.0; |
---|
132 | // pat1[][0] ha total gap. |
---|
133 | for( pat1=pat1bk+1; pat1->len != -1; pat1++ ) |
---|
134 | { |
---|
135 | for( pat2=pat2bk+1; pat2->len != -1; pat2++ ) |
---|
136 | { |
---|
137 | if( pat1->len+offset1 == pat2->len+offset2 ) |
---|
138 | { |
---|
139 | val += pat1->freq * pat2->freq; |
---|
140 | } |
---|
141 | } |
---|
142 | } |
---|
143 | // fprintf( stderr, "nocount=%f\n", val ); |
---|
144 | return( val ); |
---|
145 | } |
---|
146 | #endif |
---|
147 | |
---|
148 | |
---|
149 | |
---|
150 | #if 1 // tditeration |
---|
151 | float imp_match_out_scH( int i1, int j1 ) |
---|
152 | { |
---|
153 | // fprintf( stderr, "imp+match = %f\n", impmtx[i1][j1] * fastathreshold ); |
---|
154 | // fprintf( stderr, "val = %f\n", impmtx[i1][j1] ); |
---|
155 | return( impmtx[i1][j1] ); |
---|
156 | } |
---|
157 | #endif |
---|
158 | |
---|
159 | static void imp_match_out_veadH( float *imp, int i1, int lgth2 ) |
---|
160 | { |
---|
161 | #if FASTMATCHCALC |
---|
162 | float *pt = impmtx[i1]; |
---|
163 | while( lgth2-- ) |
---|
164 | *imp++ += *pt++; |
---|
165 | #else |
---|
166 | int j; |
---|
167 | float *pt = impmtx[i1]; |
---|
168 | for( j=0; j<lgth2; j++ ) |
---|
169 | *imp++ += pt[j]; |
---|
170 | #endif |
---|
171 | } |
---|
172 | static void imp_match_out_vead_tateH( float *imp, int j1, int lgth1 ) |
---|
173 | { |
---|
174 | int i; |
---|
175 | for( i=0; i<lgth1; i++ ) |
---|
176 | *imp++ += impmtx[i][j1]; |
---|
177 | } |
---|
178 | |
---|
179 | #if 1 // tbfast.c kara yobareru. |
---|
180 | void imp_match_init_strictH( float *imp, int clus1, int clus2, int lgth1, int lgth2, char **seq1, char **seq2, double *eff1, double *eff2, LocalHom ***localhom, int forscore ) |
---|
181 | { |
---|
182 | int i, j, k1, k2, tmpint, start1, start2, end1, end2; |
---|
183 | static TLS int impalloclen = 0; |
---|
184 | float effij; |
---|
185 | double effijx; |
---|
186 | char *pt, *pt1, *pt2; |
---|
187 | static TLS char *nocount1 = NULL; |
---|
188 | static TLS char *nocount2 = NULL; |
---|
189 | LocalHom *tmpptr; |
---|
190 | |
---|
191 | if( impalloclen < lgth1 + 2 || impalloclen < lgth2 + 2 ) |
---|
192 | { |
---|
193 | if( impmtx ) FreeFloatMtx( impmtx ); |
---|
194 | if( nocount1 ) free( nocount1 ); |
---|
195 | if( nocount2 ) free( nocount2 ); |
---|
196 | impalloclen = MAX( lgth1, lgth2 ) + 2; |
---|
197 | impmtx = AllocateFloatMtx( impalloclen, impalloclen ); |
---|
198 | nocount1 = AllocateCharVec( impalloclen ); |
---|
199 | nocount2 = AllocateCharVec( impalloclen ); |
---|
200 | } |
---|
201 | |
---|
202 | for( i=0; i<lgth1; i++ ) |
---|
203 | { |
---|
204 | for( j=0; j<clus1; j++ ) |
---|
205 | if( seq1[j][i] == '-' ) break; |
---|
206 | if( j != clus1 ) nocount1[i] = 1; |
---|
207 | else nocount1[i] = 0; |
---|
208 | } |
---|
209 | for( i=0; i<lgth2; i++ ) |
---|
210 | { |
---|
211 | for( j=0; j<clus2; j++ ) |
---|
212 | if( seq2[j][i] == '-' ) break; |
---|
213 | if( j != clus2 ) nocount2[i] = 1; |
---|
214 | else nocount2[i] = 0; |
---|
215 | } |
---|
216 | |
---|
217 | #if 0 |
---|
218 | fprintf( stderr, "nocount2 =\n" ); |
---|
219 | for( i = 0; i<impalloclen; i++ ) |
---|
220 | { |
---|
221 | fprintf( stderr, "nocount2[%d] = %d (%c)\n", i, nocount2[i], seq2[0][i] ); |
---|
222 | } |
---|
223 | #endif |
---|
224 | |
---|
225 | |
---|
226 | |
---|
227 | #if 0 |
---|
228 | fprintf( stderr, "eff1 in _init_strict = \n" ); |
---|
229 | for( i=0; i<clus1; i++ ) |
---|
230 | fprintf( stderr, "eff1[] = %f\n", eff1[i] ); |
---|
231 | for( i=0; i<clus2; i++ ) |
---|
232 | fprintf( stderr, "eff2[] = %f\n", eff2[i] ); |
---|
233 | #endif |
---|
234 | |
---|
235 | for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ ) |
---|
236 | impmtx[i][j] = 0.0; |
---|
237 | effijx = fastathreshold; |
---|
238 | for( i=0; i<clus1; i++ ) |
---|
239 | { |
---|
240 | for( j=0; j<clus2; j++ ) |
---|
241 | { |
---|
242 | effij = (float)( eff1[i] * eff2[j] * effijx ); |
---|
243 | tmpptr = localhom[i][j]; |
---|
244 | while( tmpptr ) |
---|
245 | { |
---|
246 | // fprintf( stderr, "start1 = %d\n", tmpptr->start1 ); |
---|
247 | // fprintf( stderr, "end1 = %d\n", tmpptr->end1 ); |
---|
248 | // fprintf( stderr, "i = %d, seq1 = \n%s\n", i, seq1[i] ); |
---|
249 | // fprintf( stderr, "j = %d, seq2 = \n%s\n", j, seq2[j] ); |
---|
250 | pt = seq1[i]; |
---|
251 | tmpint = -1; |
---|
252 | while( *pt != 0 ) |
---|
253 | { |
---|
254 | if( *pt++ != '-' ) tmpint++; |
---|
255 | if( tmpint == tmpptr->start1 ) break; |
---|
256 | } |
---|
257 | start1 = pt - seq1[i] - 1; |
---|
258 | |
---|
259 | if( tmpptr->start1 == tmpptr->end1 ) end1 = start1; |
---|
260 | else |
---|
261 | { |
---|
262 | #if MACHIGAI |
---|
263 | while( *pt != 0 ) |
---|
264 | { |
---|
265 | // fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, tmpptr->end1, pt-seq1[i] ); |
---|
266 | if( tmpint == tmpptr->end1 ) break; |
---|
267 | if( *pt++ != '-' ) tmpint++; |
---|
268 | } |
---|
269 | end1 = pt - seq1[i] - 0; |
---|
270 | #else |
---|
271 | while( *pt != 0 ) |
---|
272 | { |
---|
273 | // fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, tmpptr->end1, pt-seq1[i] ); |
---|
274 | if( *pt++ != '-' ) tmpint++; |
---|
275 | if( tmpint == tmpptr->end1 ) break; |
---|
276 | } |
---|
277 | end1 = pt - seq1[i] - 1; |
---|
278 | #endif |
---|
279 | } |
---|
280 | |
---|
281 | pt = seq2[j]; |
---|
282 | tmpint = -1; |
---|
283 | while( *pt != 0 ) |
---|
284 | { |
---|
285 | if( *pt++ != '-' ) tmpint++; |
---|
286 | if( tmpint == tmpptr->start2 ) break; |
---|
287 | } |
---|
288 | start2 = pt - seq2[j] - 1; |
---|
289 | if( tmpptr->start2 == tmpptr->end2 ) end2 = start2; |
---|
290 | else |
---|
291 | { |
---|
292 | #if MACHIGAI |
---|
293 | while( *pt != 0 ) |
---|
294 | { |
---|
295 | if( tmpint == tmpptr->end2 ) break; |
---|
296 | if( *pt++ != '-' ) tmpint++; |
---|
297 | } |
---|
298 | end2 = pt - seq2[j] - 0; |
---|
299 | #else |
---|
300 | while( *pt != 0 ) |
---|
301 | { |
---|
302 | if( *pt++ != '-' ) tmpint++; |
---|
303 | if( tmpint == tmpptr->end2 ) break; |
---|
304 | } |
---|
305 | end2 = pt - seq2[j] - 1; |
---|
306 | #endif |
---|
307 | } |
---|
308 | // fprintf( stderr, "start1 = %d (%c), end1 = %d (%c), start2 = %d (%c), end2 = %d (%c)\n", start1, seq1[i][start1], end1, seq1[i][end1], start2, seq2[j][start2], end2, seq2[j][end2] ); |
---|
309 | // fprintf( stderr, "step 0\n" ); |
---|
310 | if( end1 - start1 != end2 - start2 ) |
---|
311 | { |
---|
312 | // fprintf( stderr, "CHUUI!!, start1 = %d, end1 = %d, start2 = %d, end2 = %d\n", start1, end1, start2, end2 ); |
---|
313 | } |
---|
314 | |
---|
315 | #if 1 |
---|
316 | k1 = start1; k2 = start2; |
---|
317 | pt1 = seq1[i] + k1; |
---|
318 | pt2 = seq2[j] + k2; |
---|
319 | while( *pt1 && *pt2 ) |
---|
320 | { |
---|
321 | if( *pt1 != '-' && *pt2 != '-' ) |
---|
322 | { |
---|
323 | // œÅ€ß€òÆóœÅ€Ë€«€±€Ê€€€è€Š€ËÃí°Õ€·€Æ²Œ€µ€€¡£ |
---|
324 | // impmtx[k1][k2] += tmpptr->wimportance * fastathreshold; |
---|
325 | // impmtx[k1][k2] += tmpptr->importance * effij; |
---|
326 | impmtx[k1][k2] += tmpptr->fimportance * effij; |
---|
327 | // fprintf( stderr, "#### impmtx[k1][k2] = %f, tmpptr->fimportance=%f, effij=%f\n", impmtx[k1][k2], tmpptr->fimportance, effij ); |
---|
328 | // fprintf( stderr, "mark, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); |
---|
329 | // fprintf( stderr, "%d (%c) - %d (%c) - %f\n", k1, *pt1, k2, *pt2, tmpptr->fimportance * effij ); |
---|
330 | k1++; k2++; |
---|
331 | pt1++; pt2++; |
---|
332 | } |
---|
333 | else if( *pt1 != '-' && *pt2 == '-' ) |
---|
334 | { |
---|
335 | // fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); |
---|
336 | k2++; pt2++; |
---|
337 | } |
---|
338 | else if( *pt1 == '-' && *pt2 != '-' ) |
---|
339 | { |
---|
340 | // fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); |
---|
341 | k1++; pt1++; |
---|
342 | } |
---|
343 | else if( *pt1 == '-' && *pt2 == '-' ) |
---|
344 | { |
---|
345 | // fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); |
---|
346 | k1++; pt1++; |
---|
347 | k2++; pt2++; |
---|
348 | } |
---|
349 | if( k1 > end1 || k2 > end2 ) break; |
---|
350 | } |
---|
351 | #else |
---|
352 | while( k1 <= end1 && k2 <= end2 ) |
---|
353 | { |
---|
354 | fprintf( stderr, "k1,k2=%d,%d - ", k1, k2 ); |
---|
355 | if( !nocount1[k1] && !nocount2[k2] ) |
---|
356 | { |
---|
357 | impmtx[k1][k2] += tmpptr->wimportance * eff1[i] * eff2[j] * fastathreshold; |
---|
358 | fprintf( stderr, "marked\n" ); |
---|
359 | } |
---|
360 | else |
---|
361 | fprintf( stderr, "no count\n" ); |
---|
362 | k1++; k2++; |
---|
363 | } |
---|
364 | #endif |
---|
365 | tmpptr = tmpptr->next; |
---|
366 | } |
---|
367 | } |
---|
368 | } |
---|
369 | #if 0 |
---|
370 | if( clus1 == 1 && clus2 == 6 ) |
---|
371 | { |
---|
372 | fprintf( stderr, "\n" ); |
---|
373 | fprintf( stderr, "seq1[0] = %s\n", seq1[0] ); |
---|
374 | fprintf( stderr, "seq2[0] = %s\n", seq2[0] ); |
---|
375 | fprintf( stderr, "impmtx = \n" ); |
---|
376 | for( k2=0; k2<lgth2; k2++ ) |
---|
377 | fprintf( stderr, "%6.3f ", (double)k2 ); |
---|
378 | fprintf( stderr, "\n" ); |
---|
379 | for( k1=0; k1<lgth1; k1++ ) |
---|
380 | { |
---|
381 | fprintf( stderr, "%d ", k1 ); |
---|
382 | for( k2=0; k2<3; k2++ ) |
---|
383 | fprintf( stderr, "%2.1f ", impmtx[k1][k2] ); |
---|
384 | fprintf( stderr, "\n" ); |
---|
385 | } |
---|
386 | exit( 1 ); |
---|
387 | } |
---|
388 | #endif |
---|
389 | } |
---|
390 | #endif |
---|
391 | |
---|
392 | |
---|
393 | static void match_calc( float *match, float **cpmx1, float **cpmx2, int i1, int lgth2, float **floatwork, int **intwork, int initialize ) |
---|
394 | { |
---|
395 | #if FASTMATCHCALC |
---|
396 | int j, l; |
---|
397 | float scarr[26]; |
---|
398 | float **cpmxpd = floatwork; |
---|
399 | int **cpmxpdn = intwork; |
---|
400 | float *matchpt, *cpmxpdpt, **cpmxpdptpt; |
---|
401 | int *cpmxpdnpt, **cpmxpdnptpt; |
---|
402 | if( initialize ) |
---|
403 | { |
---|
404 | int count = 0; |
---|
405 | for( j=0; j<lgth2; j++ ) |
---|
406 | { |
---|
407 | count = 0; |
---|
408 | for( l=0; l<26; l++ ) |
---|
409 | { |
---|
410 | if( cpmx2[l][j] ) |
---|
411 | { |
---|
412 | cpmxpd[j][count] = cpmx2[l][j]; |
---|
413 | cpmxpdn[j][count] = l; |
---|
414 | count++; |
---|
415 | } |
---|
416 | } |
---|
417 | cpmxpdn[j][count] = -1; |
---|
418 | } |
---|
419 | } |
---|
420 | |
---|
421 | { |
---|
422 | for( l=0; l<26; l++ ) |
---|
423 | { |
---|
424 | scarr[l] = 0.0; |
---|
425 | for( j=0; j<26; j++ ) |
---|
426 | scarr[l] += n_dis[j][l] * cpmx1[j][i1]; |
---|
427 | } |
---|
428 | matchpt = match; |
---|
429 | cpmxpdnptpt = cpmxpdn; |
---|
430 | cpmxpdptpt = cpmxpd; |
---|
431 | while( lgth2-- ) |
---|
432 | { |
---|
433 | *matchpt = 0.0; |
---|
434 | cpmxpdnpt = *cpmxpdnptpt++; |
---|
435 | cpmxpdpt = *cpmxpdptpt++; |
---|
436 | while( *cpmxpdnpt>-1 ) |
---|
437 | *matchpt += scarr[*cpmxpdnpt++] * *cpmxpdpt++; |
---|
438 | matchpt++; |
---|
439 | } |
---|
440 | } |
---|
441 | #else |
---|
442 | int j, k, l; |
---|
443 | float scarr[26]; |
---|
444 | float **cpmxpd = floatwork; |
---|
445 | int **cpmxpdn = intwork; |
---|
446 | // simple |
---|
447 | if( initialize ) |
---|
448 | { |
---|
449 | int count = 0; |
---|
450 | for( j=0; j<lgth2; j++ ) |
---|
451 | { |
---|
452 | count = 0; |
---|
453 | for( l=0; l<26; l++ ) |
---|
454 | { |
---|
455 | if( cpmx2[l][j] ) |
---|
456 | { |
---|
457 | cpmxpd[count][j] = cpmx2[l][j]; |
---|
458 | cpmxpdn[count][j] = l; |
---|
459 | count++; |
---|
460 | } |
---|
461 | } |
---|
462 | cpmxpdn[count][j] = -1; |
---|
463 | } |
---|
464 | } |
---|
465 | for( l=0; l<26; l++ ) |
---|
466 | { |
---|
467 | scarr[l] = 0.0; |
---|
468 | for( k=0; k<26; k++ ) |
---|
469 | scarr[l] += n_dis[k][l] * cpmx1[k][i1]; |
---|
470 | } |
---|
471 | for( j=0; j<lgth2; j++ ) |
---|
472 | { |
---|
473 | match[j] = 0.0; |
---|
474 | for( k=0; cpmxpdn[k][j]>-1; k++ ) |
---|
475 | match[j] += scarr[cpmxpdn[k][j]] * cpmxpd[k][j]; |
---|
476 | } |
---|
477 | #endif |
---|
478 | } |
---|
479 | |
---|
480 | static void Atracking_localhom( float *impwmpt, float *lasthorizontalw, float *lastverticalw, |
---|
481 | char **seq1, char **seq2, |
---|
482 | char **mseq1, char **mseq2, |
---|
483 | float **cpmx1, float **cpmx2, |
---|
484 | int **ijp, int icyc, int jcyc ) |
---|
485 | { |
---|
486 | int i, j, l, iin, jin, ifi, jfi, lgth1, lgth2, k; |
---|
487 | float wm; |
---|
488 | char *gaptable1, *gt1bk; |
---|
489 | char *gaptable2, *gt2bk; |
---|
490 | lgth1 = strlen( seq1[0] ); |
---|
491 | lgth2 = strlen( seq2[0] ); |
---|
492 | gt1bk = AllocateCharVec( lgth1+lgth2+1 ); |
---|
493 | gt2bk = AllocateCharVec( lgth1+lgth2+1 ); |
---|
494 | |
---|
495 | #if 0 |
---|
496 | for( i=0; i<lgth1; i++ ) |
---|
497 | { |
---|
498 | fprintf( stderr, "lastverticalw[%d] = %f\n", i, lastverticalw[i] ); |
---|
499 | } |
---|
500 | #endif |
---|
501 | |
---|
502 | if( outgap == 1 ) |
---|
503 | ; |
---|
504 | else |
---|
505 | { |
---|
506 | wm = lastverticalw[0]; |
---|
507 | for( i=0; i<lgth1; i++ ) |
---|
508 | { |
---|
509 | if( lastverticalw[i] >= wm ) |
---|
510 | { |
---|
511 | wm = lastverticalw[i]; |
---|
512 | iin = i; jin = lgth2-1; |
---|
513 | ijp[lgth1][lgth2] = +( lgth1 - i ); |
---|
514 | } |
---|
515 | } |
---|
516 | for( j=0; j<lgth2; j++ ) |
---|
517 | { |
---|
518 | if( lasthorizontalw[j] >= wm ) |
---|
519 | { |
---|
520 | wm = lasthorizontalw[j]; |
---|
521 | iin = lgth1-1; jin = j; |
---|
522 | ijp[lgth1][lgth2] = -( lgth2 - j ); |
---|
523 | } |
---|
524 | } |
---|
525 | } |
---|
526 | |
---|
527 | for( i=0; i<lgth1+1; i++ ) |
---|
528 | { |
---|
529 | ijp[i][0] = i + 1; |
---|
530 | } |
---|
531 | for( j=0; j<lgth2+1; j++ ) |
---|
532 | { |
---|
533 | ijp[0][j] = -( j + 1 ); |
---|
534 | } |
---|
535 | |
---|
536 | gaptable1 = gt1bk + lgth1+lgth2; |
---|
537 | *gaptable1 = 0; |
---|
538 | gaptable2 = gt2bk + lgth1+lgth2; |
---|
539 | *gaptable2 = 0; |
---|
540 | |
---|
541 | iin = lgth1; jin = lgth2; |
---|
542 | *impwmpt = 0.0; |
---|
543 | for( k=0; k<=lgth1+lgth2; k++ ) |
---|
544 | { |
---|
545 | if( ijp[iin][jin] < 0 ) |
---|
546 | { |
---|
547 | ifi = iin-1; jfi = jin+ijp[iin][jin]; |
---|
548 | } |
---|
549 | else if( ijp[iin][jin] > 0 ) |
---|
550 | { |
---|
551 | ifi = iin-ijp[iin][jin]; jfi = jin-1; |
---|
552 | } |
---|
553 | else |
---|
554 | { |
---|
555 | ifi = iin-1; jfi = jin-1; |
---|
556 | } |
---|
557 | l = iin - ifi; |
---|
558 | while( --l ) |
---|
559 | { |
---|
560 | *--gaptable1 = 'o'; |
---|
561 | *--gaptable2 = '-'; |
---|
562 | k++; |
---|
563 | } |
---|
564 | l= jin - jfi; |
---|
565 | while( --l ) |
---|
566 | { |
---|
567 | *--gaptable1 = '-'; |
---|
568 | *--gaptable2 = 'o'; |
---|
569 | k++; |
---|
570 | } |
---|
571 | if( iin == lgth1 || jin == lgth2 ) |
---|
572 | ; |
---|
573 | else |
---|
574 | { |
---|
575 | *impwmpt += imp_match_out_scH( iin, jin ); |
---|
576 | |
---|
577 | // fprintf( stderr, "impwm = %f (iin=%d, jin=%d) seq1=%c, seq2=%c\n", *impwmpt, iin, jin, seq1[0][iin], seq2[0][jin] ); |
---|
578 | } |
---|
579 | if( iin <= 0 || jin <= 0 ) break; |
---|
580 | *--gaptable1 = 'o'; |
---|
581 | *--gaptable2 = 'o'; |
---|
582 | k++; |
---|
583 | iin = ifi; jin = jfi; |
---|
584 | } |
---|
585 | |
---|
586 | for( i=0; i<icyc; i++ ) gapireru( mseq1[i], seq1[i], gaptable1 ); |
---|
587 | for( j=0; j<jcyc; j++ ) gapireru( mseq2[j], seq2[j], gaptable2 ); |
---|
588 | |
---|
589 | free( gt1bk ); |
---|
590 | free( gt2bk ); |
---|
591 | } |
---|
592 | |
---|
593 | static float Atracking( float *lasthorizontalw, float *lastverticalw, |
---|
594 | char **seq1, char **seq2, |
---|
595 | char **mseq1, char **mseq2, |
---|
596 | float **cpmx1, float **cpmx2, |
---|
597 | int **ijp, int icyc, int jcyc ) |
---|
598 | { |
---|
599 | int i, j, l, iin, jin, ifi, jfi, lgth1, lgth2, k; |
---|
600 | float wm; |
---|
601 | char *gaptable1, *gt1bk; |
---|
602 | char *gaptable2, *gt2bk; |
---|
603 | lgth1 = strlen( seq1[0] ); |
---|
604 | lgth2 = strlen( seq2[0] ); |
---|
605 | |
---|
606 | gt1bk = AllocateCharVec( lgth1+lgth2+1 ); |
---|
607 | gt2bk = AllocateCharVec( lgth1+lgth2+1 ); |
---|
608 | |
---|
609 | #if 0 |
---|
610 | for( i=0; i<lgth1; i++ ) |
---|
611 | { |
---|
612 | fprintf( stderr, "lastverticalw[%d] = %f\n", i, lastverticalw[i] ); |
---|
613 | } |
---|
614 | #endif |
---|
615 | |
---|
616 | if( outgap == 1 ) |
---|
617 | ; |
---|
618 | else |
---|
619 | { |
---|
620 | wm = lastverticalw[0]; |
---|
621 | for( i=0; i<lgth1; i++ ) |
---|
622 | { |
---|
623 | if( lastverticalw[i] >= wm ) |
---|
624 | { |
---|
625 | wm = lastverticalw[i]; |
---|
626 | iin = i; jin = lgth2-1; |
---|
627 | ijp[lgth1][lgth2] = +( lgth1 - i ); |
---|
628 | } |
---|
629 | } |
---|
630 | for( j=0; j<lgth2; j++ ) |
---|
631 | { |
---|
632 | if( lasthorizontalw[j] >= wm ) |
---|
633 | { |
---|
634 | wm = lasthorizontalw[j]; |
---|
635 | iin = lgth1-1; jin = j; |
---|
636 | ijp[lgth1][lgth2] = -( lgth2 - j ); |
---|
637 | } |
---|
638 | } |
---|
639 | } |
---|
640 | |
---|
641 | for( i=0; i<lgth1+1; i++ ) |
---|
642 | { |
---|
643 | ijp[i][0] = i + 1; |
---|
644 | } |
---|
645 | for( j=0; j<lgth2+1; j++ ) |
---|
646 | { |
---|
647 | ijp[0][j] = -( j + 1 ); |
---|
648 | } |
---|
649 | |
---|
650 | gaptable1 = gt1bk + lgth1+lgth2; |
---|
651 | *gaptable1 = 0; |
---|
652 | gaptable2 = gt2bk + lgth1+lgth2; |
---|
653 | *gaptable2 = 0; |
---|
654 | |
---|
655 | iin = lgth1; jin = lgth2; |
---|
656 | for( k=0; k<=lgth1+lgth2; k++ ) |
---|
657 | { |
---|
658 | if( ijp[iin][jin] < 0 ) |
---|
659 | { |
---|
660 | ifi = iin-1; jfi = jin+ijp[iin][jin]; |
---|
661 | } |
---|
662 | else if( ijp[iin][jin] > 0 ) |
---|
663 | { |
---|
664 | ifi = iin-ijp[iin][jin]; jfi = jin-1; |
---|
665 | } |
---|
666 | else |
---|
667 | { |
---|
668 | ifi = iin-1; jfi = jin-1; |
---|
669 | } |
---|
670 | l = iin - ifi; |
---|
671 | while( --l ) |
---|
672 | { |
---|
673 | *--gaptable1 = 'o'; |
---|
674 | *--gaptable2 = '-'; |
---|
675 | k++; |
---|
676 | } |
---|
677 | l= jin - jfi; |
---|
678 | while( --l ) |
---|
679 | { |
---|
680 | *--gaptable1 = '-'; |
---|
681 | *--gaptable2 = 'o'; |
---|
682 | k++; |
---|
683 | } |
---|
684 | if( iin <= 0 || jin <= 0 ) break; |
---|
685 | *--gaptable1 = 'o'; |
---|
686 | *--gaptable2 = 'o'; |
---|
687 | k++; |
---|
688 | iin = ifi; jin = jfi; |
---|
689 | } |
---|
690 | |
---|
691 | for( i=0; i<icyc; i++ ) gapireru( mseq1[i], seq1[i], gaptable1 ); |
---|
692 | for( j=0; j<jcyc; j++ ) gapireru( mseq2[j], seq2[j], gaptable2 ); |
---|
693 | |
---|
694 | free( gt1bk ); |
---|
695 | free( gt2bk ); |
---|
696 | |
---|
697 | return( 0.0 ); |
---|
698 | } |
---|
699 | |
---|
700 | float H__align( char **seq1, char **seq2, double *eff1, double *eff2, int icyc, int jcyc, int alloclen, LocalHom ***localhom, float *impmatch, char *sgap1, char *sgap2, char *egap1, char *egap2 ) |
---|
701 | /* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */ |
---|
702 | { |
---|
703 | // int k; |
---|
704 | register int i, j; |
---|
705 | int lasti, lastj; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */ |
---|
706 | int lgth1, lgth2; |
---|
707 | int resultlen; |
---|
708 | float wm = 0.0; /* int ?????? */ |
---|
709 | float g; |
---|
710 | float *currentw, *previousw; |
---|
711 | // float fpenalty = (float)penalty; |
---|
712 | #if USE_PENALTY_EX |
---|
713 | float fpenalty_ex = (float)penalty_ex; |
---|
714 | #endif |
---|
715 | #if 1 |
---|
716 | float *wtmp; |
---|
717 | int *ijppt; |
---|
718 | float *mjpt, *prept, *curpt; |
---|
719 | int *mpjpt; |
---|
720 | #endif |
---|
721 | static TLS float mi, *m; |
---|
722 | static TLS int **ijp; |
---|
723 | static TLS int mpi, *mp; |
---|
724 | static TLS float *w1, *w2; |
---|
725 | static TLS float *match; |
---|
726 | static TLS float *initverticalw; /* kufuu sureba iranai */ |
---|
727 | static TLS float *lastverticalw; /* kufuu sureba iranai */ |
---|
728 | static TLS char **mseq1; |
---|
729 | static TLS char **mseq2; |
---|
730 | static TLS char **mseq; |
---|
731 | static TLS Gappat **gappat1; |
---|
732 | static TLS Gappat **gappat2; |
---|
733 | static TLS float *digf1; |
---|
734 | static TLS float *digf2; |
---|
735 | static TLS float *diaf1; |
---|
736 | static TLS float *diaf2; |
---|
737 | static TLS float *gapz1; |
---|
738 | static TLS float *gapz2; |
---|
739 | static TLS float *gapf1; |
---|
740 | static TLS float *gapf2; |
---|
741 | static TLS float *ogcp1g; |
---|
742 | static TLS float *ogcp2g; |
---|
743 | static TLS float *fgcp1g; |
---|
744 | static TLS float *fgcp2g; |
---|
745 | static TLS float *ogcp1; |
---|
746 | static TLS float *ogcp2; |
---|
747 | static TLS float *fgcp1; |
---|
748 | static TLS float *fgcp2; |
---|
749 | static TLS float **cpmx1; |
---|
750 | static TLS float **cpmx2; |
---|
751 | static TLS int **intwork; |
---|
752 | static TLS float **floatwork; |
---|
753 | static TLS int orlgth1 = 0, orlgth2 = 0; |
---|
754 | float fpenalty = (float)penalty; |
---|
755 | float tmppenal; |
---|
756 | float cumpenal; |
---|
757 | float *fgcp2pt; |
---|
758 | float *ogcp2pt; |
---|
759 | float fgcp1va; |
---|
760 | float ogcp1va; |
---|
761 | int maegap; |
---|
762 | |
---|
763 | |
---|
764 | |
---|
765 | #if 0 |
---|
766 | fprintf( stderr, "#### eff in SA+++align\n" ); |
---|
767 | fprintf( stderr, "#### seq1[0] = %s\n", seq1[0] ); |
---|
768 | fprintf( stderr, "#### strlen( seq1[0] ) = %d\n", strlen( seq1[0] ) ); |
---|
769 | for( i=0; i<icyc; i++ ) fprintf( stderr, "eff1[%d] = %f\n", i, eff1[i] ); |
---|
770 | #endif |
---|
771 | if( orlgth1 == 0 ) |
---|
772 | { |
---|
773 | mseq1 = AllocateCharMtx( njob, 0 ); |
---|
774 | mseq2 = AllocateCharMtx( njob, 0 ); |
---|
775 | } |
---|
776 | |
---|
777 | |
---|
778 | lgth1 = strlen( seq1[0] ); |
---|
779 | lgth2 = strlen( seq2[0] ); |
---|
780 | #if 0 |
---|
781 | if( lgth1 == 0 || lgth2 == 0 ) |
---|
782 | { |
---|
783 | fprintf( stderr, "WARNING (Aalignmm): lgth1=%d, lgth2=%d\n", lgth1, lgth2 ); |
---|
784 | } |
---|
785 | #endif |
---|
786 | |
---|
787 | if( lgth1 > orlgth1 || lgth2 > orlgth2 ) |
---|
788 | { |
---|
789 | int ll1, ll2; |
---|
790 | |
---|
791 | if( orlgth1 > 0 && orlgth2 > 0 ) |
---|
792 | { |
---|
793 | FreeFloatVec( w1 ); |
---|
794 | FreeFloatVec( w2 ); |
---|
795 | FreeFloatVec( match ); |
---|
796 | FreeFloatVec( initverticalw ); |
---|
797 | FreeFloatVec( lastverticalw ); |
---|
798 | |
---|
799 | FreeFloatVec( m ); |
---|
800 | FreeIntVec( mp ); |
---|
801 | |
---|
802 | FreeCharMtx( mseq ); |
---|
803 | |
---|
804 | free( gappat1 ); |
---|
805 | free( gappat2 ); |
---|
806 | FreeFloatVec( digf1 ); |
---|
807 | FreeFloatVec( digf2 ); |
---|
808 | FreeFloatVec( diaf1 ); |
---|
809 | FreeFloatVec( diaf2 ); |
---|
810 | FreeFloatVec( gapz1 ); |
---|
811 | FreeFloatVec( gapz2 ); |
---|
812 | FreeFloatVec( gapf1 ); |
---|
813 | FreeFloatVec( gapf2 ); |
---|
814 | FreeFloatVec( ogcp1 ); |
---|
815 | FreeFloatVec( ogcp2 ); |
---|
816 | FreeFloatVec( fgcp1 ); |
---|
817 | FreeFloatVec( fgcp2 ); |
---|
818 | FreeFloatVec( ogcp1g ); |
---|
819 | FreeFloatVec( ogcp2g ); |
---|
820 | FreeFloatVec( fgcp1g ); |
---|
821 | FreeFloatVec( fgcp2g ); |
---|
822 | |
---|
823 | |
---|
824 | FreeFloatMtx( cpmx1 ); |
---|
825 | FreeFloatMtx( cpmx2 ); |
---|
826 | |
---|
827 | FreeFloatMtx( floatwork ); |
---|
828 | FreeIntMtx( intwork ); |
---|
829 | } |
---|
830 | |
---|
831 | ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100; |
---|
832 | ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100; |
---|
833 | |
---|
834 | #if DEBUG |
---|
835 | fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 ); |
---|
836 | #endif |
---|
837 | |
---|
838 | w1 = AllocateFloatVec( ll2+2 ); |
---|
839 | w2 = AllocateFloatVec( ll2+2 ); |
---|
840 | match = AllocateFloatVec( ll2+2 ); |
---|
841 | |
---|
842 | initverticalw = AllocateFloatVec( ll1+2 ); |
---|
843 | lastverticalw = AllocateFloatVec( ll1+2 ); |
---|
844 | |
---|
845 | m = AllocateFloatVec( ll2+2 ); |
---|
846 | mp = AllocateIntVec( ll2+2 ); |
---|
847 | |
---|
848 | mseq = AllocateCharMtx( njob, ll1+ll2 ); |
---|
849 | |
---|
850 | digf1 = AllocateFloatVec( ll1+2 ); |
---|
851 | digf2 = AllocateFloatVec( ll2+2 ); |
---|
852 | diaf1 = AllocateFloatVec( ll1+2 ); |
---|
853 | diaf2 = AllocateFloatVec( ll2+2 ); |
---|
854 | gappat1 = (Gappat **)calloc( ll1+2, sizeof( Gappat * ) ); |
---|
855 | gappat2 = (Gappat **)calloc( ll2+2, sizeof( Gappat * ) ); |
---|
856 | gapz1 = AllocateFloatVec( ll1+2 ); |
---|
857 | gapz2 = AllocateFloatVec( ll2+2 ); |
---|
858 | gapf1 = AllocateFloatVec( ll1+2 ); |
---|
859 | gapf2 = AllocateFloatVec( ll2+2 ); |
---|
860 | ogcp1 = AllocateFloatVec( ll1+2 ); |
---|
861 | ogcp2 = AllocateFloatVec( ll2+2 ); |
---|
862 | fgcp1 = AllocateFloatVec( ll1+2 ); |
---|
863 | fgcp2 = AllocateFloatVec( ll2+2 ); |
---|
864 | ogcp1g = AllocateFloatVec( ll1+2 ); |
---|
865 | ogcp2g = AllocateFloatVec( ll2+2 ); |
---|
866 | fgcp1g = AllocateFloatVec( ll1+2 ); |
---|
867 | fgcp2g = AllocateFloatVec( ll2+2 ); |
---|
868 | |
---|
869 | cpmx1 = AllocateFloatMtx( 26, ll1+2 ); |
---|
870 | cpmx2 = AllocateFloatMtx( 26, ll2+2 ); |
---|
871 | |
---|
872 | #if FASTMATCHCALC |
---|
873 | floatwork = AllocateFloatMtx( MAX( ll1, ll2 )+2, 26 ); |
---|
874 | intwork = AllocateIntMtx( MAX( ll1, ll2 )+2, 27 ); |
---|
875 | #else |
---|
876 | floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 ); |
---|
877 | intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 ); |
---|
878 | #endif |
---|
879 | |
---|
880 | #if DEBUG |
---|
881 | fprintf( stderr, "succeeded\n" ); |
---|
882 | #endif |
---|
883 | |
---|
884 | orlgth1 = ll1 - 100; |
---|
885 | orlgth2 = ll2 - 100; |
---|
886 | } |
---|
887 | |
---|
888 | |
---|
889 | for( i=0; i<icyc; i++ ) |
---|
890 | { |
---|
891 | mseq1[i] = mseq[i]; |
---|
892 | seq1[i][lgth1] = 0; |
---|
893 | } |
---|
894 | for( j=0; j<jcyc; j++ ) |
---|
895 | { |
---|
896 | mseq2[j] = mseq[icyc+j]; |
---|
897 | seq2[j][lgth2] = 0; |
---|
898 | } |
---|
899 | |
---|
900 | |
---|
901 | if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 ) |
---|
902 | { |
---|
903 | int ll1, ll2; |
---|
904 | |
---|
905 | if( commonAlloc1 && commonAlloc2 ) |
---|
906 | { |
---|
907 | FreeIntMtx( commonIP ); |
---|
908 | } |
---|
909 | |
---|
910 | ll1 = MAX( orlgth1, commonAlloc1 ); |
---|
911 | ll2 = MAX( orlgth2, commonAlloc2 ); |
---|
912 | |
---|
913 | #if DEBUG |
---|
914 | fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 ); |
---|
915 | #endif |
---|
916 | |
---|
917 | commonIP = AllocateIntMtx( ll1+10, ll2+10 ); |
---|
918 | |
---|
919 | #if DEBUG |
---|
920 | fprintf( stderr, "succeeded\n\n" ); |
---|
921 | #endif |
---|
922 | |
---|
923 | commonAlloc1 = ll1; |
---|
924 | commonAlloc2 = ll2; |
---|
925 | } |
---|
926 | ijp = commonIP; |
---|
927 | |
---|
928 | #if 0 |
---|
929 | { |
---|
930 | float t = 0.0; |
---|
931 | for( i=0; i<icyc; i++ ) |
---|
932 | t += eff1[i]; |
---|
933 | fprintf( stderr, "## totaleff = %f\n", t ); |
---|
934 | } |
---|
935 | #endif |
---|
936 | |
---|
937 | cpmx_calc_new( seq1, cpmx1, eff1, lgth1, icyc ); |
---|
938 | cpmx_calc_new( seq2, cpmx2, eff2, lgth2, jcyc ); |
---|
939 | |
---|
940 | if( sgap1 ) |
---|
941 | { |
---|
942 | new_OpeningGapCount_zure( ogcp1g, icyc, seq1, eff1, lgth1, sgap1, egap1 ); |
---|
943 | new_OpeningGapCount_zure( ogcp2g, jcyc, seq2, eff2, lgth2, sgap2, egap1 ); |
---|
944 | new_FinalGapCount_zure( fgcp1g, icyc, seq1, eff1, lgth1, sgap1, egap1 ); |
---|
945 | new_FinalGapCount_zure( fgcp2g, jcyc, seq2, eff2, lgth2, sgap1, egap2 ); |
---|
946 | getdigapfreq_part( digf1, icyc, seq1, eff1, lgth1, sgap1, egap1 ); // sgap1 ha iranai ? |
---|
947 | getdigapfreq_part( digf2, jcyc, seq2, eff2, lgth2, sgap2, egap2 ); // sgap1 ha iranai ? |
---|
948 | getdiaminofreq_part( diaf1, icyc, seq1, eff1, lgth1, sgap1, egap1 ); // sgap1 ha iranai ? |
---|
949 | getdiaminofreq_part( diaf2, jcyc, seq2, eff2, lgth2, sgap1, egap2 ); // sgap1 ha iranai ? |
---|
950 | getgapfreq( gapf1, icyc, seq1, eff1, lgth1 ); // atode |
---|
951 | getgapfreq( gapf2, jcyc, seq2, eff2, lgth2 ); // atode |
---|
952 | getgapfreq_zure( gapz1, icyc, seq1, eff1, lgth1 ); // atode |
---|
953 | getgapfreq_zure( gapz2, jcyc, seq2, eff2, lgth2 ); // atode |
---|
954 | } |
---|
955 | else |
---|
956 | { |
---|
957 | st_OpeningGapCount( ogcp1g, icyc, seq1, eff1, lgth1 ); |
---|
958 | st_OpeningGapCount( ogcp2g, jcyc, seq2, eff2, lgth2 ); |
---|
959 | st_FinalGapCount_zure( fgcp1g, icyc, seq1, eff1, lgth1 ); |
---|
960 | st_FinalGapCount_zure( fgcp2g, jcyc, seq2, eff2, lgth2 ); |
---|
961 | st_getGapPattern( gappat1, icyc, seq1, eff1, lgth1 ); |
---|
962 | st_getGapPattern( gappat2, jcyc, seq2, eff2, lgth2 ); |
---|
963 | getdigapfreq_st( digf1, icyc, seq1, eff1, lgth1 ); |
---|
964 | getdigapfreq_st( digf2, jcyc, seq2, eff2, lgth2 ); |
---|
965 | getdiaminofreq_x( diaf1, icyc, seq1, eff1, lgth1 ); |
---|
966 | getdiaminofreq_x( diaf2, jcyc, seq2, eff2, lgth2 ); |
---|
967 | getgapfreq( gapf1, icyc, seq1, eff1, lgth1 ); |
---|
968 | getgapfreq( gapf2, jcyc, seq2, eff2, lgth2 ); |
---|
969 | getgapfreq_zure( gapz1, icyc, seq1, eff1, lgth1 ); |
---|
970 | getgapfreq_zure( gapz2, jcyc, seq2, eff2, lgth2 ); |
---|
971 | } |
---|
972 | |
---|
973 | #if 0 |
---|
974 | for( i=0; i<lgth1; i++ ) |
---|
975 | fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] ); |
---|
976 | #endif |
---|
977 | |
---|
978 | currentw = w1; |
---|
979 | previousw = w2; |
---|
980 | |
---|
981 | match_calc( initverticalw, cpmx2, cpmx1, 0, lgth1, floatwork, intwork, 1 ); |
---|
982 | if( localhom ) |
---|
983 | imp_match_out_vead_tateH( initverticalw, 0, lgth1 ); // 060306 |
---|
984 | |
---|
985 | match_calc( currentw, cpmx1, cpmx2, 0, lgth2, floatwork, intwork, 1 ); |
---|
986 | if( localhom ) |
---|
987 | imp_match_out_veadH( currentw, 0, lgth2 ); // 060306 |
---|
988 | |
---|
989 | |
---|
990 | #if 0 // -> tbfast.c |
---|
991 | if( localhom ) |
---|
992 | imp_match_calc( currentw, icyc, jcyc, lgth1, lgth2, seq1, seq2, eff1, eff2, localhom, 1, 0 ); |
---|
993 | |
---|
994 | #endif |
---|
995 | |
---|
996 | if( outgap == 1 ) |
---|
997 | { |
---|
998 | // if( g ) fprintf( stderr, "init-match penal1=%f, %c-%c\n", g, seq1[0][0], seq2[0][0] ); |
---|
999 | // initverticalw[0] += g; |
---|
1000 | // currentw[0] += g; |
---|
1001 | |
---|
1002 | // if( g ) fprintf( stderr, "init-match penal2=%f, %c-%c\n", g, seq1[0][0], seq2[0][0] ); |
---|
1003 | // initverticalw[0] += g; |
---|
1004 | // currentw[0] += g; |
---|
1005 | |
---|
1006 | for( i=1; i<lgth1+1; i++ ) |
---|
1007 | { |
---|
1008 | // initverticalw[i] += ( ogcp1[0] + fgcp1[i-1] ) ; |
---|
1009 | |
---|
1010 | tmppenal = 0.0; |
---|
1011 | // tmppenal = ( (1.0-gapf2[j-1])*(1.0-ogcp1g[i]+fgcp1g[i]) + gapf2[j-1]*(1.0-digf1[i]-diaf1[i]) ) * 0.5 * fpenalty; // mada |
---|
1012 | // tmppenal = ( (1.0-gapz2[0])*(1.0-ogcp1g[0]+0.0) + gapz2[0]*(1.0-digf1[0] - diaf1[0]) ) * 0.5 * fpenalty; // mada |
---|
1013 | // tmppenal = ( (1.0-0.0)*(1.0-ogcp1g[0]+0.0) + 0.0*(1.0-0.0-0.0) ) * 0.5 * fpenalty; // mada |
---|
1014 | // tmppenal = 0.5 * fpenalty; |
---|
1015 | // tmppenal -= ( (1.0-0.0) * (1.0-diaf1[0]) + 0.0 ) * 0.5 * fpenalty; // 0. |
---|
1016 | // tmppenal -= ( (1.0-gapf2[j-1]) * ogcp1g[i] + gapf2[j-1] ) * 0.5 * fpenalty; |
---|
1017 | // fprintf( stderr, "0,0<-%d,%d, tmppenal 1 = %f\n", i, 0, tmppenal ); |
---|
1018 | initverticalw[i] += tmppenal; |
---|
1019 | |
---|
1020 | tmppenal = diaf1[i] * ( 1.0 - gapf2[0] ) * fpenalty; |
---|
1021 | if( gappat1[i][0].freq ) |
---|
1022 | { |
---|
1023 | tmppenal += ( gappat1[i][0].freq ) * ( 1.0 - gapf2[0] ) * fpenalty; |
---|
1024 | tmppenal -= ( countnocountx( gappat2[0], diaf2[0], gappat1[i], i, 1 ) ) * fpenalty; |
---|
1025 | } |
---|
1026 | // tmppenal = ( (1.0-gapf2[j])*(1.0-fgcp1g[i]+ogcp1g[i]) + gapf2[j]*(1.0-digf1[i]-diaf1[i]) ) * 0.5 * fpenalty; // mada |
---|
1027 | // tmppenal = ( (1.0-gapz2[1])*(1.0-fgcp1g[i]+ogcp1g[i]) + gapz2[1]*(1.0-digf1[i]-diaf1[i]) ) * 0.5 * fpenalty; // mada |
---|
1028 | // tmppenal = ( (1.0-gapf2[0])*(1.0-fgcp1g[i]+ogcp1g[i]) + gapf2[0]*(1.0-digf1[i]-diaf1[i]) ) * 0.5 * fpenalty; // mada |
---|
1029 | // tmppenal = 0.5 * fpenalty; |
---|
1030 | // tmppenal -= ( (1.0-gapf2[0]) * (1.0-diaf1[i]) + gapf2[0] ) * 0.5 * fpenalty; |
---|
1031 | // tmppenal -= ( (1.0-gapf2[j]) * fgcp1g[i] + gapf2[j] ) * 0.5 * fpenalty; |
---|
1032 | initverticalw[i] += tmppenal; |
---|
1033 | // fprintf( stderr, "0,0<-%d,%d, tmppenal 2 = %f, cumpenal=%f, fgcp1g[i]=%f, ogcp1g[i]=%f\n", i, 0, tmppenal, cumpenal, fgcp1g[i], ogcp1g[i] ); |
---|
1034 | |
---|
1035 | } |
---|
1036 | cumpenal = 0.0; |
---|
1037 | for( j=1; j<lgth2+1; j++ ) |
---|
1038 | { |
---|
1039 | // currentw[j] += ( ogcp2[0] + fgcp2[j-1] ) ; |
---|
1040 | |
---|
1041 | tmppenal = 0.0; |
---|
1042 | // tmppenal = ( (1.0-gapf1[i-1])*(1.0-ogcp2g[j]+fgcp2g[j]) + gapf1[i-1]*(1.0-digf2[j]-diaf2[j]) ) * 0.5 * fpenalty; // mada |
---|
1043 | // tmppenal = ( (1.0-gapz1[0])*(1.0-ogcp2g[0]+0.0) + gapz1[0]*(1.0-digf2[j]-diaf2[j]) ) * 0.5 * fpenalty; // mada |
---|
1044 | // tmppenal = ( (1.0-0.0)*(1.0-ogcp2g[0]+0.0) + 0.0*(1.0-0.0-0.0) ) * 0.5 * fpenalty; // mada |
---|
1045 | // tmppenal = 0.5 * fpenalty; |
---|
1046 | // tmppenal -= ( (1.0-0.0) * (1.0-diaf2[0]) + 0.0 ) * 0.5 * fpenalty; // 0. |
---|
1047 | // tmppenal -= ( (1.0-gapf1[0]) * fgcp2g[j] + gapf1[0] ) * 0.5 * fpenalty; |
---|
1048 | // fprintf( stderr, "0,0<-%d,%d, tmppenal 3 = %f\n", 0, j, tmppenal ); |
---|
1049 | currentw[j] += tmppenal; |
---|
1050 | |
---|
1051 | tmppenal = diaf2[j] * ( 1.0 - gapf1[0] ) * fpenalty; |
---|
1052 | if( gappat2[j][0].freq ) |
---|
1053 | { |
---|
1054 | tmppenal += ( gappat2[j][0].freq ) * ( 1.0 - gapf1[0] ) * fpenalty; |
---|
1055 | tmppenal -= ( countnocountx( gappat1[0], diaf1[0], gappat2[j], j, 1 ) ) * fpenalty; |
---|
1056 | } |
---|
1057 | // tmppenal = ( (1.0-gapf1[i])*(1.0-fgcp2g[j]+ogcp2g[j]) + gapf1[i]*(1.0-digf2[j]-diaf2[j]) ) * 0.5 * fpenalty; // mada |
---|
1058 | // tmppenal = ( (1.0-gapz1[1])*(1.0-fgcp2g[j]+ogcp2g[j]) + gapz1[1]*(1.0-digf2[j]-diaf2[j]) ) * 0.5 * fpenalty; // mada |
---|
1059 | // tmppenal = ( (1.0-gapf1[0])*(1.0-fgcp2g[j]+ogcp2g[j]) + gapf1[0]*(1.0-digf2[j]-diaf2[j]) ) * 0.5 * fpenalty; // mada |
---|
1060 | // tmppenal = 0.5 * fpenalty; |
---|
1061 | // tmppenal -= ( (1.0-gapf1[0]) * (1.0-diaf2[j]) + gapf1[0] ) * 0.5 * fpenalty; |
---|
1062 | // tmppenal -= ( (1.0-gapf1[0]) * ogcp2g[j] + gapf1[i-1] ) * 0.5 * fpenalty; |
---|
1063 | // fprintf( stderr, "0,0<-%d,%d, tmppenal 4 = %f\n", 0, j, tmppenal ); |
---|
1064 | currentw[j] += tmppenal; |
---|
1065 | } |
---|
1066 | } |
---|
1067 | #if OUTGAP0TRY |
---|
1068 | else |
---|
1069 | { |
---|
1070 | for( j=1; j<lgth2+1; j++ ) |
---|
1071 | currentw[j] -= offset * j / 2.0; |
---|
1072 | for( i=1; i<lgth1+1; i++ ) |
---|
1073 | initverticalw[i] -= offset * i / 2.0; |
---|
1074 | } |
---|
1075 | #endif |
---|
1076 | |
---|
1077 | m[0] = 0.0; // iranai |
---|
1078 | for( j=1; j<lgth2+1; ++j ) |
---|
1079 | { |
---|
1080 | // m[j] = currentw[j-1] + ogcp1[1]; |
---|
1081 | mp[j] = 0; |
---|
1082 | |
---|
1083 | tmppenal = 0.0; |
---|
1084 | // tmppenal = ( (1.0-gapz2[j])*(1.0-ogcp1g[1]+fgcp1g[1]) + gapz2[j]*(1.0-digf1[1]-diaf1[1]) ) * 0.5 * fpenalty; // mada |
---|
1085 | // tmppenal = ( (1.0-gapf2[j-1])*(1.0-ogcp1g[1]+fgcp1g[1]) + gapf2[j-1]*(1.0-digf1[1]-diaf1[1]) ) * 0.5 * fpenalty; // mada |
---|
1086 | // tmppenal = ( (1.0-gapf2[j-1])*(1.0-ogcp1g[i]+fgcp1g[i]) + gapf2[j-1]*(1.0-digf1[i]-diaf1[i]) ) * 0.5 * fpenalty; // mada |
---|
1087 | // tmppenal = 0.5 * fpenalty; |
---|
1088 | // tmppenal -= ( (1.0-0.0) * (1.0-0.0) + 0.0 ) * 0.5 * fpenalty; |
---|
1089 | // tmppenal -= ( (1.0-gapf2[-1]) * (1.0-diaf1[0]) + gapf2[-1] ) * 0.5 * fpenalty; |
---|
1090 | // if( tmppenal ) fprintf( stderr, "%c=%c, end j tmppenal=%f\n", seq1[0][0], seq2[0][j-1], tmppenal ); |
---|
1091 | m[j] = currentw[j-1] + tmppenal + fpenalty * 10000; |
---|
1092 | // m[j] = currentw[j-1] + ogcp1[1]; |
---|
1093 | } |
---|
1094 | if( lgth2 == 0 ) |
---|
1095 | lastverticalw[0] = 0.0; // Falign kara yobaretatoki kounarukanousei ari |
---|
1096 | else |
---|
1097 | lastverticalw[0] = currentw[lgth2-1]; |
---|
1098 | |
---|
1099 | if( outgap ) lasti = lgth1+1; else lasti = lgth1; |
---|
1100 | |
---|
1101 | #if XXXXXXX |
---|
1102 | fprintf( stderr, "currentw = \n" ); |
---|
1103 | for( i=0; i<lgth1+1; i++ ) |
---|
1104 | { |
---|
1105 | fprintf( stderr, "%5.2f ", currentw[i] ); |
---|
1106 | } |
---|
1107 | fprintf( stderr, "\n" ); |
---|
1108 | fprintf( stderr, "initverticalw = \n" ); |
---|
1109 | for( i=0; i<lgth2+1; i++ ) |
---|
1110 | { |
---|
1111 | fprintf( stderr, "%5.2f ", initverticalw[i] ); |
---|
1112 | } |
---|
1113 | fprintf( stderr, "\n" ); |
---|
1114 | fprintf( stderr, "fcgp\n" ); |
---|
1115 | for( i=0; i<lgth1; i++ ) |
---|
1116 | fprintf( stderr, "fgcp1[%d]=%f\n", i, ogcp1[i] ); |
---|
1117 | for( i=0; i<lgth2; i++ ) |
---|
1118 | fprintf( stderr, "fgcp2[%d]=%f\n", i, ogcp2[i] ); |
---|
1119 | #endif |
---|
1120 | |
---|
1121 | for( i=1; i<lasti; i++ ) |
---|
1122 | { |
---|
1123 | wtmp = previousw; |
---|
1124 | previousw = currentw; |
---|
1125 | currentw = wtmp; |
---|
1126 | |
---|
1127 | previousw[0] = initverticalw[i-1]; |
---|
1128 | |
---|
1129 | match_calc( currentw, cpmx1, cpmx2, i, lgth2, floatwork, intwork, 0 ); |
---|
1130 | #if XXXXXXX |
---|
1131 | fprintf( stderr, "\n" ); |
---|
1132 | fprintf( stderr, "i=%d\n", i ); |
---|
1133 | fprintf( stderr, "currentw = \n" ); |
---|
1134 | for( j=0; j<lgth2; j++ ) |
---|
1135 | { |
---|
1136 | fprintf( stderr, "%5.2f ", currentw[j] ); |
---|
1137 | } |
---|
1138 | fprintf( stderr, "\n" ); |
---|
1139 | #endif |
---|
1140 | if( localhom ) |
---|
1141 | { |
---|
1142 | // fprintf( stderr, "Calling imp_match_calc (o) lgth = %d, i = %d\n", lgth1, i ); |
---|
1143 | #if 0 |
---|
1144 | imp_match_out_veadH( currentw, i, lgth2 ); |
---|
1145 | #else |
---|
1146 | imp_match_out_veadH( currentw, i, lgth2 ); |
---|
1147 | #endif |
---|
1148 | } |
---|
1149 | #if XXXXXXX |
---|
1150 | fprintf( stderr, "\n" ); |
---|
1151 | fprintf( stderr, "i=%d\n", i ); |
---|
1152 | fprintf( stderr, "currentw = \n" ); |
---|
1153 | for( j=0; j<lgth2; j++ ) |
---|
1154 | { |
---|
1155 | fprintf( stderr, "%5.2f ", currentw[j] ); |
---|
1156 | } |
---|
1157 | fprintf( stderr, "\n" ); |
---|
1158 | #endif |
---|
1159 | currentw[0] = initverticalw[i]; |
---|
1160 | |
---|
1161 | |
---|
1162 | // mi = previousw[0] + ogcp2[1]; // machigai |
---|
1163 | tmppenal = 0.0; |
---|
1164 | // tmppenal = ( (1.0-gapz1[i])*(1.0-ogcp2g[1]+fgcp2g[1]) + gapz1[i]*(1.0-digf2[1]-diaf2[1]) ) * 0.5 * fpenalty; // mada |
---|
1165 | // tmppenal = ( (1.0-gapf1[i-1])*(1.0-ogcp2g[1]+fgcp2g[1]) + gapf1[i-1]*(1.0-digf2[1]-diaf2[1]) ) * 0.5 * fpenalty; // mada |
---|
1166 | // fprintf( stderr, "%c=%c, end i tmppenal=%f, ogcp2g[1]=%f\n", seq1[0][i-1], seq2[0][0], tmppenal, ogcp2g[1] ); |
---|
1167 | // mi = previousw[0] + tmppenal; |
---|
1168 | mi = previousw[0] + tmppenal + fpenalty * 10000; |
---|
1169 | |
---|
1170 | mpi = 0; |
---|
1171 | ijppt = ijp[i] + 1; |
---|
1172 | mjpt = m + 1; |
---|
1173 | prept = previousw; |
---|
1174 | curpt = currentw + 1; |
---|
1175 | mpjpt = mp + 1; |
---|
1176 | fgcp2pt = fgcp2; |
---|
1177 | ogcp2pt = ogcp2 + 1; |
---|
1178 | fgcp1va = fgcp1[i-1]; |
---|
1179 | ogcp1va = ogcp1[i]; |
---|
1180 | lastj = lgth2+1; |
---|
1181 | for( j=1; j<lastj; j++ ) |
---|
1182 | { |
---|
1183 | wm = *prept; |
---|
1184 | |
---|
1185 | if( gappat2[j][0].freq ) |
---|
1186 | { |
---|
1187 | g = diaf1[i] * gappat2[j][0].freq * fpenalty; |
---|
1188 | // if( seq1[0][i] == 'D' && seq2[0][j] == 'D' ) |
---|
1189 | // if( g ) fprintf( stderr, "match penal1=%f, %c-%c\n", g/fpenalty, seq1[0][i], seq2[0][j] ); |
---|
1190 | wm += g; |
---|
1191 | } |
---|
1192 | |
---|
1193 | if( gappat1[i][0].freq ) |
---|
1194 | { |
---|
1195 | g = diaf2[j] * gappat1[i][0].freq * fpenalty; |
---|
1196 | // if( seq1[0][i] == 'D' && seq2[0][j] == 'D' ) |
---|
1197 | // if( g ) fprintf( stderr, "match penal2=%f, %c-%c\n", g/fpenalty, seq1[0][i], seq2[0][j] ); |
---|
1198 | wm += g; |
---|
1199 | } |
---|
1200 | { |
---|
1201 | g = ( (gappat1[i][0].freq) * (gappat2[j][0].freq) ) * fpenalty; |
---|
1202 | // if( seq1[0][i] == 'D' && seq2[0][j] == 'D' ) |
---|
1203 | // if( g ) fprintf( stderr, "match penal3=%f, %c-%c\n", g/fpenalty, seq1[0][i], seq2[0][j] ); |
---|
1204 | wm += g; |
---|
1205 | } |
---|
1206 | if( 0 ) |
---|
1207 | { |
---|
1208 | maegap = ijp[i-1][j-1]; |
---|
1209 | // if( seq1[0][i] == 'Y' && seq2[0][j] == 'Y' ) |
---|
1210 | // fprintf( stderr, "i,j=%d,%d, maegap=%d\n", i, j, maegap ); |
---|
1211 | maegap = 0; |
---|
1212 | |
---|
1213 | if( maegap == 0 ) |
---|
1214 | { |
---|
1215 | g = ( countnocountmatchx( gappat1[i], gappat2[j], 0, 0, 0 ) ) * fpenalty; |
---|
1216 | // if( seq1[0][i] == 'D' && seq2[0][j] == 'D' ) |
---|
1217 | // fprintf( stderr, "kanwa0 %c-%c, i,j=%d,%d, g/fpenalty=%f, nocount=%f\n", seq1[0][i], seq2[0][j], i, j, g/fpenalty, countnocountmatchx( gappat1[i], gappat2[j], 0, -maegap, 1 ) ); |
---|
1218 | } |
---|
1219 | #if 0 // atta houga yoi hazu |
---|
1220 | else if( maegap < 0 ) // i jump |
---|
1221 | { |
---|
1222 | g = ( countnocountmatchx( gappat1[i], gappat2[j], 0, -maegap, 0 ) ) * fpenalty; |
---|
1223 | if( seq1[0][i] == 'Y' && seq2[0][j] == 'Y' ) |
---|
1224 | { |
---|
1225 | fprintf( stderr, "i-jumped, offset1=%d\n", maegap ); |
---|
1226 | fprintf( stderr, "kanwa1 %c-%c, i,j=%d,%d, g/fpenalty=%f, nocount=%f, nocount=%f\n", seq1[0][i], seq2[0][j], i, j, g/fpenalty, countnocountmatchx( gappat1[i], gappat2[j], 0, -maegap, 0 ) ); |
---|
1227 | } |
---|
1228 | } |
---|
1229 | else // j jump |
---|
1230 | { |
---|
1231 | g = ( countnocountmatchx( gappat1[i], gappat2[j], maegap, 0, 0 ) ) * fpenalty; |
---|
1232 | if( seq1[0][i] == 'Y' && seq2[0][j] == 'Y' ) |
---|
1233 | { |
---|
1234 | fprintf( stderr, "j-jumped, offset1=%d\n", maegap ); |
---|
1235 | fprintf( stderr, "kanwa2, %c-%c, i,j=%d,%d, g/fpenalty=%f, nocount=%f\n", seq1[0][i], seq2[0][j], i, j, g/fpenalty, countnocountmatchx( gappat1[i], gappat2[j], 0, -maegap, 1 ) ); |
---|
1236 | } |
---|
1237 | } |
---|
1238 | #endif |
---|
1239 | wm -= g; |
---|
1240 | } |
---|
1241 | |
---|
1242 | *ijppt = 0; |
---|
1243 | |
---|
1244 | #if 0 |
---|
1245 | fprintf( stderr, "%5.0f->", wm ); |
---|
1246 | #endif |
---|
1247 | #if 0 |
---|
1248 | fprintf( stderr, "%5.0f?", g ); |
---|
1249 | #endif |
---|
1250 | // tmppenal = fpenalty; |
---|
1251 | tmppenal = diaf2[j] * ( 1.0 - gapf1[i] ) * fpenalty; |
---|
1252 | if( gappat2[j][0].freq ) |
---|
1253 | { |
---|
1254 | tmppenal += ( gappat2[j][0].freq ) * ( 1.0 - gapf1[i] ) * fpenalty; |
---|
1255 | // tmppenal -= ( countnocountx( gappat1[i], diaf1[i], gappat2[j], j-mpi-1, 0 ) ) * fpenalty; |
---|
1256 | maegap = ijp[i-1][mpi]; |
---|
1257 | maegap = 0; |
---|
1258 | if( maegap == 0 ) |
---|
1259 | { |
---|
1260 | tmppenal -= ( countnocountx( gappat1[i], diaf1[i], gappat2[j], j-mpi-1, 0 ) ) * fpenalty; |
---|
1261 | } |
---|
1262 | #if 0 // attahouga yoi hazu |
---|
1263 | else if( maegap < 0 ) // i jump |
---|
1264 | { |
---|
1265 | maegap = -maegap; |
---|
1266 | tmppenal -= ( countnocountxx( gappat1[i], diaf1[i], gappat2[j], j-mpi-1+maegap, 0 ) ) * fpenalty; |
---|
1267 | } |
---|
1268 | else // j jump |
---|
1269 | { |
---|
1270 | tmppenal -= ( countnocountxx( gappat1[i], diaf1[i], gappat2[j], j-mpi-1, maegap ) ) * fpenalty; |
---|
1271 | } |
---|
1272 | #endif |
---|
1273 | } |
---|
1274 | if( (g=mi+tmppenal) > wm ) |
---|
1275 | { |
---|
1276 | // if( seq1[0][i] == 'A' && seq2[0][j] == 'A' ) fprintf( stderr, "jump i start=%f (i,j=%d,%d, *ijppt=%d, digf2[j]=%f, diaf2[j]=%f), %c-%c\n", g-mi, i, j, -(j-mpi), digf2[j], diaf2[j], seq1[0][i], seq2[0][j] ); |
---|
1277 | wm = g; |
---|
1278 | *ijppt = -( j - mpi ); |
---|
1279 | } |
---|
1280 | if( (g=*prept) >= mi ) |
---|
1281 | { |
---|
1282 | // fprintf( stderr, "jump i end=%f, %c-%c\n", g-*prept, seq1[0][i-1], seq2[0][j-1] ); |
---|
1283 | mi = g; |
---|
1284 | mpi = j-1; |
---|
1285 | } |
---|
1286 | else if( j != 1 ) |
---|
1287 | { |
---|
1288 | // mi += ( ogcp2g[j-0] + fgcp2g[j] ) * fpenalty * 0.5; |
---|
1289 | // fprintf( stderr, "%c%c/%c%c exp, og=%f,fg=%f\n", '=', '=', seq2[0][j-1], seq2[0][j], ogcp2g[j-0] * fpenalty*0.5, fgcp2g[j] * fpenalty*0.5 ); |
---|
1290 | } |
---|
1291 | #if USE_PENALTY_EX |
---|
1292 | mi += fpenalty_ex; |
---|
1293 | #endif |
---|
1294 | |
---|
1295 | #if 0 |
---|
1296 | fprintf( stderr, "%5.0f?", g ); |
---|
1297 | #endif |
---|
1298 | |
---|
1299 | // tmppenal = fpenalty; |
---|
1300 | tmppenal = diaf1[i] * ( 1.0 - gapf2[j] ) * fpenalty; |
---|
1301 | if( gappat1[i][0].freq ) |
---|
1302 | { |
---|
1303 | tmppenal += ( gappat1[i][0].freq ) * ( 1.0 - gapf2[j] ) * fpenalty; |
---|
1304 | // tmppenal -= ( countnocountx( gappat2[j], diaf2[j], gappat1[i], i-*mpjpt-1, 1 ) ) * fpenalty; |
---|
1305 | maegap = ijp[*mpjpt][j-1]; |
---|
1306 | if( maegap == 0 ) |
---|
1307 | { |
---|
1308 | tmppenal -= ( countnocountx( gappat2[j], diaf2[j], gappat1[i], i-*mpjpt-1, 1 ) ) * fpenalty; |
---|
1309 | } |
---|
1310 | #if 0 // attahouga yoi hazu |
---|
1311 | else if( maegap > 0 ) // j jump |
---|
1312 | { |
---|
1313 | tmppenal -= ( countnocountxx( gappat2[j], diaf2[j], gappat1[i], i-*mpjpt-1+maegap, 0 ) ) * fpenalty; |
---|
1314 | } |
---|
1315 | else // i jump |
---|
1316 | { |
---|
1317 | maegap = -maegap; |
---|
1318 | tmppenal -= ( countnocountxx( gappat2[j], diaf2[j], gappat1[i], i-*mpjpt-1, maegap ) ) * fpenalty; |
---|
1319 | } |
---|
1320 | #endif |
---|
1321 | } |
---|
1322 | if( (g=*mjpt+tmppenal) > wm ) |
---|
1323 | { |
---|
1324 | // if( seq1[0][i] == 'S' && seq2[0][j] == 'S' ) fprintf( stderr, "jump j start at %d, %d, g=%f, %c-%c\n", i, j, g-*mjpt, seq1[0][i], seq2[0][j] ); |
---|
1325 | wm = g; |
---|
1326 | *ijppt = +( i - *mpjpt ); |
---|
1327 | } |
---|
1328 | if( (g=*prept) >= *mjpt ) |
---|
1329 | { |
---|
1330 | // fprintf( stderr, "jump j end=%f, %c-%c\n", g-*prept, seq1[0][i-1], seq2[0][j-1] ); |
---|
1331 | *mjpt = g; |
---|
1332 | *mpjpt = i-1; |
---|
1333 | } |
---|
1334 | else if( i != 1 ) |
---|
1335 | { |
---|
1336 | // m[j] += ( ogcp1g[i-0] + fgcp1g[i] ) * fpenalty * 0.5; |
---|
1337 | // fprintf( stderr, "%c%c/%c%c exp, og=%f,fg=%f\n", seq1[0][i-1], seq1[0][i], '=', '=', ogcp1g[i-0] * fpenalty*0.5, fgcp1g[i] * fpenalty*0.5 ); |
---|
1338 | } |
---|
1339 | #if USE_PENALTY_EX |
---|
1340 | m[j] += fpenalty_ex; |
---|
1341 | #endif |
---|
1342 | |
---|
1343 | #if 0 |
---|
1344 | fprintf( stderr, "%5.0f ", wm ); |
---|
1345 | #endif |
---|
1346 | *curpt++ += wm; |
---|
1347 | ijppt++; |
---|
1348 | mjpt++; |
---|
1349 | prept++; |
---|
1350 | mpjpt++; |
---|
1351 | fgcp2pt++; |
---|
1352 | ogcp2pt++; |
---|
1353 | } |
---|
1354 | lastverticalw[i] = currentw[lgth2-1]; |
---|
1355 | } |
---|
1356 | |
---|
1357 | // fprintf( stderr, "wm = %f\n", wm ); |
---|
1358 | |
---|
1359 | #if OUTGAP0TRY |
---|
1360 | if( !outgap ) |
---|
1361 | { |
---|
1362 | for( j=1; j<lgth2+1; j++ ) |
---|
1363 | currentw[j] -= offset * ( lgth2 - j ) / 2.0; |
---|
1364 | for( i=1; i<lgth1+1; i++ ) |
---|
1365 | lastverticalw[i] -= offset * ( lgth1 - i / 2.0); |
---|
1366 | } |
---|
1367 | #endif |
---|
1368 | |
---|
1369 | /* |
---|
1370 | fprintf( stderr, "\n" ); |
---|
1371 | for( i=0; i<icyc; i++ ) fprintf( stderr,"%s\n", seq1[i] ); |
---|
1372 | fprintf( stderr, "#####\n" ); |
---|
1373 | for( j=0; j<jcyc; j++ ) fprintf( stderr,"%s\n", seq2[j] ); |
---|
1374 | fprintf( stderr, "====>" ); |
---|
1375 | for( i=0; i<icyc; i++ ) strcpy( mseq1[i], seq1[i] ); |
---|
1376 | for( j=0; j<jcyc; j++ ) strcpy( mseq2[j], seq2[j] ); |
---|
1377 | */ |
---|
1378 | if( localhom ) |
---|
1379 | { |
---|
1380 | Atracking_localhom( impmatch, currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc ); |
---|
1381 | } |
---|
1382 | else |
---|
1383 | Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc ); |
---|
1384 | |
---|
1385 | // fprintf( stderr, "### impmatch = %f\n", *impmatch ); |
---|
1386 | |
---|
1387 | resultlen = strlen( mseq1[0] ); |
---|
1388 | if( alloclen < resultlen || resultlen > N ) |
---|
1389 | { |
---|
1390 | fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N ); |
---|
1391 | ErrorExit( "LENGTH OVER!\n" ); |
---|
1392 | } |
---|
1393 | |
---|
1394 | |
---|
1395 | for( i=0; i<icyc; i++ ) strcpy( seq1[i], mseq1[i] ); |
---|
1396 | for( j=0; j<jcyc; j++ ) strcpy( seq2[j], mseq2[j] ); |
---|
1397 | /* |
---|
1398 | fprintf( stderr, "\n" ); |
---|
1399 | for( i=0; i<icyc; i++ ) fprintf( stderr, "%s\n", mseq1[i] ); |
---|
1400 | fprintf( stderr, "#####\n" ); |
---|
1401 | for( j=0; j<jcyc; j++ ) fprintf( stderr, "%s\n", mseq2[j] ); |
---|
1402 | */ |
---|
1403 | |
---|
1404 | fprintf( stderr, "wm = %f\n", wm ); |
---|
1405 | |
---|
1406 | for( i=0; i<lgth1+1; i++ ) |
---|
1407 | { |
---|
1408 | free( gappat1[i] ); |
---|
1409 | gappat1[i] = NULL; |
---|
1410 | } |
---|
1411 | for( i=0; i<lgth2+1; i++ ) |
---|
1412 | { |
---|
1413 | free( gappat2[i] ); |
---|
1414 | gappat2[i] = NULL; |
---|
1415 | } |
---|
1416 | |
---|
1417 | return( wm ); |
---|
1418 | } |
---|
1419 | |
---|