1 | #include "mltaln.h" |
---|
2 | #include "dp.h" |
---|
3 | |
---|
4 | #define DEBUG 0 |
---|
5 | |
---|
6 | static void match_calc( float *match, float **cpmx1, float **cpmx2, int i1, int lgth2, float **floatwork, int **intwork, int initialize ) |
---|
7 | { |
---|
8 | int j, k, l; |
---|
9 | float scarr[26]; |
---|
10 | float **cpmxpd = floatwork; |
---|
11 | int **cpmxpdn = intwork; |
---|
12 | int count = 0; |
---|
13 | |
---|
14 | if( initialize ) |
---|
15 | { |
---|
16 | for( j=0; j<lgth2; j++ ) |
---|
17 | { |
---|
18 | count = 0; |
---|
19 | for( l=0; l<26; l++ ) |
---|
20 | { |
---|
21 | if( cpmx2[l][j] ) |
---|
22 | { |
---|
23 | cpmxpd[count][j] = cpmx2[l][j]; |
---|
24 | cpmxpdn[count][j] = l; |
---|
25 | count++; |
---|
26 | } |
---|
27 | } |
---|
28 | cpmxpdn[count][j] = -1; |
---|
29 | } |
---|
30 | } |
---|
31 | |
---|
32 | for( l=0; l<26; l++ ) |
---|
33 | { |
---|
34 | scarr[l] = 0.0; |
---|
35 | for( k=0; k<26; k++ ) |
---|
36 | scarr[l] += n_dis[k][l] * cpmx1[k][i1]; |
---|
37 | } |
---|
38 | for( j=0; j<lgth2; j++ ) |
---|
39 | { |
---|
40 | match[j] = 0; |
---|
41 | for( k=0; cpmxpdn[k][j] > -1; k++ ) |
---|
42 | match[j] += scarr[cpmxpdn[k][j]] * cpmxpd[k][j]; |
---|
43 | } |
---|
44 | } |
---|
45 | |
---|
46 | static float Atracking( float *lasthorizontalw, float *lastverticalw, |
---|
47 | char **seq1, char **seq2, |
---|
48 | char **mseq1, char **mseq2, |
---|
49 | float **cpmx1, float **cpmx2, |
---|
50 | int **ijp, int icyc, int jcyc ) |
---|
51 | { |
---|
52 | int i, j, k, l, iin, jin, ifi, jfi, lgth1, lgth2; |
---|
53 | // char gap[] = "-"; |
---|
54 | char *gap; |
---|
55 | float wm; |
---|
56 | gap = newgapstr; |
---|
57 | lgth1 = strlen( seq1[0] ); |
---|
58 | lgth2 = strlen( seq2[0] ); |
---|
59 | |
---|
60 | #if DEBUG |
---|
61 | for( i=0; i<lgth1; i++ ) |
---|
62 | { |
---|
63 | fprintf( stderr, "lastverticalw[%d] = %f\n", i, lastverticalw[i] ); |
---|
64 | } |
---|
65 | #endif |
---|
66 | |
---|
67 | if( outgap == 1 ) |
---|
68 | ; |
---|
69 | else |
---|
70 | { |
---|
71 | wm = lastverticalw[0]; |
---|
72 | for( i=0; i<lgth1; i++ ) |
---|
73 | { |
---|
74 | if( lastverticalw[i] >= wm ) |
---|
75 | { |
---|
76 | wm = lastverticalw[i]; |
---|
77 | iin = i; jin = lgth2-1; |
---|
78 | ijp[lgth1][lgth2] = +( lgth1 - i ); |
---|
79 | } |
---|
80 | } |
---|
81 | for( j=0; j<lgth2; j++ ) |
---|
82 | { |
---|
83 | if( lasthorizontalw[j] >= wm ) |
---|
84 | { |
---|
85 | wm = lasthorizontalw[j]; |
---|
86 | iin = lgth1-1; jin = j; |
---|
87 | ijp[lgth1][lgth2] = -( lgth2 - j ); |
---|
88 | } |
---|
89 | } |
---|
90 | } |
---|
91 | |
---|
92 | for( i=0; i<lgth1+1; i++ ) |
---|
93 | { |
---|
94 | ijp[i][0] = i + 1; |
---|
95 | } |
---|
96 | for( j=0; j<lgth2+1; j++ ) |
---|
97 | { |
---|
98 | ijp[0][j] = -( j + 1 ); |
---|
99 | } |
---|
100 | |
---|
101 | for( i=0; i<icyc; i++ ) |
---|
102 | { |
---|
103 | mseq1[i] += lgth1+lgth2; |
---|
104 | *mseq1[i] = 0; |
---|
105 | } |
---|
106 | for( j=0; j<jcyc; j++ ) |
---|
107 | { |
---|
108 | mseq2[j] += lgth1+lgth2; |
---|
109 | *mseq2[j] = 0; |
---|
110 | } |
---|
111 | iin = lgth1; jin = lgth2; |
---|
112 | for( k=0; k<=lgth1+lgth2; k++ ) |
---|
113 | { |
---|
114 | if( ijp[iin][jin] < 0 ) |
---|
115 | { |
---|
116 | ifi = iin-1; jfi = jin+ijp[iin][jin]; |
---|
117 | } |
---|
118 | else if( ijp[iin][jin] > 0 ) |
---|
119 | { |
---|
120 | ifi = iin-ijp[iin][jin]; jfi = jin-1; |
---|
121 | } |
---|
122 | else |
---|
123 | { |
---|
124 | ifi = iin-1; jfi = jin-1; |
---|
125 | } |
---|
126 | l = iin - ifi; |
---|
127 | while( --l ) |
---|
128 | { |
---|
129 | for( i=0; i<icyc; i++ ) |
---|
130 | *--mseq1[i] = seq1[i][ifi+l]; |
---|
131 | for( j=0; j<jcyc; j++ ) |
---|
132 | *--mseq2[j] = *gap; |
---|
133 | k++; |
---|
134 | } |
---|
135 | l= jin - jfi; |
---|
136 | while( --l ) |
---|
137 | { |
---|
138 | for( i=0; i<icyc; i++ ) |
---|
139 | *--mseq1[i] = *gap; |
---|
140 | for( j=0; j<jcyc; j++ ) |
---|
141 | *--mseq2[j] = seq2[j][jfi+l]; |
---|
142 | k++; |
---|
143 | } |
---|
144 | if( iin <= 0 || jin <= 0 ) break; |
---|
145 | for( i=0; i<icyc; i++ ) |
---|
146 | *--mseq1[i] = seq1[i][ifi]; |
---|
147 | for( j=0; j<jcyc; j++ ) |
---|
148 | *--mseq2[j] = seq2[j][jfi]; |
---|
149 | k++; |
---|
150 | iin = ifi; jin = jfi; |
---|
151 | } |
---|
152 | return( 0.0 ); |
---|
153 | } |
---|
154 | |
---|
155 | |
---|
156 | float Aalign( char **seq1, char **seq2, double *eff1, double *eff2, int icyc, int jcyc, int alloclen ) |
---|
157 | /* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */ |
---|
158 | { |
---|
159 | register int i, j; |
---|
160 | int lasti; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */ |
---|
161 | int lgth1, lgth2; |
---|
162 | int resultlen; |
---|
163 | float wm = 0.0; /* int ?????? */ |
---|
164 | float g; |
---|
165 | float x; |
---|
166 | static TLS float mi, *m; |
---|
167 | static TLS int **ijp; |
---|
168 | static TLS int mpi, *mp; |
---|
169 | static TLS float *currentw; |
---|
170 | static TLS float *previousw; |
---|
171 | static TLS float *match; |
---|
172 | static TLS float *initverticalw; /* kufuu sureba iranai */ |
---|
173 | static TLS float *lastverticalw; /* kufuu sureba iranai */ |
---|
174 | static TLS char **mseq1; |
---|
175 | static TLS char **mseq2; |
---|
176 | static TLS char **mseq; |
---|
177 | static TLS float **cpmx1; |
---|
178 | static TLS float **cpmx2; |
---|
179 | static TLS int **intwork; |
---|
180 | static TLS float **floatwork; |
---|
181 | static TLS int orlgth1 = 0, orlgth2 = 0; |
---|
182 | |
---|
183 | #if DEBUG |
---|
184 | fprintf( stderr, "eff in SA+++align\n" ); |
---|
185 | for( i=0; i<icyc; i++ ) fprintf( stderr, "eff1[%d] = %f\n", i, eff1[i] ); |
---|
186 | #endif |
---|
187 | if( orlgth1 == 0 ) |
---|
188 | { |
---|
189 | mseq1 = AllocateCharMtx( njob, 1 ); |
---|
190 | mseq2 = AllocateCharMtx( njob, 1 ); /* by J. Thompson */ |
---|
191 | } |
---|
192 | |
---|
193 | lgth1 = strlen( seq1[0] ); |
---|
194 | lgth2 = strlen( seq2[0] ); |
---|
195 | |
---|
196 | if( lgth1 > orlgth1 || lgth2 > orlgth2 ) |
---|
197 | { |
---|
198 | int ll1, ll2; |
---|
199 | |
---|
200 | if( orlgth1 > 0 && orlgth2 > 0 ) |
---|
201 | { |
---|
202 | FreeFloatVec( currentw ); |
---|
203 | FreeFloatVec( previousw ); |
---|
204 | FreeFloatVec( match ); |
---|
205 | FreeFloatVec( initverticalw ); |
---|
206 | FreeFloatVec( lastverticalw ); |
---|
207 | |
---|
208 | FreeFloatVec( m ); |
---|
209 | FreeIntVec( mp ); |
---|
210 | |
---|
211 | FreeCharMtx( mseq ); |
---|
212 | |
---|
213 | FreeFloatMtx( cpmx1 ); |
---|
214 | FreeFloatMtx( cpmx2 ); |
---|
215 | |
---|
216 | FreeFloatMtx( floatwork ); |
---|
217 | FreeIntMtx( intwork ); |
---|
218 | } |
---|
219 | |
---|
220 | ll1 = MAX( (int)(1.1*lgth1), orlgth1 ) + 100; |
---|
221 | ll2 = MAX( (int)(1.1*lgth2), orlgth2 ) + 100; |
---|
222 | |
---|
223 | fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 ); |
---|
224 | |
---|
225 | currentw = AllocateFloatVec( ll2+2 ); |
---|
226 | previousw = AllocateFloatVec( ll2+2 ); |
---|
227 | match = AllocateFloatVec( ll2+2 ); |
---|
228 | |
---|
229 | initverticalw = AllocateFloatVec( ll1+2 ); |
---|
230 | lastverticalw = AllocateFloatVec( ll1+2 ); |
---|
231 | |
---|
232 | m = AllocateFloatVec( ll2+2 ); |
---|
233 | mp = AllocateIntVec( ll2+2 ); |
---|
234 | |
---|
235 | mseq = AllocateCharMtx( njob, ll1+ll2 ); |
---|
236 | |
---|
237 | cpmx1 = AllocateFloatMtx( 26, ll1+2 ); |
---|
238 | cpmx2 = AllocateFloatMtx( 26, ll2+2 ); |
---|
239 | |
---|
240 | floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 ); |
---|
241 | intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 ); |
---|
242 | |
---|
243 | fprintf( stderr, "succeeded\n" ); |
---|
244 | |
---|
245 | orlgth1 = ll1; |
---|
246 | orlgth2 = ll2; |
---|
247 | } |
---|
248 | |
---|
249 | for( i=0; i<icyc; i++ ) mseq1[i] = mseq[i]; |
---|
250 | for( j=0; j<jcyc; j++ ) mseq2[j] = mseq[icyc+j]; |
---|
251 | |
---|
252 | |
---|
253 | if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 ) |
---|
254 | { |
---|
255 | int ll1, ll2; |
---|
256 | |
---|
257 | if( commonAlloc1 && commonAlloc2 ) |
---|
258 | { |
---|
259 | FreeIntMtx( commonIP ); |
---|
260 | } |
---|
261 | |
---|
262 | ll1 = MAX( orlgth1, commonAlloc1 ); |
---|
263 | ll2 = MAX( orlgth2, commonAlloc2 ); |
---|
264 | |
---|
265 | fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 ); |
---|
266 | |
---|
267 | commonIP = AllocateIntMtx( ll1+10, ll2+10 ); |
---|
268 | |
---|
269 | fprintf( stderr, "succeeded\n\n" ); |
---|
270 | |
---|
271 | commonAlloc1 = ll1; |
---|
272 | commonAlloc2 = ll2; |
---|
273 | } |
---|
274 | ijp = commonIP; |
---|
275 | |
---|
276 | cpmx_calc( seq1, cpmx1, eff1, strlen( seq1[0] ), icyc ); |
---|
277 | cpmx_calc( seq2, cpmx2, eff2, strlen( seq2[0] ), jcyc ); |
---|
278 | |
---|
279 | match_calc( initverticalw, cpmx2, cpmx1, 0, lgth1, floatwork, intwork, 1 ); |
---|
280 | match_calc( currentw, cpmx1, cpmx2, 0, lgth2, floatwork, intwork, 1 ); |
---|
281 | |
---|
282 | if( outgap == 1 ) |
---|
283 | { |
---|
284 | for( i=1; i<lgth1+1; i++ ) |
---|
285 | { |
---|
286 | initverticalw[i] += penalty * 0.5; |
---|
287 | } |
---|
288 | for( j=1; j<lgth2+1; j++ ) |
---|
289 | { |
---|
290 | currentw[j] += penalty * 0.5; |
---|
291 | } |
---|
292 | } |
---|
293 | |
---|
294 | for( j=0; j<lgth2+1; ++j ) |
---|
295 | { |
---|
296 | m[j] = currentw[j-1] + penalty * 0.5; mp[j] = 0; |
---|
297 | } |
---|
298 | |
---|
299 | lastverticalw[0] = currentw[lgth2-1]; |
---|
300 | |
---|
301 | if( outgap ) lasti = lgth1+1; else lasti = lgth1; |
---|
302 | |
---|
303 | for( i=1; i<lasti; i++ ) |
---|
304 | { |
---|
305 | |
---|
306 | floatncpy( previousw, currentw, lgth2+1 ); |
---|
307 | previousw[0] = initverticalw[i-1]; |
---|
308 | |
---|
309 | match_calc( currentw, cpmx1, cpmx2, i, lgth2, floatwork, intwork, 0 ); |
---|
310 | currentw[0] = initverticalw[i]; |
---|
311 | |
---|
312 | mi = previousw[0] + penalty * 0.5; mpi = 0; |
---|
313 | for( j=1; j<lgth2+1; j++ ) |
---|
314 | { |
---|
315 | wm = previousw[j-1]; |
---|
316 | ijp[i][j] = 0; |
---|
317 | |
---|
318 | g = penalty * 0.5; |
---|
319 | x = mi + g; |
---|
320 | if( x > wm ) |
---|
321 | { |
---|
322 | wm = x; |
---|
323 | ijp[i][j] = -( j - mpi ); |
---|
324 | } |
---|
325 | g = penalty * 0.5; |
---|
326 | x = previousw[j-1] + g; |
---|
327 | if( mi <= x ) |
---|
328 | { |
---|
329 | mi = x; |
---|
330 | mpi = j-1; |
---|
331 | } |
---|
332 | |
---|
333 | g = penalty * 0.5; |
---|
334 | x = m[j] + g; |
---|
335 | if( x > wm ) |
---|
336 | { |
---|
337 | wm = x; |
---|
338 | ijp[i][j] = +( i - mp[j] ); |
---|
339 | } |
---|
340 | g = penalty * 0.5; |
---|
341 | x = previousw[j-1] + g; |
---|
342 | if( m[j] <= x ) |
---|
343 | { |
---|
344 | m[j] = x; |
---|
345 | mp[j] = i-1; |
---|
346 | } |
---|
347 | currentw[j] += wm; |
---|
348 | } |
---|
349 | lastverticalw[i] = currentw[lgth2-1]; |
---|
350 | } |
---|
351 | /* |
---|
352 | fprintf( stderr, "\n" ); |
---|
353 | for( i=0; i<icyc; i++ ) fprintf( stderr,"%s\n", seq1[i] ); |
---|
354 | fprintf( stderr, "#####\n" ); |
---|
355 | for( j=0; j<jcyc; j++ ) fprintf( stderr,"%s\n", seq2[j] ); |
---|
356 | fprintf( stderr, "====>" ); |
---|
357 | for( i=0; i<icyc; i++ ) strcpy( mseq1[i], seq1[i] ); |
---|
358 | for( j=0; j<jcyc; j++ ) strcpy( mseq2[j], seq2[j] ); |
---|
359 | */ |
---|
360 | Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc ); |
---|
361 | |
---|
362 | resultlen = strlen( mseq1[0] ); |
---|
363 | if( alloclen < resultlen || resultlen > N ) |
---|
364 | { |
---|
365 | fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N ); |
---|
366 | ErrorExit( "LENGTH OVER!\n" ); |
---|
367 | } |
---|
368 | |
---|
369 | for( i=0; i<icyc; i++ ) strcpy( seq1[i], mseq1[i] ); |
---|
370 | for( j=0; j<jcyc; j++ ) strcpy( seq2[j], mseq2[j] ); |
---|
371 | /* |
---|
372 | fprintf( stderr, "\n" ); |
---|
373 | for( i=0; i<icyc; i++ ) fprintf( stderr, "%s\n", mseq1[i] ); |
---|
374 | fprintf( stderr, "#####\n" ); |
---|
375 | for( j=0; j<jcyc; j++ ) fprintf( stderr, "%s\n", mseq2[j] ); |
---|
376 | */ |
---|
377 | return( wm ); |
---|
378 | } |
---|