8
#define USE_PENALTY_EX 0
9
#define FASTMATCHCALC 1
12
static float **impmtx = NULL;
13
#if 1 // tditeration to naiveQscore_imp de tsukawareru.
14
float imp_match_out_scQ( int i1, int j1 )
16
// fprintf( stderr, "imp+match = %f\n", impmtx[i1][j1] * fastathreshold );
17
// fprintf( stderr, "val = %f\n", impmtx[i1][j1] );
18
return( impmtx[i1][j1] );
22
static void imp_match_out_veadQ( float *imp, int i1, int lgth2 )
25
float *pt = impmtx[i1];
30
float *pt = impmtx[i1];
31
for( j=0; j<lgth2; j++ )
35
static void imp_match_out_vead_tateQ( float *imp, int j1, int lgth1 )
38
for( i=0; i<lgth1; i++ )
39
*imp++ += impmtx[i][j1];
42
#if 1 // tbfast.c kara yobareru.
43
void imp_match_init_strictQ( float *imp, int clus1, int clus2, int lgth1, int lgth2, char **seq1, char **seq2, double *eff1, double *eff2, LocalHom ***localhom, int forscore )
45
int i, j, k1, k2, tmpint, start1, start2, end1, end2;
46
static int impalloclen = 0;
50
static char *nocount1 = NULL;
51
static char *nocount2 = NULL;
54
if( impalloclen < lgth1 + 2 || impalloclen < lgth2 + 2 )
56
if( impmtx ) FreeFloatMtx( impmtx );
57
if( nocount1 ) free( nocount1 );
58
if( nocount2 ) free( nocount2 );
59
impalloclen = MAX( lgth1, lgth2 ) + 2;
60
impmtx = AllocateFloatMtx( impalloclen, impalloclen );
61
nocount1 = AllocateCharVec( impalloclen );
62
nocount2 = AllocateCharVec( impalloclen );
65
for( i=0; i<lgth1; i++ )
67
for( j=0; j<clus1; j++ )
68
if( seq1[j][i] == '-' ) break;
69
if( j != clus1 ) nocount1[i] = 1;
72
for( i=0; i<lgth2; i++ )
74
for( j=0; j<clus2; j++ )
75
if( seq2[j][i] == '-' ) break;
76
if( j != clus2 ) nocount2[i] = 1;
81
fprintf( stderr, "nocount2 =\n" );
82
for( i = 0; i<impalloclen; i++ )
84
fprintf( stderr, "nocount2[%d] = %d (%c)\n", i, nocount2[i], seq2[0][i] );
91
fprintf( stderr, "eff1 in _init_strict = \n" );
92
for( i=0; i<clus1; i++ )
93
fprintf( stderr, "eff1[] = %f\n", eff1[i] );
94
for( i=0; i<clus2; i++ )
95
fprintf( stderr, "eff2[] = %f\n", eff2[i] );
98
for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ )
100
effijx = fastathreshold;
101
for( i=0; i<clus1; i++ )
103
for( j=0; j<clus2; j++ )
105
effij = (float)( eff1[i] * eff2[j] * effijx );
106
tmpptr = localhom[i][j];
109
// fprintf( stderr, "start1 = %d\n", tmpptr->start1 );
110
// fprintf( stderr, "end1 = %d\n", tmpptr->end1 );
111
// fprintf( stderr, "i = %d, seq1 = \n%s\n", i, seq1[i] );
112
// fprintf( stderr, "j = %d, seq2 = \n%s\n", j, seq2[j] );
117
if( *pt++ != '-' ) tmpint++;
118
if( tmpint == tmpptr->start1 ) break;
120
start1 = pt - seq1[i] - 1;
122
if( tmpptr->start1 == tmpptr->end1 ) end1 = start1;
128
// fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, tmpptr->end1, pt-seq1[i] );
129
if( tmpint == tmpptr->end1 ) break;
130
if( *pt++ != '-' ) tmpint++;
132
end1 = pt - seq1[i] - 0;
136
// fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, tmpptr->end1, pt-seq1[i] );
137
if( *pt++ != '-' ) tmpint++;
138
if( tmpint == tmpptr->end1 ) break;
140
end1 = pt - seq1[i] - 1;
148
if( *pt++ != '-' ) tmpint++;
149
if( tmpint == tmpptr->start2 ) break;
151
start2 = pt - seq2[j] - 1;
152
if( tmpptr->start2 == tmpptr->end2 ) end2 = start2;
158
if( tmpint == tmpptr->end2 ) break;
159
if( *pt++ != '-' ) tmpint++;
161
end2 = pt - seq2[j] - 0;
165
if( *pt++ != '-' ) tmpint++;
166
if( tmpint == tmpptr->end2 ) break;
168
end2 = pt - seq2[j] - 1;
171
// 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] );
172
// fprintf( stderr, "step 0\n" );
173
if( end1 - start1 != end2 - start2 )
175
// fprintf( stderr, "CHUUI!!, start1 = %d, end1 = %d, start2 = %d, end2 = %d\n", start1, end1, start2, end2 );
179
k1 = start1; k2 = start2;
182
while( *pt1 && *pt2 )
184
if( *pt1 != '-' && *pt2 != '-' )
186
// �Ťߤ���Ťˤ����ʤ��褦�����դ��Ʋ�������
187
// impmtx[k1][k2] += tmpptr->wimportance * fastathreshold;
188
// impmtx[k1][k2] += tmpptr->importance * effij;
189
impmtx[k1][k2] += tmpptr->fimportance * effij;
190
// fprintf( stderr, "#### impmtx[k1][k2] = %f, tmpptr->fimportance=%f, effij=%f\n", impmtx[k1][k2], tmpptr->fimportance, effij );
191
// fprintf( stderr, "mark, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 );
192
// fprintf( stderr, "%d (%c) - %d (%c) - %f\n", k1, *pt1, k2, *pt2, tmpptr->fimportance * effij );
196
else if( *pt1 != '-' && *pt2 == '-' )
198
// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 );
201
else if( *pt1 == '-' && *pt2 != '-' )
203
// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 );
206
else if( *pt1 == '-' && *pt2 == '-' )
208
// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 );
212
if( k1 > end1 || k2 > end2 ) break;
215
while( k1 <= end1 && k2 <= end2 )
217
fprintf( stderr, "k1,k2=%d,%d - ", k1, k2 );
218
if( !nocount1[k1] && !nocount2[k2] )
220
impmtx[k1][k2] += tmpptr->wimportance * eff1[i] * eff2[j] * fastathreshold;
221
fprintf( stderr, "marked\n" );
224
fprintf( stderr, "no count\n" );
228
tmpptr = tmpptr->next;
233
if( clus1 == 1 && clus2 == 6 )
235
fprintf( stderr, "\n" );
236
fprintf( stderr, "seq1[0] = %s\n", seq1[0] );
237
fprintf( stderr, "seq2[0] = %s\n", seq2[0] );
238
fprintf( stderr, "impmtx = \n" );
239
for( k2=0; k2<lgth2; k2++ )
240
fprintf( stderr, "%6.3f ", (double)k2 );
241
fprintf( stderr, "\n" );
242
for( k1=0; k1<lgth1; k1++ )
244
fprintf( stderr, "%d ", k1 );
245
for( k2=0; k2<3; k2++ )
246
fprintf( stderr, "%2.1f ", impmtx[k1][k2] );
247
fprintf( stderr, "\n" );
256
static void match_calc( float *match, float **cpmx1, float **cpmx2, int i1, int lgth2, float **floatwork, int **intwork, int initialize )
261
float **cpmxpd = floatwork;
262
int **cpmxpdn = intwork;
263
float *matchpt, *cpmxpdpt, **cpmxpdptpt;
264
int *cpmxpdnpt, **cpmxpdnptpt;
268
for( j=0; j<lgth2; j++ )
271
for( l=0; l<26; l++ )
275
cpmxpd[j][count] = cpmx2[l][j];
276
cpmxpdn[j][count] = l;
280
cpmxpdn[j][count] = -1;
285
for( l=0; l<26; l++ )
288
for( j=0; j<26; j++ )
289
scarr[l] += n_dis[j][l] * cpmx1[j][i1];
292
cpmxpdnptpt = cpmxpdn;
297
cpmxpdnpt = *cpmxpdnptpt++;
298
cpmxpdpt = *cpmxpdptpt++;
299
while( *cpmxpdnpt>-1 )
300
*matchpt += scarr[*cpmxpdnpt++] * *cpmxpdpt++;
307
float **cpmxpd = floatwork;
308
int **cpmxpdn = intwork;
313
for( j=0; j<lgth2; j++ )
316
for( l=0; l<26; l++ )
320
cpmxpd[count][j] = cpmx2[l][j];
321
cpmxpdn[count][j] = l;
325
cpmxpdn[count][j] = -1;
328
for( l=0; l<26; l++ )
331
for( k=0; k<26; k++ )
332
scarr[l] += n_dis[k][l] * cpmx1[k][i1];
334
for( j=0; j<lgth2; j++ )
337
for( k=0; cpmxpdn[k][j]>-1; k++ )
338
match[j] += scarr[cpmxpdn[k][j]] * cpmxpd[k][j];
343
static void Atracking_localhom( float *impwmpt, float *lasthorizontalw, float *lastverticalw,
344
char **seq1, char **seq2,
345
char **mseq1, char **mseq2,
346
float **cpmx1, float **cpmx2,
347
short **ijp, int icyc, int jcyc )
349
int i, j, l, iin, jin, ifi, jfi, lgth1, lgth2, k;
351
char *gaptable1, *gt1bk;
352
char *gaptable2, *gt2bk;
353
lgth1 = strlen( seq1[0] );
354
lgth2 = strlen( seq2[0] );
355
gt1bk = AllocateCharVec( lgth1+lgth2+1 );
356
gt2bk = AllocateCharVec( lgth1+lgth2+1 );
359
for( i=0; i<lgth1; i++ )
361
fprintf( stderr, "lastverticalw[%d] = %f\n", i, lastverticalw[i] );
369
wm = lastverticalw[0];
370
for( i=0; i<lgth1; i++ )
372
if( lastverticalw[i] >= wm )
374
wm = lastverticalw[i];
375
iin = i; jin = lgth2-1;
376
ijp[lgth1][lgth2] = +( lgth1 - i );
379
for( j=0; j<lgth2; j++ )
381
if( lasthorizontalw[j] >= wm )
383
wm = lasthorizontalw[j];
384
iin = lgth1-1; jin = j;
385
ijp[lgth1][lgth2] = -( lgth2 - j );
390
for( i=0; i<lgth1+1; i++ )
394
for( j=0; j<lgth2+1; j++ )
396
ijp[0][j] = -( j + 1 );
399
gaptable1 = gt1bk + lgth1+lgth2;
401
gaptable2 = gt2bk + lgth1+lgth2;
404
iin = lgth1; jin = lgth2;
406
for( k=0; k<=lgth1+lgth2; k++ )
408
if( ijp[iin][jin] < 0 )
410
ifi = iin-1; jfi = jin+ijp[iin][jin];
412
else if( ijp[iin][jin] > 0 )
414
ifi = iin-ijp[iin][jin]; jfi = jin-1;
418
ifi = iin-1; jfi = jin-1;
434
if( iin == lgth1 || jin == lgth2 )
438
*impwmpt += imp_match_out_scQ( iin, jin );
440
// fprintf( stderr, "impwm = %f (iin=%d, jin=%d) seq1=%c, seq2=%c\n", *impwmpt, iin, jin, seq1[0][iin], seq2[0][jin] );
442
if( iin <= 0 || jin <= 0 ) break;
446
iin = ifi; jin = jfi;
449
for( i=0; i<icyc; i++ ) gapireru( mseq1[i], seq1[i], gaptable1 );
450
for( j=0; j<jcyc; j++ ) gapireru( mseq2[j], seq2[j], gaptable2 );
456
static float Atracking( float *lasthorizontalw, float *lastverticalw,
457
char **seq1, char **seq2,
458
char **mseq1, char **mseq2,
459
float **cpmx1, float **cpmx2,
460
short **ijp, int icyc, int jcyc )
462
int i, j, l, iin, jin, ifi, jfi, lgth1, lgth2, k;
464
char *gaptable1, *gt1bk;
465
char *gaptable2, *gt2bk;
466
lgth1 = strlen( seq1[0] );
467
lgth2 = strlen( seq2[0] );
469
gt1bk = AllocateCharVec( lgth1+lgth2+1 );
470
gt2bk = AllocateCharVec( lgth1+lgth2+1 );
473
for( i=0; i<lgth1; i++ )
475
fprintf( stderr, "lastverticalw[%d] = %f\n", i, lastverticalw[i] );
483
wm = lastverticalw[0];
484
for( i=0; i<lgth1; i++ )
486
if( lastverticalw[i] >= wm )
488
wm = lastverticalw[i];
489
iin = i; jin = lgth2-1;
490
ijp[lgth1][lgth2] = +( lgth1 - i );
493
for( j=0; j<lgth2; j++ )
495
if( lasthorizontalw[j] >= wm )
497
wm = lasthorizontalw[j];
498
iin = lgth1-1; jin = j;
499
ijp[lgth1][lgth2] = -( lgth2 - j );
504
for( i=0; i<lgth1+1; i++ )
508
for( j=0; j<lgth2+1; j++ )
510
ijp[0][j] = -( j + 1 );
513
gaptable1 = gt1bk + lgth1+lgth2;
515
gaptable2 = gt2bk + lgth1+lgth2;
518
iin = lgth1; jin = lgth2;
519
for( k=0; k<=lgth1+lgth2; k++ )
521
if( ijp[iin][jin] < 0 )
523
ifi = iin-1; jfi = jin+ijp[iin][jin];
525
else if( ijp[iin][jin] > 0 )
527
ifi = iin-ijp[iin][jin]; jfi = jin-1;
531
ifi = iin-1; jfi = jin-1;
547
if( iin <= 0 || jin <= 0 ) break;
551
iin = ifi; jin = jfi;
554
for( i=0; i<icyc; i++ ) gapireru( mseq1[i], seq1[i], gaptable1 );
555
for( j=0; j<jcyc; j++ ) gapireru( mseq2[j], seq2[j], gaptable2 );
563
float Q__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 )
564
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
568
int lasti, lastj; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */
571
float wm = 0.0; /* int ?????? */
573
float *currentw, *previousw;
574
// float fpenalty = (float)penalty;
576
float fpenalty_ex = (float)penalty_ex;
581
float *mjpt, *prept, *curpt;
587
static float *w1, *w2;
589
static float *initverticalw; /* kufuu sureba iranai */
590
static float *lastverticalw; /* kufuu sureba iranai */
602
static float *ogcp1g;
603
static float *ogcp2g;
604
static float *fgcp1g;
605
static float *fgcp2g;
606
static float *og_h_dg_n1_p;
607
static float *og_h_dg_n2_p;
608
static float *fg_h_dg_n1_p;
609
static float *fg_h_dg_n2_p;
610
static float *og_t_fg_h_dg_n1_p;
611
static float *og_t_fg_h_dg_n2_p;
612
static float *fg_t_og_h_dg_n1_p;
613
static float *fg_t_og_h_dg_n2_p;
614
static float *gapz_n1;
615
static float *gapz_n2;
616
static float **cpmx1;
617
static float **cpmx2;
618
static int **intwork;
619
static float **floatwork;
620
static int orlgth1 = 0, orlgth2 = 0;
621
float fpenalty = (float)penalty;
623
float *fg_t_og_h_dg_n2_p_pt;
624
float *og_t_fg_h_dg_n2_p_pt;
625
float *og_h_dg_n2_p_pt;
626
float *fg_h_dg_n2_p_pt;
631
float fg_t_og_h_dg_n1_p_va;
632
float og_t_fg_h_dg_n1_p_va;
633
float og_h_dg_n1_p_va;
634
float fg_h_dg_n1_p_va;
644
fprintf( stderr, "#### seq1[0] = %s\n", seq1[0] );
645
fprintf( stderr, "#### seq2[0] = %s\n", seq2[0] );
649
mseq1 = AllocateCharMtx( njob, 0 );
650
mseq2 = AllocateCharMtx( njob, 0 );
654
lgth1 = strlen( seq1[0] );
655
lgth2 = strlen( seq2[0] );
657
if( lgth1 == 0 || lgth2 == 0 )
659
fprintf( stderr, "WARNING (Aalignmm): lgth1=%d, lgth2=%d\n", lgth1, lgth2 );
663
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
667
if( orlgth1 > 0 && orlgth2 > 0 )
671
FreeFloatVec( match );
672
FreeFloatVec( initverticalw );
673
FreeFloatVec( lastverticalw );
680
FreeFloatVec( digf1 );
681
FreeFloatVec( digf2 );
682
FreeFloatVec( diaf1 );
683
FreeFloatVec( diaf2 );
684
FreeFloatVec( gapz1 );
685
FreeFloatVec( gapz2 );
686
FreeFloatVec( gapf1 );
687
FreeFloatVec( gapf2 );
688
FreeFloatVec( ogcp1g );
689
FreeFloatVec( ogcp2g );
690
FreeFloatVec( fgcp1g );
691
FreeFloatVec( fgcp2g );
692
FreeFloatVec( og_h_dg_n1_p );
693
FreeFloatVec( og_h_dg_n2_p );
694
FreeFloatVec( fg_h_dg_n1_p );
695
FreeFloatVec( fg_h_dg_n2_p );
696
FreeFloatVec( og_t_fg_h_dg_n1_p );
697
FreeFloatVec( og_t_fg_h_dg_n2_p );
698
FreeFloatVec( fg_t_og_h_dg_n1_p );
699
FreeFloatVec( fg_t_og_h_dg_n2_p );
700
FreeFloatVec( gapz_n1 );
701
FreeFloatVec( gapz_n2 );
703
FreeFloatMtx( cpmx1 );
704
FreeFloatMtx( cpmx2 );
706
FreeFloatMtx( floatwork );
707
FreeIntMtx( intwork );
710
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
711
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
714
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
717
w1 = AllocateFloatVec( ll2+2 );
718
w2 = AllocateFloatVec( ll2+2 );
719
match = AllocateFloatVec( ll2+2 );
721
initverticalw = AllocateFloatVec( ll1+2 );
722
lastverticalw = AllocateFloatVec( ll1+2 );
724
m = AllocateFloatVec( ll2+2 );
725
mp = AllocateIntVec( ll2+2 );
727
mseq = AllocateCharMtx( njob, ll1+ll2 );
729
digf1 = AllocateFloatVec( ll1+2 );
730
digf2 = AllocateFloatVec( ll2+2 );
731
diaf1 = AllocateFloatVec( ll1+2 );
732
diaf2 = AllocateFloatVec( ll2+2 );
733
gapz1 = AllocateFloatVec( ll1+2 );
734
gapz2 = AllocateFloatVec( ll2+2 );
735
gapf1 = AllocateFloatVec( ll1+2 );
736
gapf2 = AllocateFloatVec( ll2+2 );
737
ogcp1g = AllocateFloatVec( ll1+2 );
738
ogcp2g = AllocateFloatVec( ll2+2 );
739
fgcp1g = AllocateFloatVec( ll1+2 );
740
fgcp2g = AllocateFloatVec( ll2+2 );
741
og_h_dg_n1_p = AllocateFloatVec( ll1 + 2 );
742
og_h_dg_n2_p = AllocateFloatVec( ll2 + 2 );
743
fg_h_dg_n1_p = AllocateFloatVec( ll1 + 2 );
744
fg_h_dg_n2_p = AllocateFloatVec( ll2 + 2 );
745
og_t_fg_h_dg_n1_p = AllocateFloatVec( ll1 + 2 );
746
og_t_fg_h_dg_n2_p = AllocateFloatVec( ll2 + 2 );
747
fg_t_og_h_dg_n1_p = AllocateFloatVec( ll1 + 2 );
748
fg_t_og_h_dg_n2_p = AllocateFloatVec( ll2 + 2 );
749
gapz_n1 = AllocateFloatVec( ll1+2 );
750
gapz_n2 = AllocateFloatVec( ll2+2 );
752
cpmx1 = AllocateFloatMtx( 26, ll1+2 );
753
cpmx2 = AllocateFloatMtx( 26, ll2+2 );
754
fprintf( stderr, "ll2 = %d\n", ll2 );
757
floatwork = AllocateFloatMtx( MAX( ll1, ll2 )+2, 26 );
758
intwork = AllocateIntMtx( MAX( ll1, ll2 )+2, 27 );
760
floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 );
761
intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 );
765
fprintf( stderr, "succeeded\n" );
773
for( i=0; i<icyc; i++ )
778
for( j=0; j<jcyc; j++ )
780
mseq2[j] = mseq[icyc+j];
785
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
789
if( commonAlloc1 && commonAlloc2 )
791
FreeShortMtx( commonIP );
794
ll1 = MAX( orlgth1, commonAlloc1 );
795
ll2 = MAX( orlgth2, commonAlloc2 );
798
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
801
commonIP = AllocateShortMtx( ll1+10, ll2+10 );
804
fprintf( stderr, "succeeded\n\n" );
815
for( i=0; i<icyc; i++ )
817
fprintf( stderr, "## totaleff = %f\n", t );
821
cpmx_calc_new( seq1, cpmx1, eff1, lgth1, icyc );
822
cpmx_calc_new( seq2, cpmx2, eff2, lgth2, jcyc );
826
new_OpeningGapCount_zure( ogcp1g, icyc, seq1, eff1, lgth1, sgap1, egap1 );
827
new_OpeningGapCount_zure( ogcp2g, jcyc, seq2, eff2, lgth2, sgap2, egap2 );
828
new_FinalGapCount_zure( fgcp1g, icyc, seq1, eff1, lgth1, sgap1, egap1 );
829
new_FinalGapCount_zure( fgcp2g, jcyc, seq2, eff2, lgth2, sgap2, egap2 );
830
getdigapfreq_part( digf1, icyc, seq1, eff1, lgth1, sgap1, egap1 );
831
getdigapfreq_part( digf2, jcyc, seq2, eff2, lgth2, sgap2, egap2 );
832
getdiaminofreq_part( diaf1, icyc, seq1, eff1, lgth1, sgap1, egap1 );
833
getdiaminofreq_part( diaf2, jcyc, seq2, eff2, lgth2, sgap2, egap2 );
834
getgapfreq( gapf1, icyc, seq1, eff1, lgth1 );
835
getgapfreq( gapf2, jcyc, seq2, eff2, lgth2 );
836
getgapfreq_zure_part( gapz1, icyc, seq1, eff1, lgth1, sgap1 );
837
getgapfreq_zure_part( gapz2, jcyc, seq2, eff2, lgth2, sgap1 );
841
st_OpeningGapCount( ogcp1g, icyc, seq1, eff1, lgth1 );
842
st_OpeningGapCount( ogcp2g, jcyc, seq2, eff2, lgth2 );
843
st_FinalGapCount_zure( fgcp1g, icyc, seq1, eff1, lgth1 );
844
st_FinalGapCount_zure( fgcp2g, jcyc, seq2, eff2, lgth2 );
845
getdigapfreq_st( digf1, icyc, seq1, eff1, lgth1 );
846
getdigapfreq_st( digf2, jcyc, seq2, eff2, lgth2 );
847
getdiaminofreq_x( diaf1, icyc, seq1, eff1, lgth1 );
848
getdiaminofreq_x( diaf2, jcyc, seq2, eff2, lgth2 );
849
getgapfreq( gapf1, icyc, seq1, eff1, lgth1 );
850
getgapfreq( gapf2, jcyc, seq2, eff2, lgth2 );
851
getgapfreq_zure( gapz1, icyc, seq1, eff1, lgth1 );
852
getgapfreq_zure( gapz2, jcyc, seq2, eff2, lgth2 );
857
// fprintf( stderr, "lgth2 = %d\n", lgth2 );
858
for( i=0; i<lastj; i++ )
860
og_h_dg_n2_p[i] = ( 1.0-ogcp2g[i]-digf2[i] ) * fpenalty * 0.5;
861
fg_h_dg_n2_p[i] = ( 1.0-fgcp2g[i]-digf2[i] ) * fpenalty * 0.5;
862
og_t_fg_h_dg_n2_p[i] = (1.0-ogcp2g[i]+fgcp2g[i]-digf2[i]) * 0.5 * fpenalty;
863
fg_t_og_h_dg_n2_p[i] = (1.0-fgcp2g[i]+ogcp2g[i]-digf2[i]) * 0.5 * fpenalty;
864
gapz_n2[i] = (1.0-gapz2[i]);
867
for( i=0; i<lastj; i++ )
869
og_h_dg_n1_p[i] = ( 1.0-ogcp1g[i]-digf1[i] ) * fpenalty * 0.5;
870
fg_h_dg_n1_p[i] = ( 1.0-fgcp1g[i]-digf1[i] ) * fpenalty * 0.5;
871
og_t_fg_h_dg_n1_p[i] = (1.0-ogcp1g[i]+fgcp1g[i]-digf1[i]) * 0.5 * fpenalty;
872
fg_t_og_h_dg_n1_p[i] = (1.0-fgcp1g[i]+ogcp1g[i]-digf1[i]) * 0.5 * fpenalty;
873
gapz_n1[i] = (1.0-gapz1[i]);
879
for( i=0; i<lgth1; i++ )
880
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
886
match_calc( initverticalw, cpmx2, cpmx1, 0, lgth1, floatwork, intwork, 1 );
888
imp_match_out_vead_tateQ( initverticalw, 0, lgth1 ); // 060306
890
match_calc( currentw, cpmx1, cpmx2, 0, lgth2, floatwork, intwork, 1 );
892
imp_match_out_veadQ( currentw, 0, lgth2 ); // 060306
897
imp_match_calc( currentw, icyc, jcyc, lgth1, lgth2, seq1, seq2, eff1, eff2, localhom, 1, 0 );
908
// g += ogcp1g[0] * og_h_dg_n2_p[0];
909
g += ogcp1g[0] * ( 1.0-ogcp2g[0]-digf2[0] ) * fpenalty * 0.5;
910
// if( g ) fprintf( stderr, "init-match penal1=%f, %c-%c\n", g, seq1[0][0], seq2[0][0] );
912
// g += ogcp2g[0] * og_h_dg_n1_p[0];
913
g += ogcp2g[0] * ( 1.0-ogcp1g[0]-digf1[0] ) * fpenalty * 0.5;
914
// if( g ) fprintf( stderr, "init-match penal2=%f, %c-%c\n", g, seq1[0][0], seq2[0][0] );
916
// g += fgcp1g[0] * fg_h_dg_n2_p[0];
917
g += fgcp1g[0] * ( 1.0-fgcp2g[0]-digf2[0] ) * fpenalty * 0.5;
918
// if( g ) fprintf( stderr, "match penal1=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
920
// g += fgcp2g[0] * fg_h_dg_n1_p[0];
921
g += fgcp2g[0] * ( 1.0-fgcp1g[0]-digf1[0] ) * fpenalty * 0.5;
922
// if( g ) fprintf( stderr, "match penal2=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
925
initverticalw[0] += g;
928
for( i=1; i<lgth1+1; i++ )
930
// tmppenal = gapz_n2[0]*og_t_fg_h_dg_n1_p[0];
931
tmppenal = ( (1.0-gapz2[0])*(1.0-ogcp1g[0]+fgcp1g[0]-digf1[0]) ) * 0.5 * fpenalty; // mada
932
initverticalw[i] += tmppenal;
934
// tmppenal = gapz_n2[1]*fg_t_og_h_dg_n1_p[i];
935
tmppenal = ( (1.0-gapz2[1])*(1.0-fgcp1g[i]+ogcp1g[i]-digf1[i]) ) * 0.5 * fpenalty; // mada
936
initverticalw[i] += tmppenal;
939
for( j=1; j<lgth2+1; j++ )
941
// tmppenal = gapz_n1[0]*og_t_fg_h_dg_n2_p[0];
942
tmppenal = ( (1.0-gapz1[0])*(1.0-ogcp2g[0]+fgcp2g[0]-digf2[0]) ) * 0.5 * fpenalty; // mada
943
currentw[j] += tmppenal;
945
// tmppenal = gapz_n1[1]*fg_t_og_h_dg_n2_p[j];
946
tmppenal = ( (1.0-gapz1[1])*(1.0-fgcp2g[j]+ogcp2g[j]-digf2[j]) ) * 0.5 * fpenalty; // mada
947
currentw[j] += tmppenal;
953
for( j=1; j<lgth2+1; j++ )
954
currentw[j] -= offset * j / 2.0;
955
for( i=1; i<lgth1+1; i++ )
956
initverticalw[i] -= offset * i / 2.0;
960
m[0] = 0.0; // iranai
961
for( j=1; j<lgth2+1; ++j )
964
m[j] = currentw[j-1] + 10000 * fpenalty; //iinoka?
967
lastverticalw[0] = 0.0; // Falign kara yobaretatoki kounarukanousei ari
969
lastverticalw[0] = currentw[lgth2-1];
971
if( outgap ) lasti = lgth1+1; else lasti = lgth1;
974
fprintf( stderr, "currentw = \n" );
975
for( i=0; i<lgth1+1; i++ )
977
fprintf( stderr, "%5.2f ", currentw[i] );
979
fprintf( stderr, "\n" );
980
fprintf( stderr, "initverticalw = \n" );
981
for( i=0; i<lgth2+1; i++ )
983
fprintf( stderr, "%5.2f ", initverticalw[i] );
985
fprintf( stderr, "\n" );
986
fprintf( stderr, "fcgp\n" );
987
for( i=0; i<lgth1; i++ )
988
fprintf( stderr, "fgcp1[%d]=%f\n", i, ogcp1[i] );
989
for( i=0; i<lgth2; i++ )
990
fprintf( stderr, "fgcp2[%d]=%f\n", i, ogcp2[i] );
993
for( i=1; i<lasti; i++ )
996
previousw = currentw;
999
previousw[0] = initverticalw[i-1];
1001
match_calc( currentw, cpmx1, cpmx2, i, lgth2, floatwork, intwork, 0 );
1003
fprintf( stderr, "\n" );
1004
fprintf( stderr, "i=%d\n", i );
1005
fprintf( stderr, "currentw = \n" );
1006
for( j=0; j<lgth2; j++ )
1008
fprintf( stderr, "%5.2f ", currentw[j] );
1010
fprintf( stderr, "\n" );
1014
// fprintf( stderr, "Calling imp_match_calc (o) lgth = %d, i = %d\n", lgth1, i );
1016
imp_match_out_veadQ( currentw, i, lgth2 );
1018
imp_match_out_veadQ( currentw, i, lgth2 );
1022
fprintf( stderr, "\n" );
1023
fprintf( stderr, "i=%d\n", i );
1024
fprintf( stderr, "currentw = \n" );
1025
for( j=0; j<lgth2; j++ )
1027
fprintf( stderr, "%5.2f ", currentw[j] );
1029
fprintf( stderr, "\n" );
1031
currentw[0] = initverticalw[i];
1034
mi = previousw[0] + 10000 * fpenalty;
1039
curpt = currentw + 1;
1041
fg_t_og_h_dg_n2_p_pt = fg_t_og_h_dg_n2_p + 1;
1042
og_t_fg_h_dg_n2_p_pt = og_t_fg_h_dg_n2_p + 1;
1043
og_h_dg_n2_p_pt = og_h_dg_n2_p + 1;
1044
fg_h_dg_n2_p_pt = fg_h_dg_n2_p + 1;
1045
gapz_n2_pt0 = gapz_n2 + 1;
1046
gapz_n2_pt1 = gapz_n2 + 2;
1047
fgcp2pt = fgcp2g + 1;
1048
ogcp2pt = ogcp2g + 1;
1050
fg_t_og_h_dg_n1_p_va = fg_t_og_h_dg_n1_p[i];
1051
og_t_fg_h_dg_n1_p_va = og_t_fg_h_dg_n1_p[i];
1052
og_h_dg_n1_p_va = og_h_dg_n1_p[i];
1053
fg_h_dg_n1_p_va = fg_h_dg_n1_p[i];
1054
gapz_n1_va0 = gapz_n1[i];
1055
gapz_n1_va1 = gapz_n1[i+1];
1056
fgcp1va = fgcp1g[i];
1057
ogcp1va = ogcp1g[i];
1060
for( j=1; j<lastj; j++ )
1064
// g = ogcp1va * *og_h_dg_n2_p_pt;
1065
// g = ogcp1g[i] * og_h_dg_n2_p[j];
1066
g = ogcp1g[i] * ( 1.0-ogcp2g[j]-digf2[j] ) * fpenalty * 0.5;
1067
// if( g && i==j ) fprintf( stderr, "match penal1=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
1070
// g = *ogcp2pt * og_h_dg_n1_p_va;
1071
// g = ogcp2g[j] * og_h_dg_n1_p[i];
1072
g = ogcp2g[j] * ( 1.0-ogcp1g[i]-digf1[i] ) * fpenalty * 0.5;
1073
// if( g && i==j ) fprintf( stderr, "match penal2=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
1076
// g = fgcp1va * *fg_h_dg_n2_p_pt;
1077
// g = fgcp1g[i] * fg_h_dg_n2_p[j];
1078
g = fgcp1g[i] * ( 1.0-fgcp2g[j]-digf2[j] ) * fpenalty * 0.5;
1079
// if( g && i==j ) fprintf( stderr, "match penal3=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
1082
// g = *fgcp2pt * fg_h_dg_n1_p_va;
1083
// g = fgcp2g[j] * fg_h_dg_n1_p[i];
1084
g = fgcp2g[j] * ( 1.0-fgcp1g[i]-digf1[i] ) * fpenalty * 0.5;
1085
// if( g && i==j ) fprintf( stderr, "match penal4=%f, %c-%c\n", g, seq1[0][i], seq2[0][j] );
1091
fprintf( stderr, "%5.0f->", wm );
1094
fprintf( stderr, "%5.0f?", g );
1096
// tmppenal = gapz_n1_va1 * *fg_t_og_h_dg_n2_p_pt;
1097
// tmppenal = gapz_n1[i+1] * fg_t_og_h_dg_n2_p[j];
1098
tmppenal = ( (1.0-gapz1[i+1])*(1.0-fgcp2g[j]+ogcp2g[j]-digf2[j]) ) * 0.5 * fpenalty; // mada
1099
if( (g=mi+tmppenal) > wm )
1101
// fprintf( stderr, "jump i start=%f (j=%d, fgcp2g[j]=%f, digf2[j]=%f, diaf2[j]=%f), %c-%c\n", g-mi, j, fgcp2g[j], digf2[j], diaf2[j], seq1[0][i], seq2[0][j] );
1103
*ijppt = -( j - mpi );
1105
// tmppenal = gapz_n1_va0 * *og_t_fg_h_dg_n2_p_pt;
1106
// tmppenal = gapz_n1[i] * og_t_fg_h_dg_n2_p[j];
1107
tmppenal = ( (1.0-gapz1[i])*(1.0-ogcp2g[j]+fgcp2g[j]-digf2[j]) ) * 0.5 * fpenalty; // mada
1108
if( (g=*prept+tmppenal) >= mi )
1110
// fprintf( stderr, "jump i end=%f, %c-%c\n", g-*prept, seq1[0][i-1], seq2[0][j-1] );
1116
fprintf( stderr, "%5.0f?", g );
1118
// tmppenal = *gapz_n2_pt1 * fg_t_og_h_dg_n1_p_va;
1119
// tmppenal = gapz_n2[j+1] * fg_t_og_h_dg_n1_p[i];
1120
tmppenal = ( (1.0-gapz2[j+1])*(1.0-fgcp1g[i]+ogcp1g[i]-digf1[i]) ) * 0.5 * fpenalty; // mada
1121
if( (g=*mjpt+tmppenal) > wm )
1124
*ijppt = +( i - *mpjpt );
1126
// tmppenal = *gapz_n2_pt0 * og_t_fg_h_dg_n1_p_va;
1127
// tmppenal = gapz_n2[j] * og_t_fg_h_dg_n1_p[i];
1128
tmppenal = ( (1.0-gapz2[j])*(1.0-ogcp1g[i]+fgcp1g[i]-digf1[i]) ) * 0.5 * fpenalty; // mada
1129
if( (g=*prept+tmppenal) >= *mjpt )
1135
fprintf( stderr, "%5.0f ", wm );
1142
fg_t_og_h_dg_n2_p_pt++;
1143
og_t_fg_h_dg_n2_p_pt++;
1151
lastverticalw[i] = currentw[lgth2-1];
1154
// fprintf( stderr, "wm = %f\n", wm );
1159
for( j=1; j<lgth2+1; j++ )
1160
currentw[j] -= offset * ( lgth2 - j ) / 2.0;
1161
for( i=1; i<lgth1+1; i++ )
1162
lastverticalw[i] -= offset * ( lgth1 - i / 2.0);
1167
fprintf( stderr, "\n" );
1168
for( i=0; i<icyc; i++ ) fprintf( stderr,"%s\n", seq1[i] );
1169
fprintf( stderr, "#####\n" );
1170
for( j=0; j<jcyc; j++ ) fprintf( stderr,"%s\n", seq2[j] );
1171
fprintf( stderr, "====>" );
1172
for( i=0; i<icyc; i++ ) strcpy( mseq1[i], seq1[i] );
1173
for( j=0; j<jcyc; j++ ) strcpy( mseq2[j], seq2[j] );
1177
Atracking_localhom( impmatch, currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc );
1180
Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc );
1182
// fprintf( stderr, "### impmatch = %f\n", *impmatch );
1184
resultlen = strlen( mseq1[0] );
1185
if( alloclen < resultlen || resultlen > N )
1187
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
1188
ErrorExit( "LENGTH OVER!\n" );
1192
for( i=0; i<icyc; i++ ) strcpy( seq1[i], mseq1[i] );
1193
for( j=0; j<jcyc; j++ ) strcpy( seq2[j], mseq2[j] );
1195
fprintf( stderr, "\n" );
1196
for( i=0; i<icyc; i++ ) fprintf( stderr, "%s\n", mseq1[i] );
1197
fprintf( stderr, "#####\n" );
1198
for( j=0; j<jcyc; j++ ) fprintf( stderr, "%s\n", mseq2[j] );
1201
fprintf( stderr, "wm = %f\n", wm );