static float recalllara( char **mseq1, char **mseq2, int alloclen )
{
static FILE *fp = NULL;
static char *ungap1;
static char *ungap2;
static char *ori1;
static char *ori2;
int res;
static char com[10000];
float value;
if( fp == NULL )
{
fp = fopen( "_laraout", "r" );
if( fp == NULL )
{
fprintf( stderr, "Cannot open _laraout\n" );
exit( 1 );
}
ungap1 = AllocateCharVec( alloclen );
ungap2 = AllocateCharVec( alloclen );
ori1 = AllocateCharVec( alloclen );
ori2 = AllocateCharVec( alloclen );
}
strcpy( ori1, *mseq1 );
strcpy( ori2, *mseq2 );
fgets( com, 999, fp );
myfgets( com, 9999, fp );
strcpy( *mseq1, com );
myfgets( com, 9999, fp );
strcpy( *mseq2, com );
gappick0( ungap1, *mseq1 );
gappick0( ungap2, *mseq2 );
t2u( ungap1 );
t2u( ungap2 );
if( strcmp( ungap1, ori1 ) || strcmp( ungap2, ori2 ) )
{
fprintf( stderr, "SEQUENCE CHANGED!!\n" );
fprintf( stderr, "*mseq1 = %s\n", *mseq1 );
fprintf( stderr, "ungap1 = %s\n", ungap1 );
fprintf( stderr, "ori1 = %s\n", ori1 );
fprintf( stderr, "*mseq2 = %s\n", *mseq2 );
fprintf( stderr, "ungap2 = %s\n", ungap2 );
fprintf( stderr, "ori2 = %s\n", ori2 );
exit( 1 );
}
value = (float)naivepairscore11( *mseq1, *mseq2, penalty );
// fclose( fp ); // saigo dake yatta houga yoi.
return( value );
}
void foldrna( int nseq1, int nseq2, char **seq1, char **seq2, double *eff1, double *eff2, RNApair ***grouprna1, RNApair ***grouprna2, float **impmtx, int *gapmap1, int *gapmap2, RNApair *additionalpair )
{
int i, j;
// int ui, uj;
// int uiup, ujup;
int uido, ujdo;
static TLS char **useq1, **useq2;
static TLS char **oseq1, **oseq2, **oseq1r, **oseq2r, *odir1, *odir2;
static TLS RNApair **pairprob1, **pairprob2;
static TLS RNApair *pairpt1, *pairpt2;
int lgth1 = strlen( seq1[0] );
int lgth2 = strlen( seq2[0] );
static TLS float **impmtx2;
static TLS float **map;
// double lenfac;
float prob;
int **sgapmap1, **sgapmap2;
char *nogapdum;
float **tbppmtx;
// fprintf( stderr, "nseq1=%d, lgth1=%d\n", nseq1, lgth1 );
useq1 = AllocateCharMtx( nseq1, lgth1+10 );
useq2 = AllocateCharMtx( nseq2, lgth2+10 );
oseq1 = AllocateCharMtx( nseq1, lgth1+10 );
oseq2 = AllocateCharMtx( nseq2, lgth2+10 );
oseq1r = AllocateCharMtx( nseq1, lgth1+10 );
oseq2r = AllocateCharMtx( nseq2, lgth2+10 );
odir1 = AllocateCharVec( lgth1+10 );
odir2 = AllocateCharVec( lgth2+10 );
sgapmap1 = AllocateIntMtx( nseq1, lgth1+1 );
sgapmap2 = AllocateIntMtx( nseq2, lgth2+1 );
nogapdum = AllocateCharVec( MAX( lgth1, lgth2 ) );
pairprob1 = (RNApair **)calloc( lgth1, sizeof( RNApair *) );
pairprob2 = (RNApair **)calloc( lgth2, sizeof( RNApair *) );
map = AllocateFloatMtx( lgth1, lgth2 );
impmtx2 = AllocateFloatMtx( lgth1, lgth2 );
tbppmtx = AllocateFloatMtx( lgth1, lgth2 );
for( i=0; i<nseq1; i++ ) strcpy( useq1[i], seq1[i] );
for( i=0; i<nseq2; i++ ) strcpy( useq2[i], seq2[i] );
for( i=0; i<nseq1; i++ ) strcpy( oseq1[i], seq1[i] );
for( i=0; i<nseq2; i++ ) strcpy( oseq2[i], seq2[i] );
for( i=0; i<nseq1; i++ ) commongappick_record( 1, useq1+i, sgapmap1[i] );
for( i=0; i<nseq2; i++ ) commongappick_record( 1, useq2+i, sgapmap2[i] );
for( i=0; i<lgth1; i++ )
{
pairprob1[i] = (RNApair *)calloc( 1, sizeof( RNApair ) );
pairprob1[i][0].bestpos = -1;
pairprob1[i][0].bestscore = -1;
}
for( i=0; i<lgth2; i++ )
{
pairprob2[i] = (RNApair *)calloc( 1, sizeof( RNApair ) );
pairprob2[i][0].bestpos = -1;
pairprob2[i][0].bestscore = -1;
}
utot( nseq1, lgth1, oseq1 );
utot( nseq2, lgth2, oseq2 );
// fprintf( stderr, "folding group1\n" );
// rnalocal( oseq1, useq1, eff1, eff1, nseq1, nseq1, lgth1+10, pair1 );
/* base-pairing probability of group 1 */
if( rnaprediction == 'r' )
rnaalifoldcall( oseq1, nseq1, pairprob1 );
else
mccaskillextract( oseq1, useq1, nseq1, pairprob1, grouprna1, sgapmap1, eff1 );
// fprintf( stderr, "folding group2\n" );
// rnalocal( oseq2, useq2, eff2, eff2, nseq2, nseq2, lgth2+10, pair2 );
/* base-pairing probability of group 2 */
if( rnaprediction == 'r' )
rnaalifoldcall( oseq2, nseq2, pairprob2 );
else
mccaskillextract( oseq2, useq2, nseq2, pairprob2, grouprna2, sgapmap2, eff2 );
#if 0
makerseq( oseq1, oseq1r, odir1, pairprob1, nseq1, lgth1 );
makerseq( oseq2, oseq2r, odir2, pairprob2, nseq2, lgth2 );
fprintf( stderr, "%s\n", odir2 );
for( i=0; i<nseq1; i++ )
{
fprintf( stdout, ">ori\n%s\n", oseq1[0] );
fprintf( stdout, ">rev\n%s\n", oseq1r[0] );
}
#endif
/* similarity score */
Lalignmm_hmout( oseq1, oseq2, eff1, eff2, nseq1, nseq2, 10000, NULL, NULL, NULL, NULL, map );
if( 1 )
{
if( RNAscoremtx == 'n' )
{
for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ )
{
// impmtx2[i][j] = osoiaveragescore( nseq1, nseq2, oseq1, oseq2, eff1, eff2, i, j ) * consweight_multi;
impmtx2[i][j] = 0.0;
}
}
else if( RNAscoremtx == 'r' )
{
for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ )
{
tbppmtx[i][j] = 1.0;
impmtx2[i][j] = 0.0;
}
for( i=0; i<lgth1; i++ ) for( pairpt1=pairprob1[i]; pairpt1->bestpos!=-1; pairpt1++ )
{
for( j=0; j<lgth2; j++ ) for( pairpt2=pairprob2[j]; pairpt2->bestpos!=-1; pairpt2++ )
{
uido = pairpt1->bestpos;
ujdo = pairpt2->bestpos;
prob = pairpt1->bestscore * pairpt2->bestscore;
if( uido > -1 && ujdo > -1 )
{
if( uido > i && j > ujdo )
{
impmtx2[i][j] += prob * pairedribosumscore53( nseq1, nseq2, oseq1, oseq2, eff1, eff2, i, j, uido, ujdo ) * consweight_multi;
tbppmtx[i][j] -= prob;
}
else if( i < uido && j < ujdo )
{
impmtx2[i][j] += prob * pairedribosumscore35( nseq1, nseq2, oseq1, oseq2, eff1, eff2, i, j, uido, ujdo ) * consweight_multi;
tbppmtx[i][j] -= prob;
}
}
}
}
for( i=0; i<lgth1; i++ )
{
for( j=0; j<lgth2; j++ )
{
impmtx2[i][j] += tbppmtx[i][j] * singleribosumscore( nseq1, nseq2, oseq1, oseq2, eff1, eff2, i, j ) * consweight_multi;
}
}
}
/* four-way consistency */
for( i=0; i<lgth1; i++ ) for( pairpt1=pairprob1[i]; pairpt1->bestpos!=-1; pairpt1++ )
{
// if( pairprob1[i] == NULL ) continue;
for( j=0; j<lgth2; j++ ) for( pairpt2=pairprob2[j]; pairpt2->bestpos!=-1; pairpt2++ )
{
// fprintf( stderr, "i=%d, j=%d, pn1=%d, pn2=%d\n", i, j, pairpt1-pairprob1[i], pairpt2-pairprob2[j] );
// if( pairprob2[j] == NULL ) continue;
uido = pairpt1->bestpos;
ujdo = pairpt2->bestpos;
prob = pairpt1->bestscore * pairpt2->bestscore;
// prob = 1.0;
// fprintf( stderr, "i=%d->uido=%d, j=%d->ujdo=%d\n", i, uido, j, ujdo );
// fprintf( stderr, "impmtx2[%d][%d] = %f\n", i, j, impmtx2[i][j] );
// if( i < uido && j > ujdo ) continue;
// if( i > uido && j < ujdo ) continue;
// posdistj = abs( ujdo-j );
// if( uido > -1 && ujdo > -1 )
if( uido > -1 && ujdo > -1 && ( ( i > uido && j > ujdo ) || ( i < uido && j < ujdo ) ) )
{
{
impmtx2[i][j] += MAX( 0, map[uido][ujdo] ) * consweight_rna * 600 * prob; // osoi
}
}
}
}
for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ )
{
impmtx[i][j] += impmtx2[i][j];
// fprintf( stderr, "fastathreshold=%f, consweight_multi=%f, consweight_rna=%f\n", fastathreshold, consweight_multi, consweight_rna );
// impmtx[i][j] *= 0.5;
}
// impmtx[0][0] += 10000.0;
// impmtx[lgth1-1][lgth2-1] += 10000.0;
#if 0
fprintf( stdout, "#impmtx2 = \n" );
for( i=0; i<lgth1; i++ )
{
for( j=0; j<lgth2; j++ )
{
fprintf( stdout, "%d %d %f\n", i, j, impmtx2[i][j] );
}
fprintf( stdout, "\n" );
}
exit( 1 );
#endif
}
FreeCharMtx( useq1 );
FreeCharMtx( useq2 );
FreeCharMtx( oseq1 );
FreeCharMtx( oseq2 );
FreeCharMtx( oseq1r );
FreeCharMtx( oseq2r );
free( odir1 );
free( odir2 );
FreeFloatMtx( impmtx2 );
FreeFloatMtx( map );
FreeIntMtx( sgapmap1 );
FreeIntMtx( sgapmap2 );
FreeFloatMtx( tbppmtx );
for( i=0; i<lgth1; i++ ) free( pairprob1[i] );
for( i=0; i<lgth2; i++ ) free( pairprob2[i] );
free( pairprob1 );
free( pairprob2 );
}
int main( int argc, char **argv )
{
int i, j;
FILE *fp, *infp;
char **seq;
int *grpseq;
char *tmpseq;
int **pointt;
static char name[M][B];
static int nlen[M];
double **mtx;
double **mtx2;
double score, score0;
static short *table1;
char b[B];
arguments( argc, argv );
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
#if 0
PreRead( stdin, &njob, &nlenmax );
#else
getnumlen( infp );
#endif
rewind( infp );
if( njob < 2 )
{
fprintf( stderr, "At least 2 sequences should be input!\n"
"Only %d sequence found.\n", njob );
exit( 1 );
}
tmpseq = AllocateCharVec( nlenmax+1 );
seq = AllocateCharMtx( njob, nlenmax+1 );
grpseq = AllocateIntVec( nlenmax+1 );
pointt = AllocateIntMtx( njob, nlenmax+1 );
mtx = AllocateDoubleMtx( njob, njob );
mtx2 = AllocateDoubleMtx( njob, njob );
pamN = NOTSPECIFIED;
#if 0
FRead( infp, name, nlen, seq );
#else
readData( infp, name, nlen, seq );
#endif
fclose( infp );
constants( njob, seq );
if( dorp == 'd' ) tsize = (int)pow( 4, 6 );
else tsize = (int)pow( 6, 6 );
maxl = 0;
for( i=0; i<njob; i++ )
{
gappick0( tmpseq, seq[i] );
nlen[i] = strlen( tmpseq );
if( nlen[i] < 6 )
{
fprintf( stderr, "Seq %d, too short, %d characters\n", i+1, nlen[i] );
exit( 1 );
}
if( nlen[i] > maxl ) maxl = nlen[i];
if( dorp == 'd' ) /* nuc */
{
seq_grp_nuc( grpseq, tmpseq );
makepointtable_nuc( pointt[i], grpseq );
}
else /* amino */
{
seq_grp( grpseq, tmpseq );
makepointtable( pointt[i], grpseq );
}
}
for( i=0; i<njob; i++ )
{
table1 = (short *)calloc( tsize, sizeof( short ) );
if( !table1 ) ErrorExit( "Cannot allocate table1\n" );
if( i % 10 == 0 )
{
fprintf( stderr, "%4d / %4d\r", i+1, njob );
}
makecompositiontable_p( table1, pointt[i] );
for( j=i; j<njob; j++ )
{
score = (double)commonsextet_p( table1, pointt[j] );
mtx[i][j] = score;
}
free( table1 );
}
for( i=0; i<njob; i++ )
{
score0 = mtx[i][i];
for( j=0; j<njob; j++ )
mtx2[i][j] = ( score0 - mtx[MIN(i,j)][MAX(i,j)] ) / score0 * 3.0;
}
for( i=0; i<njob-1; i++ ) for( j=i+1; j<njob; j++ )
{
#if TEST
double jscore;
jscore = mtx[i][j] / ( MIN( strlen( seq[i] ), strlen( seq[j] ) ) - 2 );
fprintf( stdout, "jscore = %f\n", jscore );
fprintf( stdout, "mtx2[%d][%d] = %f, mtx2[%d][%d] = %f\n", i, j, mtx2[i][j], j, i, mtx2[j][i] );
#endif
mtx2[i][j] = MIN( mtx2[i][j], mtx2[j][i] );
#if TEST
fprintf( stdout, "sonokekka mtx2[%d][%d] %f\n", i, j, mtx2[i][j] );
#endif
}
if( disopt )
{
for( i=0; i<njob; i++ )
{
sprintf( b, "=lgth = %04d", nlen[i] );
strins( b, name[i] );
}
}
fp = fopen( "hat2", "w" );
if( !fp ) ErrorExit( "Cannot open hat2." );
WriteHat2( fp, njob, name, mtx2 );
fclose( fp );
fprintf( stderr, "\n" );
SHOWVERSION;
exit( 0 );
}
int main( int argc, char *argv[] )
{
FILE *infp;
FILE *difp;
int nlenmin;
char **name;
char **seq;
char *tmpseq;
char line[100];
int *nlen;
int i;
arguments( argc, argv );
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
if( directionfile )
{
difp = fopen( directionfile, "r" );
if( !difp )
{
fprintf( stderr, "Cannot open %s\n", directionfile );
exit( 1 );
}
}
else
{
fprintf( stderr, "Give directionfile!\n" );
}
dorp = NOTSPECIFIED;
getnumlen_casepreserve( infp, &nlenmin );
fprintf( stderr, "%d x %d - %d %c\n", njob, nlenmax, nlenmin, dorp );
seq = AllocateCharMtx( njob, nlenmax+1 );
tmpseq = AllocateCharVec( MAX( B, nlenmax )+1 );
name = AllocateCharMtx( njob, B+1 );
nlen = AllocateIntVec( njob );
readData_pointer_casepreserve( infp, name, nlen, seq );
for( i=0; i<njob; i++ )
{
fgets( line, 99, difp );
if( line[0] != '_' )
{
fprintf( stderr, "Format error!\n" );
exit( 1 );
}
if( line[1] == 'R' )
{
sreverse( tmpseq, seq[i] );
strcpy( seq[i], tmpseq );
strncpy( tmpseq, name[i]+1, B-3 );
tmpseq[B-3] = 0;
strcpy( name[i]+1, "_R_" );
strcpy( name[i]+4, tmpseq );
}
else if( line[1] == 'F' )
{
;
}
else
{
fprintf( stderr, "Format error!\n" );
exit( 1 );
}
}
for( i=0; i<njob; i++ )
{
fprintf( stdout, ">%s\n", name[i]+1 );
fprintf( stdout, "%s\n", seq[i] );
}
free( nlen );
FreeCharMtx( seq );
FreeCharMtx( name );
free( tmpseq );
return( 0 );
}
static void pairalign( char **name, int nlen[M], char **seq, double *effarr, int alloclen )
{
FILE *tmpfp;
static char dumm1[B], dumm0[B];
int i, j;
char *res;
FILE *hat3p;
static double *effarr1 = NULL;
static double *effarr2 = NULL;
static char **pseq;
LocalHom **localhomtable, *tmpptr;
float pscore = 0.0; // by D.Mathog, aguess
char *aseq = NULL; // by D.Mathog
char **usedseqs = NULL; // by D.Mathog
char **usednames = NULL; // by D.Mathog
int nused;
double tsuyosa;
tsuyosa = (double)nhomologs * (nhomologs-1) / njob * TSUYOSAFACTOR;
fprintf( stderr, "tsuyosa = %f\n", tsuyosa );
localhomtable = (LocalHom **)calloc( njob, sizeof( LocalHom *) );
for( i=0; i<njob; i++)
{
localhomtable[i] = (LocalHom *)calloc( njob, sizeof( LocalHom ) );
for( j=0; j<njob; j++)
{
localhomtable[i][j].start1 = -1;
localhomtable[i][j].end1 = -1;
localhomtable[i][j].start2 = -1;
localhomtable[i][j].end2 = -1;
localhomtable[i][j].opt = -1.0;
localhomtable[i][j].next = NULL;
}
}
if( effarr1 == NULL )
{
effarr1 = AllocateDoubleVec( njob );
effarr2 = AllocateDoubleVec( njob );
pseq = AllocateCharMtx( 2, nlenmax*9+1 );
aseq = AllocateCharVec( nlenmax*9+1 );
usedseqs = AllocateCharMtx( njob, nlenmax*9+1 );
usednames = AllocateCharMtx( njob, B );
#if 0
#else
#endif
}
#if 0
fprintf( stderr, "##### fftwinsize = %d, fftthreshold = %d\n", fftWinSize, fftThreshold );
#endif
#if 0
for( i=0; i<njob; i++ )
fprintf( stderr, "TBFAST effarr[%d] = %f\n", i, effarr[i] );
#endif
// writePre( njob, name, nlen, aseq, 0 );
fprintf( stderr, "opening %s\n", pairfile );
tmpfp = fopen( pairfile, "r" );
if( !tmpfp )
{
fprintf( stderr, "Cannot open %s\n", pairfile );
exit( 1 );
}
searchKUorWA( tmpfp );
hat3p = fopen( "hat3", "w" );
if( !hat3p ) ErrorExit( "Cannot open hat3." );
nused = 0;
while( 1 )
{
res = fgets( dumm0, B-1, tmpfp );
strip( dumm0 );
if( res == NULL )
{
break;
}
load1SeqWithoutName_new( tmpfp, pseq[0] );
gappick0( aseq, pseq[0] );
i = searchused( aseq, usedseqs, nused );
if( i == -1 )
{
strcpy( usednames[nused], dumm0+1 );
strcpy( usedseqs[nused], aseq );
i = nused;
nused++;
}
fprintf( stderr, "i = %d\n", i );
res = fgets( dumm1, B-1, tmpfp );
strip( dumm1 );
if( res == NULL )
{
fprintf( stderr, "ERROR: The number of sequences in %s must be even.\n", pairfile );
exit( 1 );
}
load1SeqWithoutName_new( tmpfp, pseq[1] );
gappick0( aseq, pseq[1] );
j = searchused( aseq, usedseqs, nused );
if( j == -1 )
{
strcpy( usednames[nused], dumm1+1 );
strcpy( usedseqs[nused], aseq );
j = nused;
nused++;
}
fprintf( stderr, "j = %d\n", j );
if( strlen( pseq[0] ) != strlen( pseq[1] ) )
{
fprintf( stderr, "Not aligned, %s - %s\n", dumm0, dumm1 );
exit( 1 );
}
fprintf( stderr, "adding %d-%d\n", i, j );
putlocalhom2( pseq[0], pseq[1], localhomtable[i]+j, 0, 0, (int)pscore, strlen( pseq[0] ) );
for( tmpptr=localhomtable[i]+j; tmpptr; tmpptr=tmpptr->next )
{
if( tmpptr->opt == -1.0 ) continue;
fprintf( hat3p, "%d %d %d %6.3f %d %d %d %d %p\n", i, j, tmpptr->overlapaa, tmpptr->opt * tsuyosa, tmpptr->start1, tmpptr->end1, tmpptr->start2, tmpptr->end2, (void *)tmpptr->next );
}
}
fclose( tmpfp );
fclose( hat3p );
for( i=0; i<nused; i++ )
fprintf( stdout, ">%s\n%s\n", usednames[i], usedseqs[i] );
#if 0
fprintf( stderr, "##### writing hat3\n" );
hat3p = fopen( "hat3", "w" );
if( !hat3p ) ErrorExit( "Cannot open hat3." );
ilim = njob-1;
for( i=0; i<ilim; i++ )
{
for( j=i+1; j<njob; j++ )
{
for( tmpptr=localhomtable[i]+j; tmpptr; tmpptr=tmpptr->next )
{
if( tmpptr->opt == -1.0 ) continue;
fprintf( hat3p, "%d %d %d %6.3f %d %d %d %d %p\n", i, j, tmpptr->overlapaa, tmpptr->opt * tsuyosa, tmpptr->start1, tmpptr->end1, tmpptr->start2, tmpptr->end2, tmpptr->next );
}
}
}
fclose( hat3p );
#endif
#if DEBUG
fprintf( stderr, "calling FreeLocalHomTable\n" );
#endif
FreeLocalHomTable( localhomtable, njob );
#if DEBUG
fprintf( stderr, "done. FreeLocalHomTable\n" );
#endif
}
static void pairalign( char name[M][B], int nlen[M], char **seq, char **aseq, char **mseq1, char **mseq2, double *effarr, int alloclen )
{
int i, j, ilim;
int clus1, clus2;
int off1, off2;
float pscore = 0.0; // by D.Mathog
static char *indication1, *indication2;
FILE *hat2p, *hat3p;
static double **distancemtx;
static double *effarr1 = NULL;
static double *effarr2 = NULL;
char *pt;
char *hat2file = "hat2";
LocalHom **localhomtable, *tmpptr;
static char **pair;
int intdum;
double bunbo;
localhomtable = (LocalHom **)calloc( njob, sizeof( LocalHom *) );
for( i=0; i<njob; i++)
{
localhomtable[i] = (LocalHom *)calloc( njob, sizeof( LocalHom ) );
for( j=0; j<njob; j++)
{
localhomtable[i][j].start1 = -1;
localhomtable[i][j].end1 = -1;
localhomtable[i][j].start2 = -1;
localhomtable[i][j].end2 = -1;
localhomtable[i][j].opt = -1.0;
localhomtable[i][j].next = NULL;
localhomtable[i][j].nokori = 0;
}
}
if( effarr1 == NULL )
{
distancemtx = AllocateDoubleMtx( njob, njob );
effarr1 = AllocateDoubleVec( njob );
effarr2 = AllocateDoubleVec( njob );
indication1 = AllocateCharVec( 150 );
indication2 = AllocateCharVec( 150 );
#if 0
#else
pair = AllocateCharMtx( njob, njob );
#endif
}
#if 0
fprintf( stderr, "##### fftwinsize = %d, fftthreshold = %d\n", fftWinSize, fftThreshold );
#endif
#if 0
for( i=0; i<njob; i++ )
fprintf( stderr, "TBFAST effarr[%d] = %f\n", i, effarr[i] );
#endif
// writePre( njob, name, nlen, aseq, 0 );
for( i=0; i<njob; i++ ) for( j=0; j<njob; j++ ) pair[i][j] = 0;
for( i=0; i<njob; i++ ) pair[i][i] = 1;
if( alg == 'H' )
{
fprintf( stderr, "Calling FOLDALIGN with option '%s'\n", foldalignopt );
callfoldalign( njob, seq );
fprintf( stderr, "done.\n" );
}
if( alg == 'B' )
{
fprintf( stderr, "Calling LARA\n" );
calllara( njob, seq, "" );
fprintf( stderr, "done.\n" );
}
if( alg == 'T' )
{
fprintf( stderr, "Calling SLARA\n" );
calllara( njob, seq, "-s" );
fprintf( stderr, "done.\n" );
}
ilim = njob - 1;
for( i=0; i<ilim; i++ )
{
fprintf( stderr, "% 5d / %d\r", i, njob );
for( j=i+1; j<njob; j++ )
{
if( strlen( seq[i] ) == 0 || strlen( seq[j] ) == 0 )
{
distancemtx[i][j] = pscore;
continue;
}
strcpy( aseq[i], seq[i] );
strcpy( aseq[j], seq[j] );
clus1 = conjuctionfortbfast( pair, i, aseq, mseq1, effarr1, effarr, indication1 );
clus2 = conjuctionfortbfast( pair, j, aseq, mseq2, effarr2, effarr, indication2 );
// fprintf( stderr, "mseq1 = %s\n", mseq1[0] );
// fprintf( stderr, "mseq2 = %s\n", mseq2[0] );
#if 0
fprintf( stderr, "group1 = %.66s", indication1 );
fprintf( stderr, "\n" );
fprintf( stderr, "group2 = %.66s", indication2 );
fprintf( stderr, "\n" );
#endif
// for( l=0; l<clus1; l++ ) fprintf( stderr, "## STEP-eff for mseq1-%d %f\n", l, effarr1[l] );
#if 1
if( use_fft )
{
pscore = Falign( mseq1, mseq2, effarr1, effarr2, clus1, clus2, alloclen, &intdum, NULL, 0, NULL );
off1 = off2 = 0;
}
else
#endif
{
switch( alg )
{
case( 'a' ):
pscore = Aalign( mseq1, mseq2, effarr1, effarr2, clus1, clus2, alloclen );
off1 = off2 = 0;
break;
case( 'A' ):
pscore = G__align11( mseq1, mseq2, alloclen, NULL, 0, NULL );
off1 = off2 = 0;
break;
#if 0
case( 'V' ):
pscore = VAalign11( mseq1, mseq2, alloclen, &off1, &off2, localhomtable[i]+j );
fprintf( stderr, "i,j = %d,%d, score = %f\n", i,j, pscore );
break;
case( 'S' ):
fprintf( stderr, "aligning %d-%d\n", i, j );
pscore = suboptalign11( mseq1, mseq2, alloclen, &off1, &off2, localhomtable[i]+j );
fprintf( stderr, "i,j = %d,%d, score = %f\n", i,j, pscore );
break;
#endif
case( 'N' ):
pscore = genL__align11( mseq1, mseq2, alloclen, &off1, &off2 );
// fprintf( stderr, "pscore = %f\n", pscore );
break;
case( 'L' ):
pscore = L__align11( mseq1, mseq2, alloclen, &off1, &off2 );
// fprintf( stderr, "pscore (1) = %f\n", pscore );
// pscore = (float)naivepairscore11( *mseq1, *mseq2, penalty ); // nennnotame
// fprintf( stderr, "pscore (2) = %f\n\n", pscore );
break;
case( 'H' ):
pscore = recallpairfoldalign( mseq1, mseq2, i, j, &off1, &off2, alloclen );
break;
case( 'B' ):
case( 'T' ):
pscore = recalllara( mseq1, mseq2, alloclen );
off1 = off2 = 0;
// fprintf( stderr, "lara, pscore = %f\n", pscore );
break;
case( 's' ):
pscore = callmxscarna( mseq1, mseq2, alloclen );
off1 = off2 = 0;
// fprintf( stderr, "scarna, pscore = %f\n", pscore );
break;
case( 'M' ):
// pscore = MSalign11( mseq1, mseq2, effarr1, effarr2, clus1, clus2, alloclen, NULL, NULL, NULL, NULL );
pscore = MSalign11( mseq1, mseq2, alloclen );
// fprintf( stderr, "pscore = %f\n", pscore );
break;
ErrorExit( "ERROR IN SOURCE FILE" );
}
}
distancemtx[i][j] = pscore;
#if SCOREOUT
fprintf( stderr, "score = %10.2f (%d,%d)\n", pscore, i, j );
#endif
// fprintf( stderr, "pslocal = %d\n", pslocal );
// offset = makelocal( *mseq1, *mseq2, pslocal );
#if 0
fprintf( stderr, "off1 = %d, off2 = %d\n", off1, off2 );
fprintf( stderr, ">%d\n%s\n>%d\n%s\n>\n", i, mseq1[0], j, mseq2[0] );
#endif
// putlocalhom2( mseq1[0], mseq2[0], localhomtable[i]+j, countamino( *mseq1, off1 ), countamino( *mseq2, off2 ), pscore, strlen( mseq1[0] ) );
// fprintf( stderr, "pscore = %f\n", pscore );
if( alg == 'H' )
// if( alg == 'H' || alg == 's' || alg == 'B' ) // next version
putlocalhom_ext( mseq1[0], mseq2[0], localhomtable[i]+j, off1, off2, (int)pscore, strlen( mseq1[0] ) );
else if( alg != 'S' && alg != 'V' )
putlocalhom2( mseq1[0], mseq2[0], localhomtable[i]+j, off1, off2, (int)pscore, strlen( mseq1[0] ) );
}
}
for( i=0; i<njob; i++ )
{
pscore = 0.0;
for( pt=seq[i]; *pt; pt++ )
pscore += amino_dis[(int)*pt][(int)*pt];
distancemtx[i][i] = pscore;
}
ilim = njob-1;
for( i=0; i<ilim; i++ )
{
for( j=i+1; j<njob; j++ )
{
bunbo = MIN( distancemtx[i][i], distancemtx[j][j] );
if( bunbo == 0.0 )
distancemtx[i][j] = 2.0;
else
distancemtx[i][j] = ( 1.0 - distancemtx[i][j] / bunbo ) * 2.0;
}
}
hat2p = fopen( hat2file, "w" );
if( !hat2p ) ErrorExit( "Cannot open hat2." );
WriteHat2( hat2p, njob, name, distancemtx );
fclose( hat2p );
fprintf( stderr, "##### writing hat3\n" );
hat3p = fopen( "hat3", "w" );
if( !hat3p ) ErrorExit( "Cannot open hat3." );
ilim = njob-1;
for( i=0; i<ilim; i++ )
{
for( j=i+1; j<njob; j++ )
{
for( tmpptr=localhomtable[i]+j; tmpptr; tmpptr=tmpptr->next )
{
if( tmpptr->opt == -1.0 ) continue;
fprintf( hat3p, "%d %d %d %7.5f %d %d %d %d %p\n", i, j, tmpptr->overlapaa, tmpptr->opt, tmpptr->start1, tmpptr->end1, tmpptr->start2, tmpptr->end2, (void *)tmpptr->next );
}
}
}
fclose( hat3p );
#if DEBUG
fprintf( stderr, "calling FreeLocalHomTable\n" );
#endif
FreeLocalHomTable( localhomtable, njob );
#if DEBUG
fprintf( stderr, "done. FreeLocalHomTable\n" );
#endif
}
void GroupAlign( int nseq1, int nseq2, char **name, int *nlen, char **seq, char **aseq, char **mseq1, char **mseq2, int ***topol, double **len, double *eff, int alloclen )
{
int i;
int clus1, clus2;
int s1, s2;
float pscore;
static char **name1, **name2;
double *effarr = eff;
double *effarr1 = NULL;
double *effarr2 = NULL;
static char *indication1, *indication2;
float dumfl = 0.0;
int intdum;
#if DEBUG
double time1, time2;
#endif
// fprintf( stderr, "in GroupAlign fftWinSize = %d\n", fftWinSize );
// fprintf( stderr, "in GroupAlign fftThreshold = %d\n", fftThreshold );
if( effarr1 == NULL )
{
name1 = AllocateCharMtx( nseq1, B );
name2 = AllocateCharMtx( nseq2, B );
indication1 = AllocateCharVec( 150 );
indication2 = AllocateCharVec( 150 );
effarr1 = AllocateDoubleVec( njob );
effarr2 = AllocateDoubleVec( njob );
#if 0
#else
#endif
}
for( i=0; i<njob; i++ ) strcpy( aseq[i], seq[i] );
s1 = 0;
s2 = nseq1;
// fprintf( stdout, "nseq1 = %d\n", nseq1 );
clus1 = conjuctionforgaln( 0, nseq1, aseq, mseq1, effarr1, effarr, name, name1, indication1 );
clus2 = conjuctionforgaln( nseq1, njob, aseq, mseq2, effarr2, effarr, name, name2, indication2 );
/*
fprintf( stderr, "before align all\n" );
display( aseq, njob );
fprintf( stderr, "\n" );
fprintf( stderr, "before align 1 %s \n", indication1 );
display( mseq1, clus1 );
fprintf( stderr, "\n" );
fprintf( stderr, "before align 2 %s \n", indication2 );
display( mseq2, clus2 );
fprintf( stderr, "\n" );
*/
commongappick( nseq1, mseq1 );
commongappick( nseq2, mseq2 );
#if DEBUG
time1 = getrusage_sec();
fprintf( stderr, "entering Falign\n" );
#endif
if( use_fft )
{
if( alg == 'M' )
pscore = Falign_udpari_long( NULL, NULL, n_dis_consweight_multi, mseq1, mseq2, effarr1, effarr2, NULL, NULL, clus1, clus2, alloclen, &intdum );
else
pscore = Falign( NULL, NULL, n_dis_consweight_multi, mseq1, mseq2, effarr1, effarr2, NULL, NULL, clus1, clus2, alloclen, &intdum, NULL, 0, NULL );
}
else
{
if( alg == 'M' )
pscore = MSalignmm( n_dis_consweight_multi, mseq1, mseq2, effarr1, effarr2, clus1, clus2, alloclen, NULL, NULL, NULL, NULL, NULL, 0, NULL, outgap, outgap );
else
pscore = A__align( n_dis_consweight_multi, mseq1, mseq2, effarr1, effarr2, clus1, clus2, alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL, NULL, 0, NULL, outgap, outgap );
}
#if DEBUG
time2 = getrusage_sec();
fprintf( stdout, "### %d - %d, %f\n", clus1, clus2, time2-time1 );
fflush( stdout );
#endif
/*
fprintf( stderr, "after align 1 %s \n", indication1 );
display( mseq1, clus1 );
fprintf( stderr, "\n" );
fprintf( stderr, "after align 2 %s \n", indication2 );
display( mseq2, clus2 );
fprintf( stderr, "\n" );
*/
fprintf( stderr, "group-to-group %s /%s %f\n", indication1, indication2, pscore );
if( disp ) display( aseq, njob );
fprintf( stderr, "\n" );
/*
trap = fopen( "pre", "r+" );
if( !trap ) ErrorExit( 1 );
WriteGapFill( trap, njob, name, nlen, aseq );
fclose( trap );
fprintf( stdout, "nseq1 = %d\n", nseq1 );
*/
}
void part_imp_match_init( float *imp, int clus1, int clus2, int lgth1, int lgth2, char **seq1, char **seq2, double *eff1, double *eff2, LocalHom ***localhom )
{
int dif, i, j, k1, k2, tmpint, start1, start2, end1, end2;
static int impalloclen = 0;
char *pt;
static char *nocount1 = NULL;
static char *nocount2 = NULL;
if( impalloclen < lgth1 || impalloclen < lgth2 )
{
if( impmtx ) FreeFloatMtx( impmtx );
if( nocount1 ) free( nocount1 );
if( nocount2 ) free( nocount2 );
impalloclen = MAX( lgth1, lgth2 ) + 2;
impmtx = AllocateFloatMtx( impalloclen, impalloclen );
nocount1 = AllocateCharVec( impalloclen );
nocount2 = AllocateCharVec( impalloclen );
impalloclen -= 2;
}
for( i=0; i<lgth1; i++ )
{
for( j=0; j<clus1; j++ )
if( seq1[j][i] == '-' ) break;
if( j != clus1 ) nocount1[i] = 1;
else nocount1[i] = 0;
}
for( i=0; i<lgth2; i++ )
{
for( j=0; j<clus2; j++ )
if( seq2[j][i] == '-' ) break;
if( j != clus2 ) nocount2[i] = 1;
else nocount2[i] = 0;
}
#if 0
fprintf( stderr, "nocount2 =\n" );
for( i = 0; i<impalloclen; i++ )
{
fprintf( stderr, "nocount2[%d] = %d (%c)\n", i, nocount2[i], seq2[0][i] );
}
#endif
for( i=0; i<lgth1; i++ ) for( j=0; j<lgth2; j++ )
impmtx[i][j] = 0.0;
for( i=0; i<clus1; i++ )
{
fprintf( stderr, "i = %d, seq1 = %s\n", i, seq1[i] );
for( j=0; j<clus2; j++ )
{
fprintf( stderr, "start1 = %d\n", localhom[i][j]->start1 );
fprintf( stderr, "end1 = %d\n", localhom[i][j]->end1 );
fprintf( stderr, "j = %d, seq2 = %s\n", j, seq2[j] );
pt = seq1[i];
tmpint = -1;
while( *pt != 0 )
{
if( *pt++ != '-' ) tmpint++;
if( tmpint == localhom[i][j]->start1 ) break;
}
start1 = pt - seq1[i] - 1;
while( *pt != 0 )
{
// fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, localhom[i][j].end1, pt-seq1[i] );
if( *pt++ != '-' ) tmpint++;
if( tmpint == localhom[i][j]->end1 ) break;
}
end1 = pt - seq1[i] - 1;
pt = seq2[j];
tmpint = -1;
while( *pt != 0 )
{
if( *pt++ != '-' ) tmpint++;
if( tmpint == localhom[i][j]->start2 ) break;
}
start2 = pt - seq2[j] - 1;
while( *pt != 0 )
{
if( *pt++ != '-' ) tmpint++;
if( tmpint == localhom[i][j]->end2 ) break;
}
end2 = pt - seq2[j] - 1;
// fprintf( stderr, "start1 = %d, end1 = %d, start2 = %d, end2 = %d\n", start1, end1, start2, end2 );
k1 = start1;
k2 = start2;
fprintf( stderr, "step 0\n" );
while( k1 <= end1 && k2 <= end2 )
{
#if 0
if( !nocount1[k1] && !nocount2[k2] )
impmtx[k1][k2] += localhom[i][j].wimportance * eff1[i] * eff2[j];
k1++; k2++;
#else
if( !nocount1[k1] && !nocount2[k2] )
impmtx[k1][k2] += localhom[i][j]->wimportance * eff1[i] * eff2[j];
k1++; k2++;
#endif
}
dif = ( end1 - start1 ) - ( end2 - start2 );
fprintf( stderr, "dif = %d\n", dif );
if( dif > 0 )
{
do
{
fprintf( stderr, "dif = %d\n", dif );
k1 = start1;
k2 = start2 - dif;
while( k1 <= end1 && k2 <= end2 )
{
if( 0 <= k2 && start2 <= k2 && !nocount1[k1] && !nocount2[k2] )
impmtx[k1][k2] = localhom[i][j]->wimportance * eff1[i] * eff2[j];
k1++; k2++;
}
}
while( dif-- );
}
else
{
do
{
k1 = start1 + dif;
k2 = start2;
while( k1 <= end1 )
{
if( k1 >= 0 && k1 >= start1 && !nocount1[k1] && !nocount2[k2] )
impmtx[k1][k2] = localhom[i][j]->wimportance * eff1[i] * eff2[j];
k1++; k2++;
}
}
while( dif++ );
}
}
}
#if 0
fprintf( stderr, "impmtx = \n" );
for( k2=0; k2<lgth2; k2++ )
fprintf( stderr, "%6.3f ", (double)k2 );
fprintf( stderr, "\n" );
for( k1=0; k1<lgth1; k1++ )
{
fprintf( stderr, "%d", k1 );
for( k2=0; k2<lgth2; k2++ )
fprintf( stderr, "%6.3f ", impmtx[k1][k2] );
fprintf( stderr, "\n" );
}
exit( 1 );
#endif
}
static void pairalign( char name[M][B], int nlen[M], char **seq, char **aseq, char **mseq1, char **mseq2, double *equiv, double *effarr, char **strfiles, char **chainids, int alloclen )
{
int i, j, ilim;
int clus1, clus2;
int off1, off2;
float pscore = 0.0; // by D.Mathog
static char *indication1, *indication2;
FILE *hat2p, *hat3p;
static double **distancemtx;
static double *effarr1 = NULL;
static double *effarr2 = NULL;
char *pt;
char *hat2file = "hat2";
LocalHom **localhomtable, *tmpptr;
static char **pair;
// int intdum;
double bunbo;
char **checkseq;
localhomtable = (LocalHom **)calloc( njob, sizeof( LocalHom *) );
for( i=0; i<njob; i++)
{
localhomtable[i] = (LocalHom *)calloc( njob, sizeof( LocalHom ) );
for( j=0; j<njob; j++)
{
localhomtable[i][j].start1 = -1;
localhomtable[i][j].end1 = -1;
localhomtable[i][j].start2 = -1;
localhomtable[i][j].end2 = -1;
localhomtable[i][j].opt = -1.0;
localhomtable[i][j].next = NULL;
localhomtable[i][j].nokori = 0;
}
}
if( effarr1 == NULL )
{
distancemtx = AllocateDoubleMtx( njob, njob );
effarr1 = AllocateDoubleVec( njob );
effarr2 = AllocateDoubleVec( njob );
indication1 = AllocateCharVec( 150 );
indication2 = AllocateCharVec( 150 );
checkseq = AllocateCharMtx( njob, alloclen );
#if 0
#else
pair = AllocateCharMtx( njob, njob );
#endif
}
#if 0
fprintf( stderr, "##### fftwinsize = %d, fftthreshold = %d\n", fftWinSize, fftThreshold );
#endif
#if 0
for( i=0; i<njob; i++ )
fprintf( stderr, "TBFAST effarr[%d] = %f\n", i, effarr[i] );
#endif
// writePre( njob, name, nlen, aseq, 0 );
for( i=0; i<njob; i++ ) for( j=0; j<njob; j++ ) pair[i][j] = 0;
for( i=0; i<njob; i++ ) pair[i][i] = 1;
for( i=0; i<njob; i++ )
{
strcpy( checkseq[i], seq[i] );
// fprintf( stderr, "checkseq[%d] = %s\n", i, checkseq[i] );
}
ilim = njob - 1;
for( i=0; i<ilim; i++ )
{
fprintf( stderr, "% 5d / %d\r", i, njob );
for( j=i+1; j<njob; j++ )
{
#if 0
if( strlen( seq[i] ) == 0 || strlen( seq[j] ) == 0 )
{
distancemtx[i][j] = pscore;
continue;
}
#endif
strcpy( aseq[i], seq[i] );
strcpy( aseq[j], seq[j] );
clus1 = conjuctionfortbfast( pair, i, aseq, mseq1, effarr1, effarr, indication1 );
clus2 = conjuctionfortbfast( pair, j, aseq, mseq2, effarr2, effarr, indication2 );
// fprintf( stderr, "mseq1 = %s\n", mseq1[0] );
// fprintf( stderr, "mseq2 = %s\n", mseq2[0] );
#if 0
fprintf( stderr, "group1 = %.66s", indication1 );
fprintf( stderr, "\n" );
fprintf( stderr, "group2 = %.66s", indication2 );
fprintf( stderr, "\n" );
#endif
// for( l=0; l<clus1; l++ ) fprintf( stderr, "## STEP-eff for mseq1-%d %f\n", l, effarr1[l] );
#if 1
{
switch( alg )
{
case( 'T' ):
fprintf( stderr, " Calling tmalign %d-%d/%d \r", i+1, j+1, njob );
pscore = calltmalign( mseq1, mseq2, equiv, strfiles[i], chainids[i], strfiles[j], chainids[j], alloclen );
off1 = off2 = 0;
break;
case( 'R' ):
fprintf( stderr, " Calling PDP_ASH.pl %d-%d/%d \r", i+1, j+1, njob );
pscore = callrash( i, j, mseq1, mseq2, equiv, strfiles[i], chainids[i], strfiles[j], chainids[j], alloclen );
off1 = off2 = 0;
break;
ErrorExit( "ERROR IN SOURCE FILE" );
}
}
#endif
distancemtx[i][j] = pscore;
#if SCOREOUT
fprintf( stderr, "score = %10.2f (%d,%d)\n", pscore, i, j );
#endif
putlocalhom_str( mseq1[0], mseq2[0], equiv, scale, localhomtable[i]+j, off1, off2, (int)pscore, strlen( mseq1[0] ) );
#if 1
if( alreadyoutput[i] == 0 )
{
alreadyoutput[i] = 1;
gappick0( seq[i], mseq1[0] );
fprintf( stdout, ">%d_%s-%s\n%s\n", i+1, strfiles[i], chainids[i], seq[i] );
strcpy( checkseq[i], seq[i] );
}
else
{
gappick0( seq[i], mseq1[0] );
fprintf( stderr, "checking seq%d\n", i );
// fprintf( stderr, " seq=%s\n", seq[i] );
// fprintf( stderr, "checkseq=%s\n", checkseq[i] );
if( strcmp( checkseq[i], seq[i] ) )
{
fprintf( stderr, "\n\nWARNING: Sequence changed!!\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, " seq=%s\n", seq[i] );
fprintf( stderr, "checkseq=%s\n", checkseq[i] );
exit( 1 );
}
}
if( alreadyoutput[j] == 0 )
{
alreadyoutput[j] = 1;
gappick0( seq[j], mseq2[0] );
fprintf( stdout, ">%d_%s-%s\n%s\n", j+1, strfiles[j], chainids[j], seq[j] );
strcpy( checkseq[j], seq[j] );
}
else
{
gappick0( seq[j], mseq2[0] );
fprintf( stderr, "checking seq%d\n", j );
if( strcmp( checkseq[j], seq[j] ) )
{
fprintf( stderr, "\n\nWARNING: Sequence changed!!\n" );
fprintf( stderr, "j=%d\n", j );
fprintf( stderr, " seq=%s\n", seq[j] );
fprintf( stderr, "checkseq=%s\n", checkseq[j] );
exit( 1 );
}
}
#endif
}
}
for( i=0; i<njob; i++ )
{
pscore = 0.0;
for( pt=seq[i]; *pt; pt++ )
pscore += amino_dis[(int)*pt][(int)*pt];
distancemtx[i][i] = pscore;
}
ilim = njob-1;
for( i=0; i<ilim; i++ )
{
for( j=i+1; j<njob; j++ )
{
bunbo = MIN( distancemtx[i][i], distancemtx[j][j] );
if( bunbo == 0.0 )
distancemtx[i][j] = 2.0;
else
distancemtx[i][j] = ( 1.0 - distancemtx[i][j] / bunbo ) * 2.0;
}
}
hat2p = fopen( hat2file, "w" );
if( !hat2p ) ErrorExit( "Cannot open hat2." );
WriteHat2( hat2p, njob, name, distancemtx );
fclose( hat2p );
fprintf( stderr, "##### writing hat3\n" );
hat3p = fopen( "hat3", "w" );
if( !hat3p ) ErrorExit( "Cannot open hat3." );
ilim = njob-1;
for( i=0; i<ilim; i++ )
{
for( j=i+1; j<njob; j++ )
{
for( tmpptr=localhomtable[i]+j; tmpptr; tmpptr=tmpptr->next )
{
if( tmpptr->opt == -1.0 ) continue;
fprintf( hat3p, "%d %d %d %7.5f %d %d %d %d k\n", i, j, tmpptr->overlapaa, tmpptr->opt, tmpptr->start1, tmpptr->end1, tmpptr->start2, tmpptr->end2 );
}
}
}
fclose( hat3p );
#if DEBUG
fprintf( stderr, "calling FreeLocalHomTable\n" );
#endif
FreeLocalHomTable( localhomtable, njob );
#if DEBUG
fprintf( stderr, "done. FreeLocalHomTable\n" );
#endif
}
static void prepareash( FILE *fp, char ***strfiles, char ***chainids, char ***seqpt, char ***mseq1pt, char ***mseq2pt, double **equivpt, int *alloclenpt )
{
int i, res;
char *dumseq;
char line[1000];
char fname[1000];
char command[1000];
int linenum, istr, nstr;
FILE *checkfp;
char *sline;
int use[1000];
linenum = 0;
nstr = 0;
while( 1 )
{
fgets( line, 999, fp );
if( feof( fp ) ) break;
sline = strip( line );
use[linenum] = 1;
if( sline[0] == '#' || strlen( sline ) < 2 )
{
use[linenum] = 0;
linenum++;
continue;
}
extractfirstword( sline );
checkfp = fopen( sline, "r" );
if( checkfp == NULL )
{
fprintf( stderr, "Cannot open %s.\n", sline );
exit( 1 );
}
#if 0
fgets( linec, 999, checkfp );
if( strncmp( "HEADER ", linec, 7 ) )
{
fprintf( stderr, "Check the format of %s.\n", sline );
exit( 1 );
}
#endif
if( checkcbeta( checkfp ) )
{
fprintf( stderr, "%s has no C-beta atoms.\n", sline );
exit( 1 );
}
else
nstr++;
fclose( checkfp );
linenum++;
}
njob = nstr;
fprintf( stderr, "nstr = %d\n", nstr );
*strfiles = AllocateCharMtx( nstr, 1000 );
*chainids = AllocateCharMtx( nstr, 2 );
rewind( fp );
istr = 0;
linenum = 0;
while( 1 )
{
fgets( line, 999, fp );
if( feof( fp ) ) break;
sline = strip( line );
if( use[linenum++] )
{
(*chainids)[istr][0] = getchainid( sline );
(*chainids)[istr][1] = 0;
extractfirstword( sline );
sprintf( fname, "%s", sline );
cutpath( fname );
sprintf( command, "cp %s %s.pdb", sline, fname );
system( command );
sprintf( command, "perl \"%s/clean.pl\" %s.pdb", whereispairalign, fname );
res = system( command );
if( res )
{
fprintf( stderr, "error: Install clean.pl\n" );
exit( 1 );
}
strcpy( (*strfiles)[istr++], fname );
}
}
*seqpt = AllocateCharMtx( njob, nlenmax*2+1 );
*mseq1pt = AllocateCharMtx( njob, 0 );
*mseq2pt = AllocateCharMtx( njob, 0 );
*equivpt = AllocateDoubleVec( nlenmax*2+1 );
*alloclenpt = nlenmax*2;
dumseq = AllocateCharVec( nlenmax*2+1 );
alreadyoutput = AllocateIntVec( njob );
for( i=0; i<njob; i++ ) alreadyoutput[i] = 0;
for( i=0; i<istr; i++ )
{
fprintf( stderr, "i=%d\n", i );
(*seqpt)[i][0] = 0;
(*mseq1pt)[0] = (*seqpt)[i];
(*mseq2pt)[0] = dumseq;
callrash( i, i, *mseq1pt, *mseq2pt, *equivpt, (*strfiles)[i], (*chainids)[i], (*strfiles)[i], (*chainids)[i], *alloclenpt );
fprintf( stdout, ">%d_%s-%s\n%s\n", i+1, (*strfiles)[i], (*chainids)[i], (*seqpt)[i] );
alreadyoutput[i] = 1;
}
}
static void pairalign( char name[M][B], int nlen[M], char **seq, double *effarr, int alloclen )
{
int i, j;
FILE *hat3p;
float pscore = 0.0; // by D.Mathog
static double *effarr1 = NULL;
static double *effarr2 = NULL;
char *aseq;
static char **pseq;
LocalHom **localhomtable, *tmpptr;
double tsuyosa;
if( nhomologs < 1 ) nhomologs = 1; // tsuyosa=0.0 wo sakeru
tsuyosa = (double)nhomologs * nhomologs * TSUYOSAFACTOR;
fprintf( stderr, "tsuyosa = %f\n", tsuyosa );
localhomtable = (LocalHom **)calloc( njob, sizeof( LocalHom *) );
for( i=0; i<njob; i++)
{
localhomtable[i] = (LocalHom *)calloc( njob, sizeof( LocalHom ) );
for( j=0; j<njob; j++)
{
localhomtable[i][j].start1 = -1;
localhomtable[i][j].end1 = -1;
localhomtable[i][j].start2 = -1;
localhomtable[i][j].end2 = -1;
localhomtable[i][j].opt = -1.0;
localhomtable[i][j].next = NULL;
}
}
if( effarr1 == NULL )
{
effarr1 = AllocateDoubleVec( njob );
effarr2 = AllocateDoubleVec( njob );
pseq = AllocateCharMtx( 2, 0 );
aseq = AllocateCharVec( nlenmax*9+1 );
#if 0
#else
#endif
}
#if 0
fprintf( stderr, "##### fftwinsize = %d, fftthreshold = %d\n", fftWinSize, fftThreshold );
#endif
#if 0
for( i=0; i<njob; i++ )
fprintf( stderr, "TBFAST effarr[%d] = %f\n", i, effarr[i] );
#endif
// writePre( njob, name, nlen, aseq, 0 );
hat3p = fopen( "hat3", "w" );
if( !hat3p ) ErrorExit( "Cannot open hat3." );
fprintf( stderr, "\n" );
for( i=0; i<njob-1; i++ )
{
for( j=i+1; j<njob; j++ )
{
pseq[0] = seq[i];
pseq[1] = seq[j];
if( strlen( pseq[0] ) != strlen( pseq[1] ) )
{
fprintf( stderr, "## ERROR ###\n" );
fprintf( stderr, "Not aligned, %s - %s\n", name[i], name[j] );
fprintf( stderr, "## ERROR ###\n" );
exit( 1 );
}
fprintf( stderr, "adding %d-%d\r", i, j );
putlocalhom2( pseq[0], pseq[1], localhomtable[i]+j, 0, 0, (int)pscore, strlen( pseq[0] ) );
for( tmpptr=localhomtable[i]+j; tmpptr; tmpptr=tmpptr->next )
{
if( tmpptr->opt == -1.0 ) continue;
if( tmpptr->start1 == -1 ) continue;
fprintf( hat3p, "%d %d %d %6.3f %d %d %d %d k\n", i+seedoffset, j+seedoffset, tmpptr->overlapaa, tmpptr->opt * tsuyosa, tmpptr->start1, tmpptr->end1, tmpptr->start2, tmpptr->end2 );
}
}
}
fprintf( stderr, "\n" );
fclose( hat3p );
#if DEBUG
fprintf( stderr, "calling FreeLocalHomTable\n" );
#endif
FreeLocalHomTable( localhomtable, njob );
#if DEBUG
fprintf( stderr, "done. FreeLocalHomTable\n" );
#endif
}
float Falign_localhom( char **seq1, char **seq2,
double *eff1, double *eff2,
int clus1, int clus2,
int alloclen,
LocalHom ***localhom, float *totalimpmatch,
int *gapmap1, int *gapmap2,
int *chudanpt, int chudanref, int *chudanres )
{
// tditeration.c deha alloclen ha huhen nanode
// prevalloclen ha iranai.
int i, j, k, l, m, maxk;
int nlen, nlen2, nlen4;
static TLS int crossscoresize = 0;
static TLS char **tmpseq1 = NULL;
static TLS char **tmpseq2 = NULL;
static TLS char **tmpptr1 = NULL;
static TLS char **tmpptr2 = NULL;
static TLS char **tmpres1 = NULL;
static TLS char **tmpres2 = NULL;
static TLS char **result1 = NULL;
static TLS char **result2 = NULL;
#if RND
static TLS char **rndseq1 = NULL;
static TLS char **rndseq2 = NULL;
#endif
static TLS Fukusosuu **seqVector1 = NULL;
static TLS Fukusosuu **seqVector2 = NULL;
static TLS Fukusosuu **naiseki = NULL;
static TLS Fukusosuu *naisekiNoWa = NULL;
static TLS double *soukan = NULL;
static TLS double **crossscore = NULL;
int nlentmp;
static TLS int *kouho = NULL;
static TLS Segment *segment = NULL;
static TLS Segment *segment1 = NULL;
static TLS Segment *segment2 = NULL;
static TLS Segment **sortedseg1 = NULL;
static TLS Segment **sortedseg2 = NULL;
static TLS int *cut1 = NULL;
static TLS int *cut2 = NULL;
static TLS char *sgap1, *egap1, *sgap2, *egap2;
static TLS int localalloclen = 0;
int lag;
int tmpint;
int count, count0;
int len1, len2;
int totallen;
float totalscore;
float impmatch;
extern Fukusosuu *AllocateFukusosuuVec();
extern Fukusosuu **AllocateFukusosuuMtx();
if( seq1 == NULL )
{
if( result1 )
{
// fprintf( stderr, "Freeing localarrays in Falign\n" );
localalloclen = 0;
mymergesort( 0, 0, NULL );
alignableReagion( 0, 0, NULL, NULL, NULL, NULL, NULL );
fft( 0, NULL, 1 );
A__align( NULL, NULL, NULL, NULL, 0, 0, 0, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 0, NULL, 0, 0 );
G__align11( NULL, NULL, 0, 0, 0 );
partA__align( NULL, NULL, NULL, NULL, 0, 0, 0, NULL, NULL, 0, 0, 0, 0, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 0, NULL );
blockAlign2( NULL, NULL, NULL, NULL, NULL, NULL );
if( crossscore ) FreeDoubleMtx( crossscore );
FreeCharMtx( result1 );
FreeCharMtx( result2 );
FreeCharMtx( tmpres1 );
FreeCharMtx( tmpres2 );
FreeCharMtx( tmpseq1 );
FreeCharMtx( tmpseq2 );
free( sgap1 );
free( egap1 );
free( sgap2 );
free( egap2 );
free( kouho );
free( cut1 );
free( cut2 );
free( tmpptr1 );
free( tmpptr2 );
free( segment );
free( segment1 );
free( segment2 );
free( sortedseg1 );
free( sortedseg2 );
if( !kobetsubunkatsu )
{
FreeFukusosuuMtx ( seqVector1 );
FreeFukusosuuMtx ( seqVector2 );
FreeFukusosuuVec( naisekiNoWa );
FreeFukusosuuMtx( naiseki );
FreeDoubleVec( soukan );
}
}
else
{
// fprintf( stderr, "Did not allocate localarrays in Falign\n" );
}
return( 0.0 );
}
len1 = strlen( seq1[0] );
len2 = strlen( seq2[0] );
nlentmp = MAX( len1, len2 );
nlen = 1;
while( nlentmp >= nlen ) nlen <<= 1;
#if 0
fprintf( stderr, "### nlen = %d\n", nlen );
#endif
nlen2 = nlen/2; nlen4 = nlen2 / 2;
#if DEBUG
fprintf( stderr, "len1 = %d, len2 = %d\n", len1, len2 );
fprintf( stderr, "nlentmp = %d, nlen = %d\n", nlentmp, nlen );
#endif
if( !localalloclen )
{
sgap1 = AllocateCharVec( njob );
egap1 = AllocateCharVec( njob );
sgap2 = AllocateCharVec( njob );
egap2 = AllocateCharVec( njob );
kouho = AllocateIntVec( NKOUHO );
cut1 = AllocateIntVec( MAXSEG );
cut2 = AllocateIntVec( MAXSEG );
tmpptr1 = AllocateCharMtx( njob, 0 );
tmpptr2 = AllocateCharMtx( njob, 0 );
result1 = AllocateCharMtx( njob, alloclen );
result2 = AllocateCharMtx( njob, alloclen );
tmpres1 = AllocateCharMtx( njob, alloclen );
tmpres2 = AllocateCharMtx( njob, alloclen );
// crossscore = AllocateDoubleMtx( MAXSEG, MAXSEG );
segment = (Segment *)calloc( MAXSEG, sizeof( Segment ) );
segment1 = (Segment *)calloc( MAXSEG, sizeof( Segment ) );
segment2 = (Segment *)calloc( MAXSEG, sizeof( Segment ) );
sortedseg1 = (Segment **)calloc( MAXSEG, sizeof( Segment * ) );
sortedseg2 = (Segment **)calloc( MAXSEG, sizeof( Segment * ) );
if( !( segment && segment1 && segment2 && sortedseg1 && sortedseg2 ) )
ErrorExit( "Allocation error\n" );
if ( scoremtx == -1 ) n20or4or2 = 4;
else if( fftscore == 1 ) n20or4or2 = 2;
else n20or4or2 = 20;
}
if( localalloclen < nlen )
{
if( localalloclen )
{
#if 1
if( !kobetsubunkatsu )
{
FreeFukusosuuMtx ( seqVector1 );
FreeFukusosuuMtx ( seqVector2 );
FreeFukusosuuVec( naisekiNoWa );
FreeFukusosuuMtx( naiseki );
FreeDoubleVec( soukan );
}
FreeCharMtx( tmpseq1 );
FreeCharMtx( tmpseq2 );
#endif
#if RND
FreeCharMtx( rndseq1 );
FreeCharMtx( rndseq2 );
#endif
}
tmpseq1 = AllocateCharMtx( njob, nlen );
tmpseq2 = AllocateCharMtx( njob, nlen );
if( !kobetsubunkatsu )
{
naisekiNoWa = AllocateFukusosuuVec( nlen );
naiseki = AllocateFukusosuuMtx( n20or4or2, nlen );
seqVector1 = AllocateFukusosuuMtx( n20or4or2+1, nlen+1 );
seqVector2 = AllocateFukusosuuMtx( n20or4or2+1, nlen+1 );
soukan = AllocateDoubleVec( nlen+1 );
}
#if RND
rndseq1 = AllocateCharMtx( njob, nlen );
rndseq2 = AllocateCharMtx( njob, nlen );
for( i=0; i<njob; i++ )
{
generateRndSeq( rndseq1[i], nlen );
generateRndSeq( rndseq2[i], nlen );
}
#endif
localalloclen = nlen;
}
for( j=0; j<clus1; j++ ) strcpy( tmpseq1[j], seq1[j] );
for( j=0; j<clus2; j++ ) strcpy( tmpseq2[j], seq2[j] );
#if 0
fftfp = fopen( "input_of_Falign", "w" );
fprintf( fftfp, "nlen = %d\n", nlen );
fprintf( fftfp, "seq1: ( %d sequences ) \n", clus1 );
for( i=0; i<clus1; i++ )
fprintf( fftfp, "%s\n", seq1[i] );
fprintf( fftfp, "seq2: ( %d sequences ) \n", clus2 );
for( i=0; i<clus2; i++ )
fprintf( fftfp, "%s\n", seq2[i] );
fclose( fftfp );
system( "less input_of_Falign < /dev/tty > /dev/tty" );
#endif
if( !kobetsubunkatsu )
{
fprintf( stderr, "FFT ... " );
for( j=0; j<n20or4or2; j++ ) vec_init( seqVector1[j], nlen );
if( fftscore && scoremtx != -1 )
{
for( i=0; i<clus1; i++ )
{
seq_vec_2( seqVector1[0], polarity, eff1[i], tmpseq1[i] );
seq_vec_2( seqVector1[1], volume, eff1[i], tmpseq1[i] );
}
}
else
{
#if 0
for( i=0; i<clus1; i++ ) for( j=0; j<n20or4or2; j++ )
seq_vec( seqVector1[j], amino[j], eff1[i], tmpseq1[i] );
#else
for( i=0; i<clus1; i++ )
seq_vec_3( seqVector1, eff1[i], tmpseq1[i] );
#endif
}
#if RND
for( i=0; i<clus1; i++ )
{
vec_init2( seqVector1, rndseq1[i], eff1[i], len1, nlen );
}
#endif
#if 0
fftfp = fopen( "seqVec", "w" );
fprintf( fftfp, "before transform\n" );
for( k=0; k<n20or4or2; k++ )
{
fprintf( fftfp, "nlen=%d\n", nlen );
fprintf( fftfp, "%c\n", amino[k] );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%f %f\n", seqVector1[k][l].R, seqVector1[k][l].I );
}
fclose( fftfp );
system( "less seqVec < /dev/tty > /dev/tty" );
#endif
for( j=0; j<n20or4or2; j++ ) vec_init( seqVector2[j], nlen );
if( fftscore && scoremtx != -1 )
{
for( i=0; i<clus2; i++ )
{
seq_vec_2( seqVector2[0], polarity, eff2[i], tmpseq2[i] );
seq_vec_2( seqVector2[1], volume, eff2[i], tmpseq2[i] );
}
}
else
{
#if 0
for( i=0; i<clus2; i++ ) for( j=0; j<n20or4or2; j++ )
seq_vec( seqVector2[j], amino[j], eff2[i], tmpseq2[i] );
#else
for( i=0; i<clus2; i++ )
seq_vec_3( seqVector2, eff2[i], tmpseq2[i] );
#endif
}
#if RND
for( i=0; i<clus2; i++ )
{
vec_init2( seqVector2, rndseq2[i], eff2[i], len2, nlen );
}
#endif
#if 0
fftfp = fopen( "seqVec2", "w" );
fprintf( fftfp, "before fft\n" );
for( k=0; k<n20or4or2; k++ )
{
fprintf( fftfp, "%c\n", amino[k] );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%f %f\n", seqVector2[k][l].R, seqVector2[k][l].I );
}
fclose( fftfp );
system( "less seqVec2 < /dev/tty > /dev/tty" );
#endif
for( j=0; j<n20or4or2; j++ )
{
fft( nlen, seqVector2[j], (j==0) );
fft( nlen, seqVector1[j], 0 );
}
#if 0
fftfp = fopen( "seqVec2", "w" );
fprintf( fftfp, "#after fft\n" );
for( k=0; k<n20or4or2; k++ )
{
fprintf( fftfp, "#%c\n", amino[k] );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%f %f\n", seqVector2[k][l].R, seqVector2[k][l].I );
}
fclose( fftfp );
system( "less seqVec2 < /dev/tty > /dev/tty" );
#endif
for( k=0; k<n20or4or2; k++ )
{
for( l=0; l<nlen; l++ )
calcNaiseki( naiseki[k]+l, seqVector1[k]+l, seqVector2[k]+l );
}
for( l=0; l<nlen; l++ )
{
naisekiNoWa[l].R = 0.0;
naisekiNoWa[l].I = 0.0;
for( k=0; k<n20or4or2; k++ )
{
naisekiNoWa[l].R += naiseki[k][l].R;
naisekiNoWa[l].I += naiseki[k][l].I;
}
}
#if 0
fftfp = fopen( "naisekiNoWa", "w" );
fprintf( fftfp, "#Before fft\n" );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%d %f %f\n", l, naisekiNoWa[l].R, naisekiNoWa[l].I );
fclose( fftfp );
system( "less naisekiNoWa < /dev/tty > /dev/tty " );
#endif
fft( -nlen, naisekiNoWa, 0 );
for( m=0; m<=nlen2; m++ )
soukan[m] = naisekiNoWa[nlen2-m].R;
for( m=nlen2+1; m<nlen; m++ )
soukan[m] = naisekiNoWa[nlen+nlen2-m].R;
#if 0
fftfp = fopen( "naisekiNoWa", "w" );
fprintf( fftfp, "#After fft\n" );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%d %f\n", l, naisekiNoWa[l].R );
fclose( fftfp );
fftfp = fopen( "list.plot", "w" );
fprintf( fftfp, "plot 'naisekiNoWa'\npause -1" );
fclose( fftfp );
system( "/usr/bin/gnuplot list.plot &" );
#endif
#if 0
fprintf( stderr, "frt write start\n" );
fftfp = fopen( "frt", "w" );
for( l=0; l<nlen; l++ )
fprintf( fftfp, "%d %f\n", l-nlen2, soukan[l] );
fclose( fftfp );
system( "less frt < /dev/tty > /dev/tty" );
#if 0
fftfp = fopen( "list.plot", "w" );
fprintf( fftfp, "plot 'frt'\n pause +1" );
fclose( fftfp );
system( "/usr/bin/gnuplot list.plot" );
#endif
#endif
getKouho( kouho, NKOUHO, soukan, nlen );
#if 0
for( i=0; i<NKOUHO; i++ )
{
fprintf( stderr, "kouho[%d] = %d\n", i, kouho[i] );
}
#endif
}
#if KEIKA
fprintf( stderr, "Searching anchors ... " );
#endif
count = 0;
#define CAND 0
#if CAND
fftfp = fopen( "cand", "w" );
fclose( fftfp );
#endif
if( kobetsubunkatsu )
{
maxk = 1;
kouho[0] = 0;
}
else
{
maxk = NKOUHO;
}
for( k=0; k<maxk; k++ )
{
lag = kouho[k];
zurasu2( lag, clus1, clus2, seq1, seq2, tmpptr1, tmpptr2 );
#if CAND
fftfp = fopen( "cand", "a" );
fprintf( fftfp, "Candidate No.%d lag = %d\n", k+1, lag );
fprintf( fftfp, "%s\n", tmpptr1[0] );
fprintf( fftfp, "%s\n", tmpptr2[0] );
fclose( fftfp );
#endif
tmpint = alignableReagion( clus1, clus2, tmpptr1, tmpptr2, eff1, eff2, segment+count );
if( count+tmpint > MAXSEG -3 ) ErrorExit( "TOO MANY SEGMENTS.\n" );
while( tmpint-- > 0 )
{
if( lag > 0 )
{
segment1[count].start = segment[count].start ;
segment1[count].end = segment[count].end ;
segment1[count].center = segment[count].center;
segment1[count].score = segment[count].score;
segment2[count].start = segment[count].start + lag;
segment2[count].end = segment[count].end + lag;
segment2[count].center = segment[count].center + lag;
segment2[count].score = segment[count].score ;
}
else
{
segment1[count].start = segment[count].start - lag;
segment1[count].end = segment[count].end - lag;
segment1[count].center = segment[count].center - lag;
segment1[count].score = segment[count].score ;
segment2[count].start = segment[count].start ;
segment2[count].end = segment[count].end ;
segment2[count].center = segment[count].center;
segment2[count].score = segment[count].score ;
}
#if 0
fftfp = fopen( "cand", "a" );
fprintf( fftfp, "Goukaku=%dko\n", tmpint );
fprintf( fftfp, "in 1 %d\n", segment1[count].center );
fprintf( fftfp, "in 2 %d\n", segment2[count].center );
fclose( fftfp );
#endif
segment1[count].pair = &segment2[count];
segment2[count].pair = &segment1[count];
count++;
#if 0
fprintf( stderr, "count=%d\n", count );
#endif
}
}
#if 1
if( !kobetsubunkatsu )
fprintf( stderr, "%d segments found\n", count );
#endif
if( !count && fftNoAnchStop )
ErrorExit( "Cannot detect anchor!" );
#if 0
fftfp = fopen( "fft", "a" );
fprintf( fftfp, "RESULT before sort:\n" );
for( l=0; l<count; l++ )
{
fprintf( fftfp, "cut[%d]=%d, ", l, segment1[l].center );
fprintf( fftfp, "%d score = %f\n", segment2[l].center, segment1[l].score );
}
fclose( fftfp );
#endif
#if KEIKA
fprintf( stderr, "Aligning anchors ... " );
#endif
for( i=0; i<count; i++ )
{
sortedseg1[i] = &segment1[i];
sortedseg2[i] = &segment2[i];
}
#if 0
tmpsort( count, sortedseg1 );
tmpsort( count, sortedseg2 );
qsort( sortedseg1, count, sizeof( Segment * ), segcmp );
qsort( sortedseg2, count, sizeof( Segment * ), segcmp );
#else
mymergesort( 0, count-1, sortedseg1 );
mymergesort( 0, count-1, sortedseg2 );
#endif
for( i=0; i<count; i++ ) sortedseg1[i]->number = i;
for( i=0; i<count; i++ ) sortedseg2[i]->number = i;
if( kobetsubunkatsu )
{
for( i=0; i<count; i++ )
{
cut1[i+1] = sortedseg1[i]->center;
cut2[i+1] = sortedseg2[i]->center;
}
cut1[0] = 0;
cut2[0] = 0;
cut1[count+1] = len1;
cut2[count+1] = len2;
count += 2;
}
else
{
if( crossscoresize < count+2 )
{
crossscoresize = count+2;
#if 1
fprintf( stderr, "######allocating crossscore, size = %d\n", crossscoresize );
#endif
if( crossscore ) FreeDoubleMtx( crossscore );
crossscore = AllocateDoubleMtx( crossscoresize, crossscoresize );
}
for( i=0; i<count+2; i++ ) for( j=0; j<count+2; j++ )
crossscore[i][j] = 0.0;
for( i=0; i<count; i++ )
{
crossscore[segment1[i].number+1][segment1[i].pair->number+1] = segment1[i].score;
cut1[i+1] = sortedseg1[i]->center;
cut2[i+1] = sortedseg2[i]->center;
}
#if DEBUG
fprintf( stderr, "AFTER SORT\n" );
for( i=0; i<count; i++ ) fprintf( stderr, "%d, %d\n", segment1[i].start, segment2[i].start );
#endif
crossscore[0][0] = 10000000.0;
cut1[0] = 0;
cut2[0] = 0;
crossscore[count+1][count+1] = 10000000.0;
cut1[count+1] = len1;
cut2[count+1] = len2;
count += 2;
count0 = count;
blockAlign2( cut1, cut2, sortedseg1, sortedseg2, crossscore, &count );
if( count0 > count )
{
#if 0
fprintf( stderr, "\7 REPEAT!? \n" );
#else
fprintf( stderr, "REPEAT!? \n" );
#endif
if( fftRepeatStop ) exit( 1 );
}
#if KEIKA
else fprintf( stderr, "done\n" );
#endif
}
#if 0
fftfp = fopen( "fft", "a" );
fprintf( fftfp, "RESULT after sort:\n" );
for( l=0; l<count; l++ )
{
fprintf( fftfp, "cut[%d]=%d, ", l, segment1[l].center );
fprintf( fftfp, "%d\n", segment2[l].center );
}
fclose( fftfp );
#endif
#if 0
fftfp = fopen( "fft", "a" );
fprintf( fftfp, "RESULT after sort:\n" );
for( l=0; l<count; l++ )
{
fprintf( fftfp, "cut : %d %d\n", cut1[l], cut2[l] );
}
fclose( fftfp );
#endif
#if KEIKA
fprintf( trap_g, "Devided to %d segments\n", count-1 );
fprintf( trap_g, "%d %d forg\n", MIN( clus1, clus2 ), count-1 );
#endif
totallen = 0;
for( j=0; j<clus1; j++ ) result1[j][0] = 0;
for( j=0; j<clus2; j++ ) result2[j][0] = 0;
totalscore = 0.0;
*totalimpmatch = 0.0;
for( i=0; i<count-1; i++ )
{
#if DEBUG
fprintf( stderr, "DP %03d / %03d %4d to ", i+1, count-1, totallen );
#else
#if KEIKA
fprintf( stderr, "DP %03d / %03d\r", i+1, count-1 );
#endif
#endif
if( cut1[i] )
{
getkyokaigap( sgap1, seq1, cut1[i]-1, clus1 );
getkyokaigap( sgap2, seq2, cut2[i]-1, clus2 );
}
else
{
for( j=0; j<clus1; j++ ) sgap1[j] = 'o';
for( j=0; j<clus2; j++ ) sgap2[j] = 'o';
}
if( cut1[i+1] != len1 )
{
getkyokaigap( egap1, seq1, cut1[i+1], clus1 );
getkyokaigap( egap2, seq2, cut2[i+1], clus2 );
}
else
{
for( j=0; j<clus1; j++ ) egap1[j] = 'o';
for( j=0; j<clus2; j++ ) egap2[j] = 'o';
}
for( j=0; j<clus1; j++ )
{
strncpy( tmpres1[j], seq1[j]+cut1[i], cut1[i+1]-cut1[i] );
tmpres1[j][cut1[i+1]-cut1[i]] = 0;
}
if( kobetsubunkatsu ) commongappick_record( clus1, tmpres1, gapmap1 );
for( j=0; j<clus2; j++ )
{
strncpy( tmpres2[j], seq2[j]+cut2[i], cut2[i+1]-cut2[i] );
tmpres2[j][cut2[i+1]-cut2[i]] = 0;
}
if( kobetsubunkatsu ) commongappick_record( clus2, tmpres2, gapmap2 );
#if 0
fprintf( stderr, "count = %d\n", count );
fprintf( stderr, "### reg1 = %d-%d\n", cut1[i], cut1[i+1]-1 );
fprintf( stderr, "### reg2 = %d-%d\n", cut2[i], cut2[i+1]-1 );
#endif
switch( alg )
{
case( 'a' ):
totalscore += Aalign( tmpres1, tmpres2, eff1, eff2, clus1, clus2, alloclen );
break;
case( 'Q' ):
totalscore += partQ__align( tmpres1, tmpres2, eff1, eff2, clus1, clus2, alloclen, localhom, &impmatch, cut1[i], cut1[i+1]-1, cut2[i], cut2[i+1]-1, gapmap1, gapmap2, sgap1, sgap2, egap1, egap2 );
*totalimpmatch += impmatch;
// fprintf( stderr, "*totalimpmatch in Falign_localhom = %f\n", *totalimpmatch );
break;
case( 'A' ):
totalscore += partA__align( tmpres1, tmpres2, eff1, eff2, clus1, clus2, alloclen, localhom, &impmatch, cut1[i], cut1[i+1]-1, cut2[i], cut2[i+1]-1, gapmap1, gapmap2, sgap1, sgap2, egap1, egap2, chudanpt, chudanref, chudanres );
*totalimpmatch += impmatch;
// fprintf( stderr, "*totalimpmatch in Falign_localhom = %f\n", *totalimpmatch );
break;
default:
fprintf( stderr, "alg = %c\n", alg );
ErrorExit( "ERROR IN SOURCE FILE Falign.c" );
break;
}
#ifdef enablemultithread
if( chudanres && *chudanres )
{
// fprintf( stderr, "\n\n## CHUUDAN!!! at Falign_localhom\n" );
return( -1.0 );
}
#endif
nlen = strlen( tmpres1[0] );
if( totallen + nlen > alloclen )
{
fprintf( stderr, "totallen=%d + nlen=%d > alloclen = %d\n", totallen, nlen, alloclen );
ErrorExit( "LENGTH OVER in Falign\n " );
}
for( j=0; j<clus1; j++ ) strcat( result1[j], tmpres1[j] );
for( j=0; j<clus2; j++ ) strcat( result2[j], tmpres2[j] );
totallen += nlen;
#if 0
fprintf( stderr, "%4d\r", totallen );
fprintf( stderr, "\n\n" );
for( j=0; j<clus1; j++ )
{
fprintf( stderr, "%s\n", tmpres1[j] );
}
fprintf( stderr, "-------\n" );
for( j=0; j<clus2; j++ )
{
fprintf( stderr, "%s\n", tmpres2[j] );
}
#endif
}
#if KEIKA
fprintf( stderr, "DP ... done \n" );
#endif
for( j=0; j<clus1; j++ ) strcpy( seq1[j], result1[j] );
for( j=0; j<clus2; j++ ) strcpy( seq2[j], result2[j] );
#if 0
for( j=0; j<clus1; j++ )
{
fprintf( stderr, "%s\n", result1[j] );
}
fprintf( stderr, "- - - - - - - - - - -\n" );
for( j=0; j<clus2; j++ )
{
fprintf( stderr, "%s\n", result2[j] );
}
#endif
return( totalscore );
}
