int main( int ac, char **av )
{
int *nlen;
static char **name, **seq;
double score;
extern double score_calc_for_score( int, char ** );
arguments( ac, av );
getnumlen( stdin );
rewind( stdin );
nlen = AllocateIntVec( njob );
name = AllocateCharMtx( njob, B+1 );
seq = AllocateCharMtx( njob, nlenmax+2 );
readData_pointer( stdin, name, nlen, seq );
if( !isaligned( njob, seq ) ) ErrorExit( "Not aligned." );
constants( njob, seq );
score = score_calc_for_score( njob, seq );
if( scoremtx == 0 ) score += offset;
fprintf( stdout, "score = %f\n", score );
if ( scoremtx == 0 ) fprintf( stdout, "JTT %dPAM\n", pamN );
else if( scoremtx == 1 ) fprintf( stdout, "Dayhoff( machigai ga aru )\n" );
else if( scoremtx == 2 ) fprintf( stdout, "M-Y\n" );
else if( scoremtx == -1 ) fprintf( stdout, "DNA 1:%d\n", kimuraR );
fprintf( stdout, "gap penalty = %+6.2f, %+6.2f, %+6.2f\n", (double)ppenalty/1000, (double)ppenalty_ex/1000, (double)poffset/1000 );
exit( 0 );
}
int main()
{
int i, j;
char **seq;
static char name[M][B];
static int nlen[M];
double **mtx;
FILE *fp;
int res;
scoremtx = NOTSPECIFIED;
#if 0
PreRead( stdin, &njob, &nlenmax );
#else
getnumlen( stdin );
#endif
rewind( stdin );
seq = AllocateCharMtx( njob, nlenmax+1 );
mtx = AllocateDoubleMtx( njob, njob );
#if 0
FRead( stdin, name, nlen, seq );
#else
readData( stdin, name, nlen, seq );
#endif
for( i=0; i<njob-1; i++ )
{
fprintf( stderr, "%4d/%4d\r", i+1, njob );
for( j=i+1; j<njob; j++ )
mtx[i][j] = (double)substitution_score( seq[i], seq[j] );
}
#if TEST
for( i=0; i<njob-1; i++ ) for( j=i+1; j<njob; j++ )
fprintf( stdout, "i=%d, j=%d, mtx[][] = %f\n", i, j, mtx[i][j] );
#endif
fp = fopen( "hat2", "w" );
WriteHat2( fp, njob, name, mtx );
fclose( fp );
exit( 0 );
/*
res = system( ALNDIR "/spgsdl < hat2" );
if( res ) exit( 1 );
else exit( 0 );
*/
}
int main( int argc, char *argv[] )
{
static int nlen[M];
static char **name, **seq;
static char **bseq;
static double *eff;
int i;
char c;
int alloclen;
FILE *infp;
arguments( argc, argv );
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
if( !pairfile )
{
fprintf( stderr, "Usage: %s -p pairfile -i inputfile \n", argv[0] );
exit( 1 );
}
getnumlen( infp );
rewind( infp );
if( njob < 2 )
{
fprintf( stderr, "At least 2 sequences should be input!\n"
"Only %d sequence found.\n", njob );
exit( 1 );
}
name = AllocateCharMtx( njob, B+1 );
seq = AllocateCharMtx( njob, nlenmax*9+1 );
bseq = AllocateCharMtx( njob, nlenmax*9+1 );
alloclen = nlenmax*9;
eff = AllocateDoubleVec( njob );
#if 0
Read( name, nlen, seq );
#else
readData_pointer( infp, name, nlen, seq );
#endif
fclose( infp );
constants( njob, seq );
#if 0
fprintf( stderr, "params = %d, %d, %d\n", penalty, penalty_ex, offset );
#endif
initSignalSM();
initFiles();
WriteOptions( trap_g );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illeagal character %c\n", c );
exit( 1 );
}
// writePre( njob, name, nlen, seq, 0 );
for( i=0; i<njob; i++ ) eff[i] = 1.0;
for( i=0; i<njob; i++ ) gappick0( bseq[i], seq[i] );
pairalign( name, nlen, bseq, eff, alloclen );
fprintf( trap_g, "done.\n" );
#if DEBUG
fprintf( stderr, "closing trap_g\n" );
#endif
fclose( trap_g );
#if IODEBUG
fprintf( stderr, "OSHIMAI\n" );
#endif
SHOWVERSION;
return( 0 );
}
int main( int argc, char *argv[] )
{
static int nlen[M];
static char **name, **seq;
static char **oseq;
static double **pscore;
static double *eff;
static double **node0, **node1;
static double *gapc;
static double *avgap;
double tmpavgap;
int i, j, m, goffset;
static int ***topol;
static double **len;
FILE *prep;
char c;
int corestart, coreend;
int alloclen;
int winsize;
char *pt, *ot;
double gapmin;
arguments( argc, argv );
getnumlen( stdin );
rewind( stdin );
if( njob < 2 )
{
fprintf( stderr, "At least 2 sequences should be input!\n"
"Only %d sequence found.\n", njob );
exit( 1 );
}
seq = AllocateCharMtx( njob, nlenmax*9+1 );
name = AllocateCharMtx( njob, B+1 );
oseq = AllocateCharMtx( njob, nlenmax*9+1 );
alloclen = nlenmax*9;
topol = AllocateIntCub( njob, 2, njob );
len = AllocateDoubleMtx( njob, 2 );
pscore = AllocateDoubleMtx( njob, njob );
eff = AllocateDoubleVec( njob );
node0 = AllocateDoubleMtx( njob, njob );
node1 = AllocateDoubleMtx( njob, njob );
gapc = AllocateDoubleVec( alloclen );
avgap = AllocateDoubleVec( alloclen );
#if 0
Read( name, nlen, seq );
#else
readData_pointer( stdin, name, nlen, seq );
#endif
constants( njob, seq );
#if 0
fprintf( stderr, "params = %d, %d, %d\n", penalty, penalty_ex, offset );
#endif
initSignalSM();
initFiles();
WriteOptions( trap_g );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illeagal character %c\n", c );
exit( 1 );
}
writePre( njob, name, nlen, seq, 0 );
if( tbutree == 0 )
{
for( i=1; i<njob; i++ )
{
if( nlen[i] != nlen[0] )
{
fprintf( stderr, "Input pre-aligned seqences or make hat2.\n" );
exit( 1 );
}
}
for( i=0; i<njob-1; i++ ) for( j=i+1; j<njob; j++ )
{
/*
pscore[i][j] = (double)score_calc1( seq[i], seq[j] );
*/
pscore[i][j] = (double)substitution_hosei( seq[i], seq[j] );
}
}
else
{
fprintf( stderr, "Loading 'hat2' ... " );
prep = fopen( "hat2", "r" );
if( prep == NULL ) ErrorExit( "Make hat2." );
readhat2_pointer( prep, njob, name, pscore );
fclose( prep );
fprintf( stderr, "done.\n" );
#if 0
prep = fopen( "hat2_check", "w" );
WriteHat2( prep, njob, name, pscore );
fclose( prep );
#endif
}
fprintf( stderr, "Constructing dendrogram ... " );
if( treemethod == 'x' )
supg( njob, pscore, topol, len );
else if( treemethod == 's' )
spg( njob, pscore, topol, len );
else if( treemethod == 'p' )
upg2( njob, pscore, topol, len );
else
ErrorExit( "Incorrect tree\n" );
fprintf( stderr, "done.\n" );
countnode( njob, topol, node0 );
if( tbrweight )
{
weight = 3;
#if 0
utree = 0; counteff( njob, topol, len, eff ); utree = 1;
#else
counteff_simple( njob, topol, len, eff );
#endif
}
else
{
for( i=0; i<njob; i++ ) eff[i] = 1.0;
}
for( i=0; i<nlenmax; i++ )
{
gapc[i] = 0.0;
for( j=0; j<njob; j++ )
{
if( seq[j][i] == '-' ) gapc[i] += eff[j];
}
}
gapmin = 1.0;
winsize = fftWinSize;
goffset = winsize/2;
tmpavgap = 0.0;
corestart = coreend = -1;
for( i=0; i<winsize; i++ )
{
tmpavgap += gapc[i];
}
for( i=winsize; i<nlenmax; i++ )
{
m = i - goffset;
avgap[m] = tmpavgap / winsize;
// fprintf( stdout, "%d %f %f\n", m, avgap[m], gapc[i] );
if( avgap[m] < corethr )
{
if( corestart == -1 )
corestart = i - winsize;
// fprintf( stdout, "ok, gapmin = %f, corestart = %d, coreend = %d\n", gapmin, corestart, coreend );
if( avgap[m] < gapmin )
{
gapmin = avgap[m];
}
coreend = i;
}
tmpavgap -= gapc[i-winsize];
tmpavgap += gapc[i];
}
if( corestart == -1 || coreend == -1 )
{
corestart = 0;
coreend = nlenmax-1;
}
for( i=0; i<njob; i++ )
{
pt = oseq[i];
m = winsize;
while( m-- ) *pt++ = '-';
for( j=corestart; j<=coreend; j++ )
*pt++ = seq[i][j];
m = winsize;
while( m-- ) *pt++ = '-';
*pt = 0;
ot = oseq[i]+winsize-1;
pt = seq[i]+corestart-1;
if( coreext ) m = winsize;
else m = 0;
while( m && --pt > seq[i] )
if( *pt != '-' )
{
*ot-- = *pt;
m--;
}
ot = oseq[i]+winsize+coreend-corestart+1;
pt = seq[i]+coreend;
if( coreext ) m = winsize;
else m = 0;
while( m && *(++pt) )
{
if( *pt != '-' )
{
*ot++ = *pt;
m--;
}
}
fprintf( stdout, ">%s\n", name[i] );
fprintf( stdout, "%s\n", oseq[i] );
}
exit( 1 );
SHOWVERSION;
return( 0 );
}
void rnaalifoldcall( char **seq, int nseq, RNApair **pairprob )
{
int lgth;
int i;
static TLS int *order = NULL;
static TLS char **name = NULL;
char gett[1000];
FILE *fp;
int left, right, dumm;
float prob;
static TLS int pid;
static TLS char fnamein[100];
static TLS char cmd[1000];
static TLS int *pairnum;
lgth = strlen( seq[0] );
if( order == NULL )
{
pid = (int)getpid();
sprintf( fnamein, "/tmp/_rnaalifoldin.%d", pid );
order = AllocateIntVec( njob );
name = AllocateCharMtx( njob, 10 );
for( i=0; i<njob; i++ )
{
order[i] = i;
sprintf( name[i], "seq%d", i );
}
}
pairnum = calloc( lgth, sizeof( int ) );
for( i=0; i<lgth; i++ ) pairnum[i] = 0;
fp = fopen( fnamein, "w" );
if( !fp )
{
fprintf( stderr, "Cannot open /tmp/_rnaalifoldin\n" );
exit( 1 );
}
clustalout_pointer( fp, nseq, lgth, seq, name, NULL, NULL, order, 15 );
fclose( fp );
sprintf( cmd, "RNAalifold -p %s", fnamein );
system( cmd );
fp = fopen( "alifold.out", "r" );
if( !fp )
{
fprintf( stderr, "Cannot open /tmp/_rnaalifoldin\n" );
exit( 1 );
}
#if 0
for( i=0; i<lgth; i++ ) // atode kesu
{
pairprob[i] = (RNApair *)realloc( pairprob[i], (2) * sizeof( RNApair ) ); // atode kesu
pairprob[i][1].bestscore = -1.0;
pairprob[i][1].bestpos = -1;
}
#endif
while( 1 )
{
fgets( gett, 999, fp );
if( gett[0] == '(' ) break;
if( gett[0] == '{' ) break;
if( gett[0] == '.' ) break;
if( gett[0] == ',' ) break;
if( gett[0] != ' ' ) continue;
sscanf( gett, "%d %d %d %f", &left, &right, &dumm, &prob );
left--;
right--;
#if 0
if( prob > 50.0 && prob > pairprob[left][0].bestscore )
{
pairprob[left][0].bestscore = prob;
pairprob[left][0].bestpos = right;
#else
if( prob > 0.0 )
{
pairprob[left] = (RNApair *)realloc( pairprob[left], (pairnum[left]+2) * sizeof( RNApair ) );
pairprob[left][pairnum[left]].bestscore = prob / 100.0;
pairprob[left][pairnum[left]].bestpos = right;
pairnum[left]++;
pairprob[left][pairnum[left]].bestscore = -1.0;
pairprob[left][pairnum[left]].bestpos = -1;
fprintf( stderr, "%d-%d, %f\n", left, right, prob );
pairprob[right] = (RNApair *)realloc( pairprob[right], (pairnum[right]+2) * sizeof( RNApair ) );
pairprob[right][pairnum[right]].bestscore = prob / 100.0;
pairprob[right][pairnum[right]].bestpos = left;
pairnum[right]++;
pairprob[right][pairnum[right]].bestscore = -1.0;
pairprob[right][pairnum[right]].bestpos = -1;
fprintf( stderr, "%d-%d, %f\n", left, right, prob );
#endif
}
}
fclose( fp );
sprintf( cmd, "rm -f %s", fnamein );
system( cmd );
for( i=0; i<lgth; i++ )
{
if( (right=pairprob[i][0].bestpos) > -1 )
{
pairprob[right][0].bestpos = i;
pairprob[right][0].bestscore = pairprob[i][0].bestscore;
}
}
#if 0
for( i=0; i<lgth; i++ ) // atode kesu
if( pairprob[i][0].bestscore > -1 ) pairprob[i][0].bestscore = 1.0; // atode kesu
#endif
// fprintf( stderr, "after taikakuka in rnaalifoldcall\n" );
// for( i=0; i<lgth; i++ )
// {
// fprintf( stderr, "pair of %d = %d (%f) %c:%c\n", i, pairprob[i][0].bestpos, pairprob[i][0].bestscore, seq[0][i], seq[0][pairprob[i][0].bestpos] );
// }
free( pairnum );
}
static void utot( int n, int l, char **s )
{
int i, j;
for( i=0; i<l; i++ )
{
for( j=0; j<n; j++ )
{
if ( s[j][i] == 'a' ) s[j][i] = 'a';
else if( s[j][i] == 't' ) s[j][i] = 't';
else if( s[j][i] == 'u' ) s[j][i] = 't';
else if( s[j][i] == 'g' ) s[j][i] = 'g';
else if( s[j][i] == 'c' ) s[j][i] = 'c';
else if( s[j][i] == '-' ) s[j][i] = '-';
else s[j][i] = 'n';
}
}
}
int main( int argc, char *argv[] )
{
static char com[10000];
static int *nlen;
int left, right;
int res;
static char **name, **seq, **nogap;
static int **gapmap;
static int *order;
int i, j;
FILE *infp;
RNApair ***pairprob;
RNApair **alnpairprob;
RNApair *pairprobpt;
RNApair *pt;
int *alnpairnum;
float prob;
int adpos;
arguments( argc, argv );
#ifndef enablemultithread
nthread = 0;
#endif
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
if( !whereismccaskillmea )
whereismccaskillmea = "";
getnumlen( infp );
rewind( infp );
if( dorp != 'd' )
{
fprintf( stderr, "nuc only\n" );
exit( 1 );
}
seq = AllocateCharMtx( njob, nlenmax*2+1 );
nogap = AllocateCharMtx( njob, nlenmax*2+1 );
gapmap = AllocateIntMtx( njob, nlenmax*2+1 );
order = AllocateIntVec( njob );
name = AllocateCharMtx( njob, B+1 );
nlen = AllocateIntVec( njob );
pairprob = (RNApair ***)calloc( njob, sizeof( RNApair ** ) );
alnpairprob = (RNApair **)calloc( nlenmax, sizeof( RNApair * ) );
alnpairnum = AllocateIntVec( nlenmax );
for( i=0; i<nlenmax; i++ ) alnpairnum[i] = 0;
readData_pointer( infp, name, nlen, seq );
fclose( infp );
for( i=0; i<njob; i++ )
{
pairprob[i] = (RNApair **)calloc( nlenmax, sizeof( RNApair * ) );
for( j=0; j<nlenmax; j++ )
{
pairprob[i][j] = (RNApair *)calloc( 1, sizeof( RNApair ) );
pairprob[i][j][0].bestpos = -1;
pairprob[i][j][0].bestscore = -1.0;
}
strcpy( nogap[i], seq[i] );
order[i] = i;
}
for( j=0; j<nlenmax; j++ )
{
alnpairprob[j] = (RNApair *)calloc( 1, sizeof( RNApair ) );
alnpairprob[j][0].bestpos = -1;
alnpairprob[j][0].bestscore = -1.0;
}
constants( njob, seq );
if( alg == 'G' )
fprintf( stderr, "Running DAFS (Sato et al. 2012; http://www.ncrna.org/).\n" );
else
fprintf( stderr, "Running mxscarna with the mccaskill_mea mode.\n" );
#ifdef enablemultithread
if( nthread > 0 )
{
int jobpos;
pthread_t *handle;
pthread_mutex_t mutex;
thread_arg_t *targ;
jobpos = 0;
targ = calloc( nthread, sizeof( thread_arg_t ) );
handle = calloc( nthread, sizeof( pthread_t ) );
pthread_mutex_init( &mutex, NULL );
for( i=0; i<nthread; i++ )
{
targ[i].thread_no = i;
targ[i].njob = njob;
targ[i].jobpospt = &jobpos;
targ[i].gapmap = gapmap;
targ[i].nogap = nogap;
targ[i].nlenmax = nlenmax;
targ[i].pairprob = pairprob;
targ[i].mutex = &mutex;
// athread( targ );
pthread_create( handle+i, NULL, athread, (void *)(targ+i) );
}
for( i=0; i<nthread; i++ )
{
pthread_join( handle[i], NULL );
}
pthread_mutex_destroy( &mutex );
for( i=0; i<njob; i++ )
{
fprintf( stdout, ">%d\n", i );
outmccaskill( stdout, pairprob[i], nlenmax );
}
}
else
#endif
{
for( i=0; i<njob; i++ )
{
fprintf( stderr, "%d / %d\n", i+1, njob );
commongappick_record( 1, nogap+i, gapmap[i] );
infp = fopen( "_mccaskillinorg", "w" );
// fprintf( infp, ">in\n%s\n", nogap[i] );
fprintf( infp, ">in\n" );
write1seq( infp, nogap[i] );
fclose( infp );
system( "tr -d '\\r' < _mccaskillinorg > _mccaskillin" ); // for cygwin, wakaran
if( alg == 'G' )
sprintf( com, "env PATH=%s dafs --mafft-out _mccaskillout _mccaskillin > _dum1 2>_dum", whereismccaskillmea );
else
sprintf( com, "env PATH=%s mxscarnamod -m -writebpp _mccaskillin > _mccaskillout 2>_dum", whereismccaskillmea );
res = system( com );
if( res )
{
fprintf( stderr, "ERROR IN mccaskill_mea\n" );
exit( 1 );
}
infp = fopen( "_mccaskillout", "r" );
readrawmccaskill( infp, pairprob[i], nlenmax );
fclose( infp );
fprintf( stdout, ">%d\n", i );
outmccaskill( stdout, pairprob[i], nlenmax );
}
}
for( i=0; i<njob; i++ )
{
for( j=0; j<nlen[i]; j++ ) for( pairprobpt=pairprob[i][j]; pairprobpt->bestpos!=-1; pairprobpt++ )
{
left = gapmap[i][j];
right = gapmap[i][pairprobpt->bestpos];
prob = pairprobpt->bestscore;
for( pt=alnpairprob[left]; pt->bestpos!=-1; pt++ )
if( pt->bestpos == right ) break;
if( pt->bestpos == -1 )
{
alnpairprob[left] = (RNApair *)realloc( alnpairprob[left], (alnpairnum[left]+2) * sizeof( RNApair ) );
adpos = alnpairnum[left];
alnpairnum[left]++;
alnpairprob[left][adpos].bestscore = 0.0;
alnpairprob[left][adpos].bestpos = right;
alnpairprob[left][adpos+1].bestscore = -1.0;
alnpairprob[left][adpos+1].bestpos = -1;
pt = alnpairprob[left]+adpos;
}
else
adpos = pt-alnpairprob[left];
pt->bestscore += prob;
if( pt->bestpos != right )
{
fprintf( stderr, "okashii!\n" );
exit( 1 );
}
// fprintf( stderr, "adding %d-%d, %f\n", left, right, prob );
}
}
for( i=0; i<njob; i++ )
{
for( j=0; j<nlenmax; j++ ) free( pairprob[i][j] );
free( pairprob[i] );
}
free( pairprob );
for( j=0; j<nlenmax; j++ ) free( alnpairprob[j] );
free( alnpairprob );
free( alnpairnum );
fprintf( stderr, "%d thread(s)\n", nthread );
return( 0 );
#if 0
fprintf( stdout, "result=\n" );
for( i=0; i<nlenmax; i++ ) for( pairprobpt=alnpairprob[i]; pairprobpt->bestpos!=-1; pairprobpt++ )
{
pairprobpt->bestscore /= (float)njob;
left = i;
right = pairprobpt->bestpos;
prob = pairprobpt->bestscore;
fprintf( stdout, "%d-%d, %f\n", left, right, prob );
}
return( 0 );
#endif
}
int main( int argc, char *argv[] )
{
static int nlen[M];
static char **name, **seq;
int i, j, alloclen, c;
double **mtx;
double *self;
double tmpdouble;
FILE *fp;
arguments( argc, argv );
getnumlen( stdin );
rewind( stdin );
if( njob < 2 )
{
fprintf( stderr, "At least 2 sequences should be input!\n"
"Only %d sequence found.\n", njob );
exit( 1 );
}
name = AllocateCharMtx( njob, B+1 );
seq = AllocateCharMtx( njob, nlenmax*9+1 );
mtx = AllocateDoubleMtx( njob, njob );
self = AllocateDoubleVec( njob );
alloclen = nlenmax*9;
readData_pointer( stdin, name, nlen, seq );
constants( njob, seq );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illeagal character %c\n", c );
exit( 1 );
}
for( i=0; i<njob; i++ )
{
self[i] = (double)substitution_nid( seq[i], seq[i] );
// fprintf( stdout, "self[%d] = %f\n", i, self[i] );
}
for( i=0; i<njob-1; i++ )
for( j=i+1; j<njob; j++ )
{
tmpdouble = (double)substitution_score( seq[i], seq[j] );
// fprintf( stdout, "tmpdouble = %f\n", tmpdouble );
mtx[i][j] = ( 1.0 - tmpdouble / MIN( self[i], self[j] ) );
if( mtx[i][j] < 0.95 )
mtx[i][j] = - log( 1.0 - mtx[i][j] );
else
mtx[i][j] = 3.0;
}
#if TEST
for( i=0; i<njob-1; i++ ) for( j=i+1; j<njob; j++ )
fprintf( stdout, "i=%d, j=%d, mtx[][] = %f\n", i, j, mtx[i][j] );
#endif
fp = fopen( "hat2", "w" );
WriteHat2_pointer( fp, njob, name, mtx );
fclose( fp );
exit( 0 );
return( 0 );
}
int main( int argc, char **argv )
{
int i, j;
char **seq;
static char **name;
static int nlen[M];
float *selfscore;
double **mtx;
FILE *fp;
FILE *infp;
float ssi, ssj, bunbo;
arguments( argc, argv );
#ifndef enablemultithread
nthread = 0;
#endif
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 );
seq = AllocateCharMtx( njob, nlenmax+1 );
name = AllocateCharMtx( njob, B+1 );
mtx = AllocateDoubleMtx( njob, njob );
selfscore = AllocateFloatVec( njob );
#if 0
FRead( stdin, name, nlen, seq );
#else
readData_pointer( infp, name, nlen, seq );
#endif
fclose( infp );
constants( njob, seq );
#if 0
for( i=0; i<njob-1; i++ )
{
fprintf( stderr, "%4d/%4d\r", i+1, njob );
for( j=i+1; j<njob; j++ )
mtx[i][j] = (double)substitution_hosei( seq[i], seq[j] );
// fprintf( stderr, "i=%d,j=%d, l=%d &&& %f\n", i, j, nlen[0], mtx[i][j] );
}
#else // 061003
for( i=0; i<njob; i++ )
{
selfscore[i] = (float)naivepairscore11( seq[i], seq[i], penalty );
}
#ifdef enablemultithread
if( nthread > 0 )
{
thread_arg_t *targ;
Jobtable jobpos;
pthread_t *handle;
pthread_mutex_t mutex;
jobpos.i = 0;
jobpos.j = 0;
targ = calloc( nthread, sizeof( thread_arg_t ) );
handle = calloc( nthread, sizeof( pthread_t ) );
pthread_mutex_init( &mutex, NULL );
for( i=0; i<nthread; i++ )
{
targ[i].thread_no = i;
targ[i].njob = njob;
targ[i].selfscore = selfscore;
targ[i].mtx = mtx;
targ[i].seq = seq;
targ[i].jobpospt = &jobpos;
targ[i].mutex = &mutex;
pthread_create( handle+i, NULL, athread, (void *)(targ+i) );
}
for( i=0; i<nthread; i++ )
{
pthread_join( handle[i], NULL );
}
pthread_mutex_destroy( &mutex );
}
else
#endif
{
for( i=0; i<njob-1; i++ )
{
ssi = selfscore[i];
fprintf( stderr, "%4d/%4d\r", i+1, njob );
for( j=i+1; j<njob; j++ )
{
ssj = selfscore[j];
bunbo = MIN( ssi, ssj );
if( bunbo == 0.0 )
mtx[i][j] = 1.0;
else
mtx[i][j] = 1.0 - (double)naivepairscore11( seq[i], seq[j], penalty ) / bunbo;
// mtx[i][j] = 1.0 - (double)naivepairscore11( seq[i], seq[j], penalty ) / MIN( selfscore[i], selfscore[j] );
// fprintf( stderr, "i=%d,j=%d, l=%d### %f, score = %d\n", i, j, nlen[0], mtx[i][j], naivepairscore11( seq[i], seq[j], penalty ) );
}
}
}
#endif
#if TEST
for( i=0; i<njob-1; i++ ) for( j=i+1; j<njob; j++ )
fprintf( stdout, "i=%d, j=%d, mtx[][] = %f\n", i, j, mtx[i][j] );
#endif
fp = fopen( "hat2", "w" );
WriteHat2_pointer( fp, njob, name, mtx );
fclose( fp );
#if 0
if( treeout )
{
int ***topol;
double **len;
topol = AllocateIntCub( njob, 2, njob );
len = AllocateDoubleMtx( njob, njob );
veryfastsupg_double_outtree( njob, mtx, topol, len );
}
#endif
SHOWVERSION;
exit( 0 );
/*
res = system( ALNDIR "/spgsdl < hat2" );
if( res ) exit( 1 );
else exit( 0 );
*/
}
int main( int argc, char *argv[] )
{
FILE *infp;
int nlenmin;
char **name;
char **seq;
int *nlen;
int i;
char *usual;
arguments( argc, argv );
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
// dorp = NOTSPECIFIED;
getnumlen_casepreserve( infp, &nlenmin );
fprintf( stderr, "%d x %d - %d %c\n", njob, nlenmax, nlenmin, dorp );
seq = AllocateCharMtx( njob, nlenmax+1 );
name = AllocateCharMtx( njob, B+1 );
nlen = AllocateIntVec( njob );
readData_pointer_casepreserve( infp, name, nlen, seq );
// for( i=0; i<njob; i++ ) gappick_samestring( seq[i] );
#if 0
FILE *origfp;
origfp = fopen( "_original", "w" );
if( !origfp )
{
fprintf( stderr, "Cannot open _original\n" );
exit( 1 );
}
for( i=0; i<njob; i++ )
{
nlen[i] = strlen( seq[i] );
fprintf( origfp, ">%s\n", name[i]+1 );
if( seq[i][nlen[i]-1] == '\n' ) seq[i][nlen[i]-1] = 0;
fprintf( origfp, "%s\n", seq[i] );
}
fclose( origfp );
#endif
if( dorp == 'p' )
{
usual = "ARNDCQEGHILKMFPSTWYVarndcqeghilkmfpstwyv-.";
replace_unusual( njob, seq, usual, 'X', toupper );
}
else
{
usual = "ATGCUatgcuBDHKMNRSVWYXbdhkmnrsvwyx-";
replace_unusual( njob, seq, usual, 'n', tolower );
}
for( i=0; i<njob; i++ )
{
fprintf( stdout, ">_os_%d_oe_%s\n", i+seedoffset, name[i]+1 );
fprintf( stdout, "%s\n", seq[i] );
}
free( nlen );
FreeCharMtx( seq );
FreeCharMtx( name );
return( 0 );
}
float partA__align( char **seq1, char **seq2, double *eff1, double *eff2, int icyc, int jcyc, int alloclen, LocalHom ***localhom, float *impmatch, int start1, int end1, int start2, int end2, int *gapmap1, int *gapmap2, char *sgap1, char *sgap2, char *egap1, char *egap2 )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
register int i, j;
int lasti, lastj; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */
int lgth1, lgth2;
int resultlen;
float wm = 0.0; /* int ?????? */
float g;
float *currentw, *previousw;
#if 1
float *wtmp;
int *ijppt;
float *mjpt, *prept, *curpt;
int *mpjpt;
#endif
static float mi, *m;
static int **ijp;
static int mpi, *mp;
static float *w1, *w2;
static float *match;
static float *initverticalw; /* kufuu sureba iranai */
static float *lastverticalw; /* kufuu sureba iranai */
static char **mseq1;
static char **mseq2;
static char **mseq;
static float *ogcp1;
static float *ogcp2;
static float *fgcp1;
static float *fgcp2;
static float **cpmx1;
static float **cpmx2;
static int **intwork;
static float **floatwork;
static int orlgth1 = 0, orlgth2 = 0;
float fpenalty = (float)penalty;
#if USE_PENALTY_EX
float fpenalty_ex = (float)penalty_ex;
#endif
float *fgcp2pt;
float *ogcp2pt;
float fgcp1va;
float ogcp1va;
#if 0
fprintf( stderr, "eff in SA+++align\n" );
for( i=0; i<icyc; i++ ) fprintf( stderr, "eff1[%d] = %f\n", i, eff1[i] );
#endif
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
FreeFloatVec( ogcp1 );
FreeFloatVec( ogcp2 );
FreeFloatVec( fgcp1 );
FreeFloatVec( fgcp2 );
FreeFloatMtx( cpmx1 );
FreeFloatMtx( cpmx2 );
FreeFloatMtx( floatwork );
FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
ogcp1 = AllocateFloatVec( ll1+2 );
ogcp2 = AllocateFloatVec( ll2+2 );
fgcp1 = AllocateFloatVec( ll1+2 );
fgcp2 = AllocateFloatVec( ll2+2 );
cpmx1 = AllocateFloatMtx( 26, ll1+2 );
cpmx2 = AllocateFloatMtx( 26, ll2+2 );
#if FASTMATCHCALC
floatwork = AllocateFloatMtx( MAX( ll1, ll2 )+2, 26 );
intwork = AllocateIntMtx( MAX( ll1, ll2 )+2, 26 );
#else
floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 );
intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 );
#endif
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
for( i=0; i<icyc; i++ ) mseq1[i] = mseq[i];
for( j=0; j<jcyc; j++ ) mseq2[j] = mseq[icyc+j];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijp = commonIP;
cpmx_calc_new( seq1, cpmx1, eff1, lgth1, icyc );
cpmx_calc_new( seq2, cpmx2, eff2, lgth2, jcyc );
if( sgap1 )
{
new_OpeningGapCount( ogcp1, icyc, seq1, eff1, lgth1, sgap1 );
new_OpeningGapCount( ogcp2, jcyc, seq2, eff2, lgth2, sgap2 );
new_FinalGapCount( fgcp1, icyc, seq1, eff1, lgth1, egap1 );
new_FinalGapCount( fgcp2, jcyc, seq2, eff2, lgth2, egap2 );
}
else
{
st_OpeningGapCount( ogcp1, icyc, seq1, eff1, lgth1 );
st_OpeningGapCount( ogcp2, jcyc, seq2, eff2, lgth2 );
st_FinalGapCount( fgcp1, icyc, seq1, eff1, lgth1 );
st_FinalGapCount( fgcp2, jcyc, seq2, eff2, lgth2 );
}
for( i=0; i<lgth1; i++ )
{
ogcp1[i] = 0.5 * ( 1.0 - ogcp1[i] ) * fpenalty;
fgcp1[i] = 0.5 * ( 1.0 - fgcp1[i] ) * fpenalty;
}
for( i=0; i<lgth2; i++ )
{
ogcp2[i] = 0.5 * ( 1.0 - ogcp2[i] ) * fpenalty;
fgcp2[i] = 0.5 * ( 1.0 - fgcp2[i] ) * fpenalty;
}
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc( initverticalw, cpmx2, cpmx1, 0, lgth1, floatwork, intwork, 1 );
if( localhom )
part_imp_match_out_vead_tate_gapmap( initverticalw, gapmap2[0]+start2, lgth1, start1, gapmap1 );
match_calc( currentw, cpmx1, cpmx2, 0, lgth2, floatwork, intwork, 1 );
if( localhom )
part_imp_match_out_vead_gapmap( currentw, gapmap1[0]+start1, lgth2, start2, gapmap2 );
#if 0 // -> tbfast.c
if( localhom )
imp_match_calc( currentw, icyc, jcyc, lgth1, lgth2, seq1, seq2, eff1, eff2, localhom, 1, 0 );
#endif
if( outgap == 1 )
{
for( i=1; i<lgth1+1; i++ )
{
initverticalw[i] += ( ogcp1[0] + fgcp1[i-1] ) ;
}
for( j=1; j<lgth2+1; j++ )
{
currentw[j] += ( ogcp2[0] + fgcp2[j-1] ) ;
}
}
#if OUTGAP0TRY
else
{
for( j=1; j<lgth2+1; j++ )
currentw[j] -= offset * j / 2.0;
for( i=1; i<lgth1+1; i++ )
initverticalw[i] -= offset * i / 2.0;
}
#endif
for( j=1; j<lgth2+1; ++j )
{
m[j] = currentw[j-1] + ogcp1[1]; mp[j] = 0;
}
lastverticalw[0] = currentw[lgth2-1];
if( outgap ) lasti = lgth1+1; else lasti = lgth1;
lastj = lgth2+1;
#if XXXXXXX
fprintf( stderr, "currentw = \n" );
for( i=0; i<lgth1+1; i++ )
{
fprintf( stderr, "%5.2f ", currentw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "initverticalw = \n" );
for( i=0; i<lgth2+1; i++ )
{
fprintf( stderr, "%5.2f ", initverticalw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "fcgp\n" );
for( i=0; i<lgth1; i++ )
fprintf( stderr, "fgcp1[%d]=%f\n", i, ogcp1[i] );
for( i=0; i<lgth2; i++ )
fprintf( stderr, "fgcp2[%d]=%f\n", i, ogcp2[i] );
#endif
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc( currentw, cpmx1, cpmx2, i, lgth2, floatwork, intwork, 0 );
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
if( localhom )
{
// fprintf( stderr, "Calling imp_match_calc (o) lgth = %d, i = %d\n", lgth1, i );
// imp_match_out_vead( currentw, i, lgth2 );
part_imp_match_out_vead_gapmap( currentw, gapmap1[i]+start1, lgth2, start2, gapmap2 );
}
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
currentw[0] = initverticalw[i];
mi = previousw[0] + ogcp2[1]; mpi = 0;
ijppt = ijp[i] + 1;
mjpt = m + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
fgcp2pt = fgcp2;
ogcp2pt = ogcp2+1;
fgcp1va = fgcp1[i-1];
ogcp1va = ogcp1[i];
for( j=1; j<lastj; j++ )
{
wm = *prept;
*ijppt = 0;
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
g = mi + *fgcp2pt;
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
*ijppt = -( j - mpi );
}
g = *prept + *ogcp2pt;
if( g >= mi )
{
mi = g;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
g = *mjpt + fgcp1va;
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
*ijppt = +( i - *mpjpt );
}
g = *prept + ogcp1va;
if( g >= *mjpt )
{
*mjpt = g;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
m[j] += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f ", wm );
#endif
*curpt += wm;
ijppt++;
mjpt++;
prept++;
mpjpt++;
curpt++;
fgcp2pt++;
ogcp2pt++;
}
lastverticalw[i] = currentw[lgth2-1];
}
#if OUTGAP0TRY
if( !outgap )
{
for( j=1; j<lgth2+1; j++ )
currentw[j] -= offset * ( lgth2 - j ) / 2.0;
for( i=1; i<lgth1+1; i++ )
lastverticalw[i] -= offset * ( lgth1 - i / 2.0);
}
#endif
/*
fprintf( stderr, "\n" );
for( i=0; i<icyc; i++ ) fprintf( stderr,"%s\n", seq1[i] );
fprintf( stderr, "#####\n" );
for( j=0; j<jcyc; j++ ) fprintf( stderr,"%s\n", seq2[j] );
fprintf( stderr, "====>" );
for( i=0; i<icyc; i++ ) strcpy( mseq1[i], seq1[i] );
for( j=0; j<jcyc; j++ ) strcpy( mseq2[j], seq2[j] );
*/
if( localhom )
{
Atracking_localhom( impmatch, currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc, start1, end1, start2, end2, gapmap1, gapmap2 );
}
else
Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp, icyc, jcyc );
// fprintf( stderr, "### impmatch = %f\n", *impmatch );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
for( i=0; i<icyc; i++ ) strcpy( seq1[i], mseq1[i] );
for( j=0; j<jcyc; j++ ) strcpy( seq2[j], mseq2[j] );
/*
fprintf( stderr, "\n" );
for( i=0; i<icyc; i++ ) fprintf( stderr, "%s\n", mseq1[i] );
fprintf( stderr, "#####\n" );
for( j=0; j<jcyc; j++ ) fprintf( stderr, "%s\n", mseq2[j] );
*/
return( wm );
}
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;
}
}
int main( int argc, char *argv[] )
{
static int nlen[M];
static char name[M][B], **seq;
static char **mseq1, **mseq2;
static char **aseq;
static char **bseq;
static double *eff;
static double *equiv;
char **strfiles;
char **chainids;
int i;
FILE *infp;
char c;
int alloclen;
arguments( argc, argv );
if( equivthreshold < 1 || 9 < equivthreshold )
{
fprintf( stderr, "-t n, n must be 1..9\n" );
exit( 1 );
}
if( ( equivwinsize + 1 ) % 2 != 0 )
{
fprintf( stderr, "equivwinsize = %d\n", equivwinsize );
fprintf( stderr, "It must be an odd number.\n" );
exit( 1 );
}
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
nlenmax = 10000; // tekitou
if( alg == 'R' )
prepareash( infp, &strfiles, &chainids, &seq, &mseq1, &mseq2, &equiv, &alloclen );
else if( alg == 'T' )
preparetmalign( infp, &strfiles, &chainids, &seq, &mseq1, &mseq2, &equiv, &alloclen );
fclose( infp );
aseq = AllocateCharMtx( njob, nlenmax*2+1 );
bseq = AllocateCharMtx( njob, nlenmax*2+1 );
eff = AllocateDoubleVec( njob );
for( i=0; i<njob; i++ )
{
fprintf( stderr, "str%d = %s-%s\n", i, strfiles[i], chainids[i] );
}
if( njob < 1 )
{
fprintf( stderr, "No structure found.\n" );
exit( 1 );
}
if( njob < 2 )
{
fprintf( stderr, "Only %d structure found.\n", njob );
exit( 0 );
}
if( njob > M )
{
fprintf( stderr, "The number of structures must be < %d\n", M );
fprintf( stderr, "Please try sequence-based methods for such large data.\n" );
exit( 1 );
}
#if 0
readData( infp, name, nlen, seq );
#endif
constants( njob, seq );
#if 0
fprintf( stderr, "params = %d, %d, %d\n", penalty, penalty_ex, offset );
#endif
initSignalSM();
initFiles();
WriteOptions( trap_g );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illegal character %c\n", c );
exit( 1 );
}
// writePre( njob, name, nlen, seq, 0 );
for( i=0; i<njob; i++ ) eff[i] = 1.0;
for( i=0; i<njob; i++ ) gappick0( bseq[i], seq[i] );
pairalign( name, nlen, bseq, aseq, mseq1, mseq2, equiv, eff, strfiles, chainids, alloclen );
fprintf( trap_g, "done.\n" );
#if DEBUG
fprintf( stderr, "closing trap_g\n" );
#endif
fclose( trap_g );
// writePre( njob, name, nlen, aseq, !contin );
#if 0
writeData( stdout, njob, name, nlen, aseq );
#endif
#if IODEBUG
fprintf( stderr, "OSHIMAI\n" );
#endif
SHOWVERSION;
return( 0 );
}
float L__align11( char **seq1, char **seq2, int alloclen, int *off1pt, int *off2pt )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
register int i, j;
int lasti, lastj; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */
int lgth1, lgth2;
int resultlen;
float wm = 0.0; /* int ?????? */
float g;
float *currentw, *previousw;
#if 1
float *wtmp;
int *ijppt;
float *mjpt, *prept, *curpt;
int *mpjpt;
#endif
static TLS float mi, *m;
static TLS int **ijp;
static TLS int mpi, *mp;
static TLS float *w1, *w2;
static TLS float *match;
static TLS float *initverticalw; /* kufuu sureba iranai */
static TLS float *lastverticalw; /* kufuu sureba iranai */
static TLS char **mseq1;
static TLS char **mseq2;
static TLS char **mseq;
// static TLS int **intwork;
// static TLS float **floatwork;
static TLS int orlgth1 = 0, orlgth2 = 0;
float maxwm;
int endali = 0, endalj = 0; // by D.Mathog, a guess
// int endali, endalj;
float localthr = -offset;
float localthr2 = -offset;
// float localthr = 100;
// float localthr2 = 100;
float fpenalty = (float)penalty;
float fpenalty_ex = (float)penalty_ex;
if( seq1 == NULL )
{
if( orlgth1 > 0 && orlgth2 > 0 )
{
orlgth1 = 0;
orlgth2 = 0;
free( mseq1 );
free( mseq2 );
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
}
return( 0.0 );
}
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
// FreeFloatMtx( floatwork );
// FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
// floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 );
// intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 );
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijp = commonIP;
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc( initverticalw, seq2, seq1, 0, lgth1 );
match_calc( currentw, seq1, seq2, 0, lgth2 );
lasti = lgth2+1;
for( j=1; j<lasti; ++j )
{
m[j] = currentw[j-1]; mp[j] = 0;
#if 0
if( m[j] < localthr ) m[j] = localthr2;
#endif
}
lastverticalw[0] = currentw[lgth2-1];
lasti = lgth1+1;
#if 0
fprintf( stderr, "currentw = \n" );
for( i=0; i<lgth1+1; i++ )
{
fprintf( stderr, "%5.2f ", currentw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "initverticalw = \n" );
for( i=0; i<lgth2+1; i++ )
{
fprintf( stderr, "%5.2f ", initverticalw[i] );
}
fprintf( stderr, "\n" );
#endif
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, " " );
for( j=0; j<lgth2; j++ )
fprintf( stderr, "%c ", seq2[0][j] );
fprintf( stderr, "\n" );
#endif
localstop = lgth1+lgth2+1;
maxwm = -999999999.9;
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, "%c ", seq1[0][0] );
for( j=0; j<lgth2+1; j++ )
fprintf( stderr, "%5.0f ", currentw[j] );
fprintf( stderr, "\n" );
#endif
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc( currentw, seq1, seq2, i, lgth2 );
#if DEBUG2
fprintf( stderr, "%c ", seq1[0][i] );
fprintf( stderr, "%5.0f ", currentw[0] );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
currentw[0] = initverticalw[i];
mi = previousw[0]; mpi = 0;
#if 0
if( mi < localthr ) mi = localthr2;
#endif
ijppt = ijp[i] + 1;
mjpt = m + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
lastj = lgth2+1;
for( j=1; j<lastj; j++ )
{
wm = *prept;
*ijppt = 0;
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=mi+fpenalty) > wm )
{
wm = g;
*ijppt = -( j - mpi );
}
if( *prept > mi )
{
mi = *prept;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=*mjpt+fpenalty) > wm )
{
wm = g;
*ijppt = +( i - *mpjpt );
}
if( *prept > *mjpt )
{
*mjpt = *prept;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
*mjpt += fpenalty_ex;
#endif
if( maxwm < wm )
{
maxwm = wm;
endali = i;
endalj = j;
}
#if 1
if( wm < localthr )
{
// fprintf( stderr, "stop i=%d, j=%d, curpt=%f\n", i, j, *curpt );
*ijppt = localstop;
wm = localthr2;
}
#endif
#if 0
fprintf( stderr, "%5.0f ", *curpt );
#endif
#if DEBUG2
fprintf( stderr, "%5.0f ", wm );
// fprintf( stderr, "%c-%c *ijppt = %d, localstop = %d\n", seq1[0][i], seq2[0][j], *ijppt, localstop );
#endif
*curpt++ += wm;
ijppt++;
mjpt++;
prept++;
mpjpt++;
}
#if DEBUG2
fprintf( stderr, "\n" );
#endif
lastverticalw[i] = currentw[lgth2-1];
}
#if 0
fprintf( stderr, "maxwm = %f\n", maxwm );
fprintf( stderr, "endali = %d\n", endali );
fprintf( stderr, "endalj = %d\n", endalj );
#endif
if( ijp[endali][endalj] == localstop )
{
strcpy( seq1[0], "" );
strcpy( seq2[0], "" );
*off1pt = *off2pt = 0;
fprintf( stderr, "maxwm <- 0.0 \n" );
return( 0.0 );
}
Ltracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, ijp, off1pt, off2pt, endali, endalj );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
strcpy( seq1[0], mseq1[0] );
strcpy( seq2[0], mseq2[0] );
#if 0
fprintf( stderr, "wm=%f\n", wm );
fprintf( stderr, ">\n%s\n", mseq1[0] );
fprintf( stderr, ">\n%s\n", mseq2[0] );
fprintf( stderr, "maxwm = %f\n", maxwm );
fprintf( stderr, " wm = %f\n", wm );
#endif
return( maxwm );
}
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 L__align11( double **n_dynamicmtx, float scoreoffset, char **seq1, char **seq2, int alloclen, int *off1pt, int *off2pt )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
int i, j;
int lasti, lastj; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */
int lgth1, lgth2;
int resultlen;
float wm = 0.0; /* int ?????? */
float g;
float *currentw, *previousw;
#if 1
float *wtmp;
int *ijppt;
float *mjpt, *prept, *curpt;
int *mpjpt;
#endif
static TLS float mi, *m;
static TLS int **ijp;
static TLS int mpi, *mp;
static TLS float *w1, *w2;
static TLS float *match;
static TLS float *initverticalw; /* kufuu sureba iranai */
static TLS float *lastverticalw; /* kufuu sureba iranai */
static TLS char **mseq1;
static TLS char **mseq2;
static TLS char **mseq;
// static TLS int **intwork;
// static TLS float **floatwork;
static TLS int orlgth1 = 0, orlgth2 = 0;
static TLS double **amino_dynamicmtx = NULL; // ??
float maxwm;
int endali = 0, endalj = 0; // by D.Mathog, a guess
// int endali, endalj;
float localthr = -offset + scoreoffset * 600; // 2013/12/13
float localthr2 = -offset + scoreoffset * 600; // 2013/12/13
// float localthr = -offset;
// float localthr2 = -offset;
float fpenalty = (float)penalty;
float fpenalty_ex = (float)penalty_ex;
float fpenalty_shift = (float)penalty_shift;
float fpenalty_tmp; // atode kesu
int *warpis = NULL;
int *warpjs = NULL;
int *warpi = NULL;
int *warpj = NULL;
int *prevwarpi = NULL;
int *prevwarpj = NULL;
float *wmrecords = NULL;
float *prevwmrecords = NULL;
int warpn = 0;
int warpbase;
float curm = 0.0;
float *wmrecordspt, *wmrecords1pt, *prevwmrecordspt;
int *warpipt, *warpjpt;
if( seq1 == NULL )
{
if( orlgth1 > 0 && orlgth2 > 0 )
{
orlgth1 = 0;
orlgth2 = 0;
free( mseq1 );
free( mseq2 );
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
if( amino_dynamicmtx ) FreeDoubleMtx( amino_dynamicmtx ); amino_dynamicmtx = NULL;
}
return( 0.0 );
}
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
warpbase = lgth1 + lgth2;
warpis = NULL;
warpjs = NULL;
warpn = 0;
if( trywarp )
{
wmrecords = AllocateFloatVec( lgth2+1 );
warpi = AllocateIntVec( lgth2+1 );
warpj = AllocateIntVec( lgth2+1 );
prevwmrecords = AllocateFloatVec( lgth2+1 );
prevwarpi = AllocateIntVec( lgth2+1 );
prevwarpj = AllocateIntVec( lgth2+1 );
for( i=0; i<lgth2+1; i++ ) prevwmrecords[i] = 0.0;
for( i=0; i<lgth2+1; i++ ) wmrecords[i] = 0.0;
for( i=0; i<lgth2+1; i++ ) prevwarpi[i] = -warpbase;
for( i=0; i<lgth2+1; i++ ) prevwarpj[i] = -warpbase;
for( i=0; i<lgth2+1; i++ ) warpi[i] = -warpbase;
for( i=0; i<lgth2+1; i++ ) warpj[i] = -warpbase;
}
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
if( amino_dynamicmtx ) FreeDoubleMtx( amino_dynamicmtx ); amino_dynamicmtx = NULL;
// FreeFloatMtx( floatwork );
// FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
// floatwork = AllocateFloatMtx( nalphabets, MAX( ll1, ll2 )+2 );
// intwork = AllocateIntMtx( nalphabets, MAX( ll1, ll2 )+2 );
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
amino_dynamicmtx = AllocateDoubleMtx( 0x80, 0x80 );
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
for( i=0; i<nalphabets; i++) for( j=0; j<nalphabets; j++ )
amino_dynamicmtx[(int)amino[i]][(int)amino[j]] = (double)n_dynamicmtx[i][j];
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijp = commonIP;
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc_mtx( amino_dynamicmtx, initverticalw, seq2, seq1, 0, lgth1 );
match_calc_mtx( amino_dynamicmtx, currentw, seq1, seq2, 0, lgth2 );
lasti = lgth2+1;
for( j=1; j<lasti; ++j )
{
m[j] = currentw[j-1]; mp[j] = 0;
#if 0
if( m[j] < localthr ) m[j] = localthr2;
#endif
}
lastverticalw[0] = currentw[lgth2-1];
lasti = lgth1+1;
#if 0
fprintf( stderr, "currentw = \n" );
for( i=0; i<lgth1+1; i++ )
{
fprintf( stderr, "%5.2f ", currentw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "initverticalw = \n" );
for( i=0; i<lgth2+1; i++ )
{
fprintf( stderr, "%5.2f ", initverticalw[i] );
}
fprintf( stderr, "\n" );
#endif
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, " " );
for( j=0; j<lgth2; j++ )
fprintf( stderr, "%c ", seq2[0][j] );
fprintf( stderr, "\n" );
#endif
localstop = lgth1+lgth2+1;
maxwm = -999999999.9;
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, "%c ", seq1[0][0] );
for( j=0; j<lgth2+1; j++ )
fprintf( stderr, "%5.0f ", currentw[j] );
fprintf( stderr, "\n" );
#endif
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc_mtx( amino_dynamicmtx, currentw, seq1, seq2, i, lgth2 );
#if DEBUG2
fprintf( stderr, "%c ", seq1[0][i] );
fprintf( stderr, "%5.0f ", currentw[0] );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
currentw[0] = initverticalw[i];
mi = previousw[0]; mpi = 0;
#if 0
if( mi < localthr ) mi = localthr2;
#endif
ijppt = ijp[i] + 1;
mjpt = m + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
lastj = lgth2+1;
if( trywarp )
{
prevwmrecordspt = prevwmrecords;
wmrecordspt = wmrecords+1;
wmrecords1pt = wmrecords;
warpipt = warpi + 1;
warpjpt = warpj + 1;
}
for( j=1; j<lastj; j++ )
{
wm = *prept;
*ijppt = 0;
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=mi+fpenalty) > wm )
{
wm = g;
*ijppt = -( j - mpi );
}
if( *prept > mi )
{
mi = *prept;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=*mjpt+fpenalty) > wm )
{
wm = g;
*ijppt = +( i - *mpjpt );
}
if( *prept > *mjpt )
{
*mjpt = *prept;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
*mjpt += fpenalty_ex;
#endif
if( maxwm < wm )
{
maxwm = wm;
endali = i;
endalj = j;
}
#if 1
if( wm < localthr )
{
// fprintf( stderr, "stop i=%d, j=%d, curpt=%f, localthr = %f\n", i, j, *curpt, localthr );
*ijppt = localstop;
wm = localthr2;
}
#endif
#if 0
fprintf( stderr, "%5.0f ", *curpt );
#endif
#if 0
fprintf( stderr, "wm (%d,%d) = %5.0f\n", i, j, wm );
// fprintf( stderr, "%c-%c *ijppt = %d, localstop = %d\n", seq1[0][i], seq2[0][j], *ijppt, localstop );
#endif
if( trywarp )
{
fpenalty_tmp = fpenalty_shift + fpenalty_ex * ( i - prevwarpi[j-1] + j - prevwarpj[j-1] );
// fprintf( stderr, "fpenalty_shift = %f\n", fpenalty_tmp );
// fprintf( stderr, "\n\n\nwarp to %c-%c (%d-%d) from %c-%c (%d-%d) ? prevwmrecords[%d] = %f + %f <- wm = %f\n", seq1[0][prevwarpi[j-1]], seq2[0][prevwarpj[j-1]], prevwarpi[j-1], prevwarpj[j-1], seq1[0][i], seq2[0][j], i, j, j, prevwmrecords[j-1], fpenalty_tmp, wm );
// if( (g=prevwmrecords[j-1] + fpenalty_shift )> wm )
if( ( g=*prevwmrecordspt++ + fpenalty_tmp )> wm ) // naka ha osokute kamawanai
{
// fprintf( stderr, "Yes! Warp!! from %d-%d (%c-%c) to %d-%d (%c-%c) fpenalty_tmp = %f! warpn = %d\n", i, j, seq1[0][i], seq2[0][j-1], prevwarpi[j-1], prevwarpj[j-1],seq1[0][prevwarpi[j-1]], seq2[0][prevwarpj[j-1]], fpenalty_tmp, warpn );
if( warpn && prevwarpi[j-1] == warpis[warpn-1] && prevwarpj[j-1] == warpjs[warpn-1] )
{
*ijppt = warpbase + warpn - 1;
}
else
{
*ijppt = warpbase + warpn;
warpis = realloc( warpis, sizeof(int) * ( warpn+1 ) );
warpjs = realloc( warpjs, sizeof(int) * ( warpn+1 ) );
warpis[warpn] = prevwarpi[j-1];
warpjs[warpn] = prevwarpj[j-1];
warpn++;
}
wm = g;
}
else
{
}
curm = *curpt + wm;
// fprintf( stderr, "###### curm = %f at %c-%c, i=%d, j=%d\n", curm, seq1[0][i], seq2[0][j], i, j );
// fprintf( stderr, "copy from i, j-1? %f > %f?\n", wmrecords[j-1], curm );
// if( wmrecords[j-1] > wmrecords[j] )
if( *wmrecords1pt > *wmrecordspt )
{
// fprintf( stderr, "yes\n" );
// wmrecords[j] = wmrecords[j-1];
*wmrecordspt = *wmrecords1pt;
// warpi[j] = warpi[j-1];
// warpj[j] = warpj[j-1];
*warpipt = *(warpipt-1);
*warpjpt = *(warpjpt-1);
// fprintf( stderr, "warpi[j]=%d, warpj[j]=%d wmrecords[j] = %f\n", warpi[j], warpj[j], wmrecords[j] );
}
// else
// {
// fprintf( stderr, "no\n" );
// }
// fprintf( stderr, " curm = %f at %c-%c\n", curm, seq1[0][i], seq2[0][j] );
// fprintf( stderr, " wmrecords[%d] = %f\n", j, wmrecords[j] );
// fprintf( stderr, "replace?\n" );
// if( curm > wmrecords[j] )
if( curm > *wmrecordspt )
{
// fprintf( stderr, "yes at %d-%d (%c-%c), replaced warp: warpi[j]=%d, warpj[j]=%d warpn=%d, wmrecords[j] = %f -> %f\n", i, j, seq1[0][i], seq2[0][j], i, j, warpn, wmrecords[j], curm );
// wmrecords[j] = curm;
*wmrecordspt = curm;
// warpi[j] = i;
// warpj[j] = j;
*warpipt = i;
*warpjpt = j;
}
// else
// {
// fprintf( stderr, "No! warpi[j]=%d, warpj[j]=%d wmrecords[j] = %f\n", warpi[j], warpj[j], wmrecords[j] );
// }
// fprintf( stderr, "%d-%d (%c-%c) curm = %5.0f, wmrecords[j]=%f\n", i, j, seq1[0][i], seq2[0][j], curm, wmrecords[j] );
wmrecordspt++;
wmrecords1pt++;
warpipt++;
warpjpt++;
}
*curpt++ += wm;
ijppt++;
mjpt++;
prept++;
mpjpt++;
}
#if DEBUG2
fprintf( stderr, "\n" );
#endif
lastverticalw[i] = currentw[lgth2-1];
if( trywarp )
{
fltncpy( prevwmrecords, wmrecords, lastj );
intncpy( prevwarpi, warpi, lastj );
intncpy( prevwarpj, warpj, lastj );
}
}
// fprintf( stderr, "\nwm = %f\n", wm );
if( trywarp )
{
// if( warpn ) fprintf( stderr, "warpn = %d\n", warpn );
free( wmrecords );
free( prevwmrecords );
free( warpi );
free( warpj );
free( prevwarpi );
free( prevwarpj );
}
#if 0
fprintf( stderr, "maxwm = %f\n", maxwm );
fprintf( stderr, "endali = %d\n", endali );
fprintf( stderr, "endalj = %d\n", endalj );
#endif
if( ijp[endali][endalj] == localstop )
{
strcpy( seq1[0], "" );
strcpy( seq2[0], "" );
*off1pt = *off2pt = 0;
fprintf( stderr, "maxwm <- 0.0 \n" );
return( 0.0 );
}
Ltracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, ijp, off1pt, off2pt, endali, endalj, warpis, warpjs, warpbase );
if( warpis ) free( warpis );
if( warpjs ) free( warpjs );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
strcpy( seq1[0], mseq1[0] );
strcpy( seq2[0], mseq2[0] );
#if 0
fprintf( stderr, "wm=%f\n", wm );
fprintf( stderr, ">\n%s\n", mseq1[0] );
fprintf( stderr, ">\n%s\n", mseq2[0] );
fprintf( stderr, "maxwm = %f\n", maxwm );
fprintf( stderr, " wm = %f\n", wm );
#endif
return( maxwm );
}
float G__align11( char **seq1, char **seq2, int alloclen )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
register int i, j;
int lasti; /* outgap == 0 -> lgth1, outgap == 1 -> lgth1+1 */
int lgth1, lgth2;
int resultlen;
float wm; /* int ?????? */
float g;
float *currentw, *previousw;
float fpenalty = (float)penalty;
#if USE_PENALTY_EX
float fpenalty_ex = (float)penalty_ex;
#endif
#if 1
float *wtmp;
int *ijppt;
float *mjpt, *prept, *curpt;
int *mpjpt;
#endif
static float mi, *m;
static int **ijp;
static int mpi, *mp;
static float *w1, *w2;
static float *match;
static float *initverticalw; /* kufuu sureba iranai */
static float *lastverticalw; /* kufuu sureba iranai */
static char **mseq1;
static char **mseq2;
static char **mseq;
static int **intwork;
static float **floatwork;
static int orlgth1 = 0, orlgth2 = 0;
wm = 0.0;
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
if( lgth1 <= 0 || lgth2 <= 0 )
{
fprintf( stderr, "WARNING (g11): lgth1=%d, lgth2=%d\n", lgth1, lgth2 );
}
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
FreeFloatMtx( floatwork );
FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 );
intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 );
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijp = commonIP;
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc( initverticalw, seq2, seq1, 0, lgth1 );
match_calc( currentw, seq1, seq2, 0, lgth2 );
if( outgap == 1 )
{
for( i=1; i<lgth1+1; i++ )
{
initverticalw[i] += fpenalty;
}
for( j=1; j<lgth2+1; j++ )
{
currentw[j] += fpenalty;
}
}
for( j=1; j<lgth2+1; ++j )
{
m[j] = currentw[j-1]; mp[j] = 0;
}
if( lgth2 == 0 )
lastverticalw[0] = 0.0; // lgth2==0 no toki error
else
lastverticalw[0] = currentw[lgth2-1]; // lgth2==0 no toki error
if( outgap ) lasti = lgth1+1; else lasti = lgth1;
#if XXXXXXX
fprintf( stderr, "currentw = \n" );
for( i=0; i<lgth1+1; i++ )
{
fprintf( stderr, "%5.2f ", currentw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "initverticalw = \n" );
for( i=0; i<lgth2+1; i++ )
{
fprintf( stderr, "%5.2f ", initverticalw[i] );
}
fprintf( stderr, "\n" );
#endif
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc( currentw, seq1, seq2, i, lgth2 );
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
currentw[0] = initverticalw[i];
mi = previousw[0]; mpi = 0;
ijppt = ijp[i] + 1;
mjpt = m + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
for( j=1; j<lgth2+1; j++ )
{
wm = *prept;
*ijppt = 0;
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=mi+fpenalty) > wm )
{
wm = g;
*ijppt = -( j - mpi );
}
if( (g=*prept) >= mi )
{
mi = g;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( (g=*mjpt + fpenalty) > wm )
{
wm = g;
*ijppt = +( i - *mpjpt );
}
if( (g=*prept) >= *mjpt )
{
*mjpt = g;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
m[j] += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f ", wm );
#endif
*curpt++ += wm;
ijppt++;
mjpt++;
prept++;
mpjpt++;
}
lastverticalw[i] = currentw[lgth2-1]; // lgth2==0 no toki error
}
Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, ijp );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
strcpy( seq1[0], mseq1[0] );
strcpy( seq2[0], mseq2[0] );
#if 0
fprintf( stderr, "\n" );
fprintf( stderr, ">\n%s\n", mseq1[0] );
fprintf( stderr, ">\n%s\n", mseq2[0] );
fprintf( stderr, "wm = %f\n", wm );
#endif
return( wm );
}
float MSalign11( char **seq1, char **seq2, int alloclen )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
register int i, j;
int lasti, lastj;
int iin = 0, jin = 0; // by Mathog, a guess
int lgth1, lgth2;
int resultlen;
float wm = 0.0; /* int ?????? */
float g;
float *currentw, *previousw;
float fpenalty = (float)penalty;
#if USE_PENALTY_EX
float fpenalty_ex = (float)penalty_ex;
#endif
float *maxinw = NULL, *maxinwpt = NULL; // by D.Mathog, guess
float *maxinh = NULL; // by D.Mathog, guess
#if 1
float wmmax;
float *wtmp;
int *ijppt;
float *mjpt, *prept, *curpt;
int *mpjpt;
#endif
static float mi, *m;
static int **ijp;
static int mpi, *mp;
static float *w1, *w2;
static float *match;
static float *initverticalw; /* kufuu sureba iranai */
static float *lastverticalw; /* kufuu sureba iranai */
static char **mseq1;
static char **mseq2;
static char **mseq;
static float **cpmx1;
static float **cpmx2;
static int **intwork;
static float **WMMTX;
static float **floatwork;
static int orlgth1 = 0, orlgth2 = 0;
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( maxinw );
FreeFloatVec( maxinh );
FreeFloatVec( m );
FreeIntVec( mp );
FreeCharMtx( mseq );
FreeFloatMtx( cpmx1 );
FreeFloatMtx( cpmx2 );
FreeFloatMtx( floatwork );
FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
maxinw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
maxinh = AllocateFloatVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
cpmx1 = AllocateFloatMtx( 26, ll1+2 );
cpmx2 = AllocateFloatMtx( 26, ll2+2 );
floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 );
intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 );
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
FreeFloatMtx( WMMTX );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
WMMTX = AllocateFloatMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijp = commonIP;
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc( initverticalw, seq2, seq1, 0, lgth1 );
match_calc( currentw, seq1, seq2, 0, lgth2 );
WMMTX[0][0] = initverticalw[0];
maxinh[0] = initverticalw[0];
for( i=1; i<lgth1+1; i++ )
{
initverticalw[i] += fpenalty;
WMMTX[i][0] = initverticalw[i];
if( maxinh[0] < initverticalw[i] ) maxinh[0] = initverticalw[i];
}
maxinw[0] = currentw[0];
for( j=1; j<lgth2+1; j++ )
{
currentw[j] += fpenalty;
WMMTX[0][j] = currentw[j];
if( maxinw[0] < currentw[j] ) maxinw[0] = currentw[j];
}
for( j=1; j<lgth2+1; ++j )
{
m[j] = currentw[j-1]; mp[j] = 0;
}
lastverticalw[0] = currentw[lgth2-1];
lasti = lgth1+1;
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc( currentw, seq1, seq2, i, lgth2 );
currentw[0] = initverticalw[i];
mi = previousw[0]; mpi = 0;
maxinwpt = maxinw + i;
*maxinwpt = currentw[0];
fprintf( stderr, "currentw[0] = %f, *maxinwpt = %f\n", currentw[0], maxinw[i] );
ijppt = ijp[i] + 1;
mjpt = m + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
lastj = lgth2+1;
for( j=1; j<lastj; j++ )
{
wm = *prept;
*ijppt = 0;
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
g = mi + fpenalty;
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
*ijppt = -( j - mpi );
}
g = *prept;
if( g >= mi )
{
mi = g;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
g = *mjpt + fpenalty;
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
*ijppt = +( i - *mpjpt );
}
g = *prept;
if( g >= *mjpt )
{
*mjpt = g;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
m[j] += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f ", wm );
#endif
*curpt += wm;
WMMTX[i][j] = *curpt;
if( j<lgth2 && *maxinwpt < *curpt ) *maxinwpt = *curpt;
if( j<lgth2 && maxinh[j] < *curpt ) maxinh[j] = *curpt;
// fprintf( stderr, "maxintwpt = %f\n", *maxinwpt );
ijppt++;
mjpt++;
prept++;
mpjpt++;
curpt++;
}
lastverticalw[i] = currentw[lgth2-1];
}
wmmax = -999.9;
for( i=0; i<lgth1; i++ )
{
g = lastverticalw[i];
if( g > wmmax )
{
wmmax = g;
iin = i;
jin = lgth2-1;
}
}
for( j=0; j<lgth2; j++ )
{
g = currentw[j];
if( g > wmmax )
{
wmmax = g;
iin = lgth1-1;
jin = j;
}
}
for( i=0; i<lgth1; i++ )
fprintf( stderr, "maxinw[%d] = %f\n", i, maxinw[i] );
for( j=0; j<lgth2; j++ )
fprintf( stderr, "maxinh[%d] = %f\n", j, maxinh[j] );
fprintf( stderr, "wmmax = %f (%d,%d)\n", wmmax, iin, jin );
if( iin == lgth1 - 1 && jin == lgth2 - 1 )
;
else
wmmax += fpenalty;
fprintf( stderr, "wmmax = %f\n", wmmax );
#if 0
for( i=0; i<lgth1; i++ )
{
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "% 10.2f ", WMMTX[i][j] );
}
fprintf( stderr, "\n" );
}
#endif
mseq1[0] += lgth1+lgth2;
*mseq1[0] = 0;
mseq2[0] += lgth1+lgth2;
*mseq2[0] = 0;
backdp( WMMTX, wmmax, maxinw, maxinh, lgth1, lgth2, alloclen, w1, w2, initverticalw, m, mp, iin, jin, seq1, seq2, mseq1, mseq2 );
fprintf( stderr, "\n" );
#if 1
fprintf( stderr, "\n" );
fprintf( stderr, ">MSres\n%s\n", mseq1[0] );
fprintf( stderr, ">MSres\n%s\n", mseq2[0] );
#endif
#if 0
for( i=0; i<lgth1; i++ )
{
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "% 10.2f ", WMMTX[i][j] );
}
fprintf( stderr, "\n" );
}
#endif
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
mseq1[0] += lgth1+lgth2;
*mseq1[0] = 0;
mseq2[0] += lgth1+lgth2;
*mseq2[0] = 0;
Atracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijp );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
strcpy( seq1[0], mseq1[0] );
strcpy( seq2[0], mseq2[0] );
#if 1
fprintf( stderr, "\n" );
fprintf( stderr, ">\n%s\n", mseq1[0] );
fprintf( stderr, ">\n%s\n", mseq2[0] );
#endif
return( wm );
}
float genL__align11( char **seq1, char **seq2, int alloclen, int *off1pt, int *off2pt )
/* score no keisan no sai motokaraaru gap no atukai ni mondai ga aru */
{
// int k;
register int i, j;
int lasti, lastj;
int lgth1, lgth2;
int resultlen;
float wm = 0.0; /* int ?????? */
float g;
float *currentw, *previousw;
#if 1
float *wtmp;
int *ijpipt;
int *ijpjpt;
float *mjpt, *Mjpt, *prept, *curpt;
int *mpjpt, *Mpjpt;
#endif
static TLS float mi, *m;
static TLS float Mi, *largeM;
static TLS int **ijpi;
static TLS int **ijpj;
static TLS int mpi, *mp;
static TLS int Mpi, *Mp;
static TLS float *w1, *w2;
static TLS float *match;
static TLS float *initverticalw; /* kufuu sureba iranai */
static TLS float *lastverticalw; /* kufuu sureba iranai */
static TLS char **mseq1;
static TLS char **mseq2;
static TLS char **mseq;
static TLS float **cpmx1;
static TLS float **cpmx2;
static TLS int **intwork;
static TLS float **floatwork;
static TLS int orlgth1 = 0, orlgth2 = 0;
float maxwm;
float tbk;
int tbki, tbkj;
int endali, endalj;
// float localthr = 0.0;
// float localthr2 = 0.0;
float fpenalty = (float)penalty;
float fpenalty_OP = (float)penalty_OP;
float fpenalty_ex = (float)penalty_ex;
// float fpenalty_EX = (float)penalty_EX;
float foffset = (float)offset;
float localthr = -foffset;
float localthr2 = -foffset;
if( seq1 == NULL )
{
if( orlgth1 > 0 && orlgth2 > 0 )
{
orlgth1 = 0;
orlgth2 = 0;
free( mseq1 );
free( mseq2 );
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
free( largeM );
free( Mp );
FreeCharMtx( mseq );
FreeFloatMtx( cpmx1 );
FreeFloatMtx( cpmx2 );
FreeFloatMtx( floatwork );
FreeIntMtx( intwork );
}
return( 0.0 );
}
// fprintf( stderr, "@@@@@@@@@@@@@ penalty_OP = %f, penalty_EX = %f, pelanty = %f\n", fpenalty_OP, fpenalty_EX, fpenalty );
if( orlgth1 == 0 )
{
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
}
lgth1 = strlen( seq1[0] );
lgth2 = strlen( seq2[0] );
if( lgth1 > orlgth1 || lgth2 > orlgth2 )
{
int ll1, ll2;
if( orlgth1 > 0 && orlgth2 > 0 )
{
FreeFloatVec( w1 );
FreeFloatVec( w2 );
FreeFloatVec( match );
FreeFloatVec( initverticalw );
FreeFloatVec( lastverticalw );
FreeFloatVec( m );
FreeIntVec( mp );
FreeFloatVec( largeM );
FreeIntVec( Mp );
FreeCharMtx( mseq );
FreeFloatMtx( cpmx1 );
FreeFloatMtx( cpmx2 );
FreeFloatMtx( floatwork );
FreeIntMtx( intwork );
}
ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100;
ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100;
#if DEBUG
fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 );
#endif
w1 = AllocateFloatVec( ll2+2 );
w2 = AllocateFloatVec( ll2+2 );
match = AllocateFloatVec( ll2+2 );
initverticalw = AllocateFloatVec( ll1+2 );
lastverticalw = AllocateFloatVec( ll1+2 );
m = AllocateFloatVec( ll2+2 );
mp = AllocateIntVec( ll2+2 );
largeM = AllocateFloatVec( ll2+2 );
Mp = AllocateIntVec( ll2+2 );
mseq = AllocateCharMtx( njob, ll1+ll2 );
cpmx1 = AllocateFloatMtx( nalphabets, ll1+2 );
cpmx2 = AllocateFloatMtx( nalphabets, ll2+2 );
floatwork = AllocateFloatMtx( nalphabets, MAX( ll1, ll2 )+2 );
intwork = AllocateIntMtx( nalphabets, MAX( ll1, ll2 )+2 );
#if DEBUG
fprintf( stderr, "succeeded\n" );
#endif
orlgth1 = ll1 - 100;
orlgth2 = ll2 - 100;
}
mseq1[0] = mseq[0];
mseq2[0] = mseq[1];
if( orlgth1 > commonAlloc1 || orlgth2 > commonAlloc2 )
{
int ll1, ll2;
if( commonAlloc1 && commonAlloc2 )
{
FreeIntMtx( commonIP );
FreeIntMtx( commonJP );
}
ll1 = MAX( orlgth1, commonAlloc1 );
ll2 = MAX( orlgth2, commonAlloc2 );
#if DEBUG
fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 );
#endif
commonIP = AllocateIntMtx( ll1+10, ll2+10 );
commonJP = AllocateIntMtx( ll1+10, ll2+10 );
#if DEBUG
fprintf( stderr, "succeeded\n\n" );
#endif
commonAlloc1 = ll1;
commonAlloc2 = ll2;
}
ijpi = commonIP;
ijpj = commonJP;
#if 0
for( i=0; i<lgth1; i++ )
fprintf( stderr, "ogcp1[%d]=%f\n", i, ogcp1[i] );
#endif
currentw = w1;
previousw = w2;
match_calc( initverticalw, seq2, seq1, 0, lgth1 );
match_calc( currentw, seq1, seq2, 0, lgth2 );
lasti = lgth2+1;
for( j=1; j<lasti; ++j )
{
m[j] = currentw[j-1]; mp[j] = 0;
largeM[j] = currentw[j-1]; Mp[j] = 0;
}
lastverticalw[0] = currentw[lgth2-1];
#if 0
fprintf( stderr, "currentw = \n" );
for( i=0; i<lgth1+1; i++ )
{
fprintf( stderr, "%5.2f ", currentw[i] );
}
fprintf( stderr, "\n" );
fprintf( stderr, "initverticalw = \n" );
for( i=0; i<lgth2+1; i++ )
{
fprintf( stderr, "%5.2f ", initverticalw[i] );
}
fprintf( stderr, "\n" );
#endif
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, " " );
for( j=0; j<lgth2+1; j++ )
fprintf( stderr, "%c ", seq2[0][j] );
fprintf( stderr, "\n" );
#endif
localstop = lgth1+lgth2+1;
maxwm = -999999999.9;
endali = endalj = 0;
#if DEBUG2
fprintf( stderr, "\n" );
fprintf( stderr, "%c ", seq1[0][0] );
for( j=0; j<lgth2+1; j++ )
fprintf( stderr, "%5.0f ", currentw[j] );
fprintf( stderr, "\n" );
#endif
lasti = lgth1+1;
for( i=1; i<lasti; i++ )
{
wtmp = previousw;
previousw = currentw;
currentw = wtmp;
previousw[0] = initverticalw[i-1];
match_calc( currentw, seq1, seq2, i, lgth2 );
#if DEBUG2
fprintf( stderr, "%c ", seq1[0][i] );
fprintf( stderr, "%5.0f ", currentw[0] );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
#if XXXXXXX
fprintf( stderr, "\n" );
fprintf( stderr, "i=%d\n", i );
fprintf( stderr, "currentw = \n" );
for( j=0; j<lgth2; j++ )
{
fprintf( stderr, "%5.2f ", currentw[j] );
}
fprintf( stderr, "\n" );
#endif
currentw[0] = initverticalw[i];
mi = previousw[0]; mpi = 0;
Mi = previousw[0]; Mpi = 0;
#if 0
if( mi < localthr ) mi = localthr2;
#endif
ijpipt = ijpi[i] + 1;
ijpjpt = ijpj[i] + 1;
mjpt = m + 1;
Mjpt = largeM + 1;
prept = previousw;
curpt = currentw + 1;
mpjpt = mp + 1;
Mpjpt = Mp + 1;
tbk = -999999.9;
tbki = 0;
tbkj = 0;
lastj = lgth2+1;
for( j=1; j<lastj; j++ )
{
wm = *prept;
*ijpipt = i-1;
*ijpjpt = j-1;
// fprintf( stderr, "i,j=%d,%d %c-%c\n", i, j, seq1[0][i], seq2[0][j] );
// fprintf( stderr, "wm=%f\n", wm );
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
g = mi + fpenalty;
#if 0
fprintf( stderr, "%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
// *ijpipt = i - 1;
*ijpjpt = mpi;
}
g = *prept;
if( g > mi )
{
mi = g;
mpi = j-1;
}
#if USE_PENALTY_EX
mi += fpenalty_ex;
#endif
#if 0
fprintf( stderr, "%5.0f->", wm );
#endif
g = *mjpt + fpenalty;
#if 0
fprintf( stderr, "m%5.0f?", g );
#endif
if( g > wm )
{
wm = g;
*ijpipt = *mpjpt;
*ijpjpt = j - 1; //IRU!
}
g = *prept;
if( g > *mjpt )
{
*mjpt = g;
*mpjpt = i-1;
}
#if USE_PENALTY_EX
*mjpt += fpenalty_ex;
#endif
g = tbk + fpenalty_OP;
// g = tbk;
if( g > wm )
{
wm = g;
*ijpipt = tbki;
*ijpjpt = tbkj;
// fprintf( stderr, "hit! i%d, j%d, ijpi = %d, ijpj = %d\n", i, j, *ijpipt, *ijpjpt );
}
// g = Mi;
if( Mi > tbk )
{
tbk = Mi; //error desu.
tbki = i-1;
tbkj = Mpi;
}
// g = *Mjpt;
if( *Mjpt > tbk )
{
tbk = *Mjpt;
tbki = *Mpjpt;
tbkj = j-1;
}
// tbk += fpenalty_EX;// + foffset;
// g = *prept;
if( *prept > *Mjpt )
{
*Mjpt = *prept;
*Mpjpt = i-1;
}
// *Mjpt += fpenalty_EX;// + foffset;
// g = *prept;
if( *prept > Mi )
{
Mi = *prept;
Mpi = j-1;
}
// Mi += fpenalty_EX;// + foffset;
// fprintf( stderr, "wm=%f, tbk=%f(%c-%c), mi=%f, *mjpt=%f\n", wm, tbk, seq1[0][tbki], seq2[0][tbkj], mi, *mjpt );
// fprintf( stderr, "ijp = %c,%c\n", seq1[0][abs(*ijpipt)], seq2[0][abs(*ijpjpt)] );
if( maxwm < wm )
{
maxwm = wm;
endali = i;
endalj = j;
}
#if 1
if( wm < localthr )
{
// fprintf( stderr, "stop i=%d, j=%d, curpt=%f\n", i, j, *curpt );
*ijpipt = localstop;
// *ijpjpt = localstop;
wm = localthr2;
}
#endif
#if 0
fprintf( stderr, "%5.0f ", *curpt );
#endif
#if DEBUG2
fprintf( stderr, "%5.0f ", wm );
// fprintf( stderr, "%c-%c *ijppt = %d, localstop = %d\n", seq1[0][i], seq2[0][j], *ijppt, localstop );
#endif
*curpt += wm;
ijpipt++;
ijpjpt++;
mjpt++;
Mjpt++;
prept++;
mpjpt++;
Mpjpt++;
curpt++;
}
#if DEBUG2
fprintf( stderr, "\n" );
#endif
lastverticalw[i] = currentw[lgth2-1];
}
#if DEBUG2
fprintf( stderr, "maxwm = %f\n", maxwm );
fprintf( stderr, "endali = %d\n", endali );
fprintf( stderr, "endalj = %d\n", endalj );
#endif
if( ijpi[endali][endalj] == localstop ) // && ijpj[endali][endalj] == localstop )
{
strcpy( seq1[0], "" );
strcpy( seq2[0], "" );
*off1pt = *off2pt = 0;
return( 0.0 );
}
gentracking( currentw, lastverticalw, seq1, seq2, mseq1, mseq2, cpmx1, cpmx2, ijpi, ijpj, off1pt, off2pt, endali, endalj );
// fprintf( stderr, "### impmatch = %f\n", *impmatch );
resultlen = strlen( mseq1[0] );
if( alloclen < resultlen || resultlen > N )
{
fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N );
ErrorExit( "LENGTH OVER!\n" );
}
strcpy( seq1[0], mseq1[0] );
strcpy( seq2[0], mseq2[0] );
#if 0
fprintf( stderr, "\n" );
fprintf( stderr, ">\n%s\n", mseq1[0] );
fprintf( stderr, ">\n%s\n", mseq2[0] );
#endif
return( maxwm );
}
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 );
*/
}
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 );
}
int main( int argc, char *argv[] )
{
char **argv2;
static int *nlen;
static char **name, **seq;
static char **seq1, **seq2;
static char **mseq1, **mseq2;
static char **aseq;
static char **bseq;
static double **pscore;
static double *eff;
int i, j, len1, len2;
static int ***topol;
static double **len;
FILE *gp1, *gp2;
char c;
int nlenmax1, nlenmax2, nseq1, nseq2;
int alloclen;
argv2 = arguments( argc, argv );
fprintf( stderr, "####### in galn\n" );
initFiles();
fprintf( stderr, "file1 = %s\n", argv2[0] );
fprintf( stderr, "file2 = %s\n", argv2[1] );
gp1 = fopen( argv2[0], "r" ); if( !gp1 ) ErrorExit( "cannot open file1" );
gp2 = fopen( argv2[1], "r" ); if( !gp2 ) ErrorExit( "cannot open file2" );
#if 0
PreRead( gp1, &nseq1, &nlenmax1 );
PreRead( gp2, &nseq2, &nlenmax2 );
#else
getnumlen( gp1 );
nseq1 = njob; nlenmax1 = nlenmax;
getnumlen( gp2 );
nseq2 = njob; nlenmax2 = nlenmax;
#endif
njob = nseq1 + nseq2;
nlenmax = MAX( nlenmax1, nlenmax2 );
rewind( gp1 );
rewind( gp2 );
name = AllocateCharMtx( njob, B );
nlen = AllocateIntVec( njob );
seq1 = AllocateCharMtx( nseq1, nlenmax*3 );
seq2 = AllocateCharMtx( nseq2, nlenmax*3 );
seq = AllocateCharMtx( njob, 1 );
aseq = AllocateCharMtx( njob, nlenmax*3 );
bseq = AllocateCharMtx( njob, nlenmax*3 );
mseq1 = AllocateCharMtx( njob, 1 );
mseq2 = AllocateCharMtx( njob, 1 );
alloclen = nlenmax * 3;
topol = AllocateIntCub( njob, 2, njob );
len = AllocateDoubleMtx( njob, 2 );
pscore = AllocateDoubleMtx( njob, njob );
eff = AllocateDoubleVec( njob );
#if 0
njob=nseq2; FRead( gp2, name+nseq1, nlen+nseq1, seq2 );
njob=nseq1; FRead( gp1, name, nlen, seq1 );
#else
njob=nseq2; readDataforgaln( gp2, name+nseq1, nlen+nseq1, seq2 );
njob=nseq1; readDataforgaln( gp1, name, nlen, seq1 );
#endif
njob = nseq1 + nseq2;
#if 0 // CHUUI
commongappick( nseq1, seq1 );
commongappick( nseq2, seq2 );
#endif
for( i=0; i<nseq1; i++ ) seq[i] = seq1[i];
for( i=nseq1; i<njob; i++ ) seq[i] = seq2[i-nseq1];
/*
Write( stdout, njob, name, nlen, seq );
*/
constants( njob, seq );
WriteOptions( trap_g );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illeagal character %c\n", c );
exit( 1 );
}
for( i=1; i<nseq1; i++ )
{
if( nlen[i] != nlen[0] )
ErrorExit( "group1 is not aligned." );
}
for( i=nseq1+1; i<njob; i++ )
{
if( nlen[i] != nlen[nseq1] )
ErrorExit( "group2 is not aligned." );
}
if( tbutree == 0 )
{
for( i=0; i<nseq1; i++ )
{
for( j=i+1; j<nseq1; j++ )
{
pscore[i][j] = (double)substitution_hosei( seq[i], seq[j] );
// fprintf( stderr, "%d-%d, %5.1f \n", i, j, pscore[i][j] );
}
for( j=nseq1; j<njob; j++ )
{
pscore[i][j] = 3.0;
// fprintf( stderr, "%d-%d, %5.1f \n", i, j, pscore[i][j] );
}
}
for( i=nseq1; i<njob-1; i++ )
{
for( j=i+1; j<njob; j++ )
{
pscore[i][j] = (double)substitution_hosei( seq[i], seq[j] );
// fprintf( stderr, "%d-%d, %5.1f \n", i, j, pscore[i][j] );
}
}
// fprintf( stderr, "\n" );
}
else
{
fprintf( stderr, "Not supported\n" );
exit( 1 );
#if 0
prep = fopen( "hat2", "r" );
if( prep == NULL ) ErrorExit( "Make hat2." );
readhat2( prep, njob, name, pscore );
fclose( prep );
#endif
}
fprintf( stderr, "Constructing dendrogram ... " );
if( treemethod == 'x' )
veryfastsupg( njob, pscore, topol, len );
else
ErrorExit( "Incorrect tree\n" );
fprintf( stderr, "done.\n" );
if( tbrweight )
{
weight = 3;
counteff_simple( njob, topol, len, eff );
// for( i=0; i<njob; i++ ) fprintf( stderr, "eff[%d] = %f\n", i, eff[i] );
}
else
{
for( i=0; i<njob; i++ ) eff[i] = 1.0;
}
len1 = strlen( seq[0] );
len2 = strlen( seq[nseq1] );
if( len1 > 30000 || len2 > 30000 )
{
fprintf( stderr, "\nlen1=%d, len2=%d, Switching to the memsave mode.\n", len1, len2 );
alg = 'M';
}
GroupAlign( nseq1, nseq2, name, nlen, seq, aseq, mseq1, mseq2, topol, len, eff, alloclen );
#if 0
writePre( njob, name, nlen, aseq, 1 );
#else
writeDataforgaln( stdout, njob, name, nlen, aseq );
#endif
SHOWVERSION;
return( 0 );
}
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 );
}
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
}
int main( int argc, char *argv[] )
{
static int nlen[M];
static char name[M][B], **seq;
static char **mseq1, **mseq2;
static char **aseq;
static char **bseq;
static double *eff;
int i;
FILE *infp;
char c;
int alloclen;
arguments( argc, argv );
if( inputfile )
{
infp = fopen( inputfile, "r" );
if( !infp )
{
fprintf( stderr, "Cannot open %s\n", inputfile );
exit( 1 );
}
}
else
infp = stdin;
getnumlen( infp );
rewind( infp );
if( njob < 2 )
{
fprintf( stderr, "At least 2 sequences should be input!\n"
"Only %d sequence found.\n", njob );
exit( 1 );
}
if( njob > M )
{
fprintf( stderr, "The number of sequences must be < %d\n", M );
fprintf( stderr, "Please try the splittbfast program for such large data.\n" );
exit( 1 );
}
seq = AllocateCharMtx( njob, nlenmax*9+1 );
aseq = AllocateCharMtx( njob, nlenmax*9+1 );
bseq = AllocateCharMtx( njob, nlenmax*9+1 );
mseq1 = AllocateCharMtx( njob, 0 );
mseq2 = AllocateCharMtx( njob, 0 );
alloclen = nlenmax*9;
eff = AllocateDoubleVec( njob );
#if 0
Read( name, nlen, seq );
#else
readData( infp, name, nlen, seq );
#endif
fclose( infp );
constants( njob, seq );
#if 0
fprintf( stderr, "params = %d, %d, %d\n", penalty, penalty_ex, offset );
#endif
initSignalSM();
initFiles();
WriteOptions( trap_g );
c = seqcheck( seq );
if( c )
{
fprintf( stderr, "Illegal character %c\n", c );
exit( 1 );
}
// writePre( njob, name, nlen, seq, 0 );
for( i=0; i<njob; i++ ) eff[i] = 1.0;
for( i=0; i<njob; i++ ) gappick0( bseq[i], seq[i] );
pairalign( name, nlen, bseq, aseq, mseq1, mseq2, eff, alloclen );
fprintf( trap_g, "done.\n" );
#if DEBUG
fprintf( stderr, "closing trap_g\n" );
#endif
fclose( trap_g );
// writePre( njob, name, nlen, aseq, !contin );
#if 0
writeData( stdout, njob, name, nlen, aseq );
#endif
#if IODEBUG
fprintf( stderr, "OSHIMAI\n" );
#endif
SHOWVERSION;
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
}
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 );
}
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 );
}
