#include "mltaln.h" #define DEBUG 0 #define IODEBUG 0 #define SCOREOUT 0 static int treein; static int topin; static int treeout; static int distout; static int noalign; void arguments( int argc, char *argv[] ) { int c; treein = 0; topin = 0; rnaprediction = 'm'; rnakozo = 0; nevermemsave = 0; inputfile = NULL; fftkeika = 0; constraint = 0; nblosum = 62; fmodel = 0; calledByXced = 0; devide = 0; use_fft = 0; // chuui force_fft = 0; fftscore = 1; fftRepeatStop = 0; fftNoAnchStop = 0; weight = 3; utree = 1; tbutree = 1; refine = 0; check = 1; cut = 0.0; disp = 0; outgap = 1; alg = 'A'; mix = 0; tbitr = 0; scmtd = 5; tbweight = 0; tbrweight = 3; checkC = 0; treemethod = 'X'; contin = 0; scoremtx = 1; kobetsubunkatsu = 0; dorp = NOTSPECIFIED; ppenalty = NOTSPECIFIED; ppenalty_ex = NOTSPECIFIED; poffset = NOTSPECIFIED; kimuraR = NOTSPECIFIED; pamN = NOTSPECIFIED; geta2 = GETA2; fftWinSize = NOTSPECIFIED; fftThreshold = NOTSPECIFIED; RNAppenalty = NOTSPECIFIED; RNAppenalty_ex = NOTSPECIFIED; RNApthr = NOTSPECIFIED; TMorJTT = JTT; consweight_multi = 1.0; consweight_rna = 0.0; while( --argc > 0 && (*++argv)[0] == '-' ) { while ( ( c = *++argv[0] ) ) { switch( c ) { case 'i': inputfile = *++argv; fprintf( stderr, "inputfile = %s\n", inputfile ); --argc; goto nextoption; case 'e': RNApthr = (int)( atof( *++argv ) * 1000 - 0.5 ); --argc; goto nextoption; case 'o': RNAppenalty = (int)( atof( *++argv ) * 1000 - 0.5 ); --argc; goto nextoption; case 'f': ppenalty = (int)( atof( *++argv ) * 1000 - 0.5 ); // fprintf( stderr, "ppenalty = %d\n", ppenalty ); --argc; goto nextoption; case 'g': ppenalty_ex = (int)( atof( *++argv ) * 1000 - 0.5 ); fprintf( stderr, "ppenalty_ex = %d\n", ppenalty_ex ); --argc; goto nextoption; case 'h': poffset = (int)( atof( *++argv ) * 1000 - 0.5 ); // fprintf( stderr, "poffset = %d\n", poffset ); --argc; goto nextoption; case 'k': kimuraR = atoi( *++argv ); fprintf( stderr, "kappa = %d\n", kimuraR ); --argc; goto nextoption; case 'b': nblosum = atoi( *++argv ); scoremtx = 1; fprintf( stderr, "blosum %d / kimura 200\n", nblosum ); --argc; goto nextoption; case 'j': pamN = atoi( *++argv ); scoremtx = 0; TMorJTT = JTT; fprintf( stderr, "jtt/kimura %d\n", pamN ); --argc; goto nextoption; case 'm': pamN = atoi( *++argv ); scoremtx = 0; TMorJTT = TM; fprintf( stderr, "tm %d\n", pamN ); --argc; goto nextoption; case 'l': fastathreshold = atof( *++argv ); constraint = 2; --argc; goto nextoption; case 'r': consweight_rna = atof( *++argv ); rnakozo = 1; --argc; goto nextoption; case 'c': consweight_multi = atof( *++argv ); --argc; goto nextoption; case 'R': rnaprediction = 'r'; break; case 's': RNAscoremtx = 'r'; break; #if 1 case 'a': fmodel = 1; break; #endif case 'y': distout = 1; break; case 't': treeout = 1; break; case 'T': noalign = 1; break; case 'D': dorp = 'd'; break; case 'P': dorp = 'p'; break; case 'O': fftNoAnchStop = 1; break; #if 0 case 'e': fftscore = 0; break; case 'r': fmodel = -1; break; case 'R': fftRepeatStop = 1; break; case 's': treemethod = 's'; break; #endif case 'X': treemethod = 'X'; break; case 'E': treemethod = 'E'; break; case 'q': treemethod = 'q'; break; #if 0 case 'a': alg = 'a'; break; #endif case 'Q': alg = 'Q'; break; case 'H': alg = 'H'; break; case 'A': alg = 'A'; break; case 'S': alg = 'S'; break; case 'M': alg = 'M'; break; case 'N': nevermemsave = 1; break; case 'B': break; case 'C': alg = 'C'; break; case 'F': use_fft = 1; break; case 'G': force_fft = 1; use_fft = 1; break; case 'U': treein = 1; break; case 'V': topin = 1; break; case 'u': tbrweight = 0; weight = 0; break; case 'v': tbrweight = 3; break; case 'd': disp = 1; break; #if 0 case 'o': outgap = 0; break; #endif /* Modified 01/08/27, default: user tree */ case 'J': tbutree = 0; break; /* modification end. */ case 'z': fftThreshold = atoi( *++argv ); --argc; goto nextoption; case 'w': fftWinSize = atoi( *++argv ); --argc; goto nextoption; case 'Z': checkC = 1; break; default: fprintf( stderr, "illegal option %c\n", c ); argc = 0; break; } } nextoption: ; } if( argc == 1 ) { cut = atof( (*argv) ); argc--; } if( argc != 0 ) { fprintf( stderr, "options: Check source file !\n" ); exit( 1 ); } if( tbitr == 1 && outgap == 0 ) { fprintf( stderr, "conflicting options : o, m or u\n" ); exit( 1 ); } if( alg == 'C' && outgap == 0 ) { fprintf( stderr, "conflicting options : C, o\n" ); exit( 1 ); } } void treebase( int nlen[M], char **aseq, char **mseq1, char **mseq2, int ***topol, double *effarr, int *alloclen, LocalHom **localhomtable, RNApair ***singlerna, double *effarr_kozo ) { int i, l; int len1, len2; int clus1, clus2; float pscore, tscore; static char *indication1, *indication2; static double *effarr1 = NULL; static double *effarr2 = NULL; static double *effarr1_kozo = NULL; static double *effarr2_kozo = NULL; static LocalHom ***localhomshrink = NULL; static int *fftlog; int m1, m2; float dumfl = 0.0; int ffttry; RNApair ***grouprna1, ***grouprna2; if( rnakozo && rnaprediction == 'm' ) { grouprna1 = (RNApair ***)calloc( njob, sizeof( RNApair ** ) ); grouprna2 = (RNApair ***)calloc( njob, sizeof( RNApair ** ) ); } else { grouprna1 = grouprna2 = NULL; } if( effarr1 == NULL ) { fftlog = AllocateIntVec( njob ); effarr1 = AllocateDoubleVec( njob ); effarr2 = AllocateDoubleVec( njob ); indication1 = AllocateCharVec( 150 ); indication2 = AllocateCharVec( 150 ); #if 0 #else if( constraint ) { localhomshrink = (LocalHom ***)calloc( njob, sizeof( LocalHom ** ) ); for( i=0; i 66 ) fprintf( stderr, "..." ); fprintf( stderr, "\n" ); fprintf( stderr, "group2 = %.66s", indication2 ); if( strlen( indication2 ) > 66 ) fprintf( stderr, "..." ); fprintf( stderr, "\n" ); #endif // for( i=0; i 10000 || len2 > 10000 ) ) ) { fprintf( stderr, "\nlen1=%d, len2=%d, Switching to the memsave mode.\n", len1, len2 ); alg = 'M'; if( commonIP ) FreeIntMtx( commonIP ); commonAlloc1 = 0; commonAlloc2 = 0; } // if( fftlog[m1] && fftlog[m2] ) ffttry = ( nlen[m1] > clus1 && nlen[m2] > clus2 ); if( fftlog[m1] && fftlog[m2] ) ffttry = ( nlen[m1] > clus1 && nlen[m2] > clus2 && clus1 < 1000 && clus2 < 1000 ); else ffttry = 0; // ffttry = ( nlen[m1] > clus1 && nlen[m2] > clus2 && clus1 < 5000 && clus2 < 5000 ); // v6.708 // fprintf( stderr, "f=%d, len1/fftlog[m1]=%f, clus1=%d, len2/fftlog[m2]=%f, clus2=%d\n", ffttry, (float)len1/fftlog[m1], clus1, (float)len2/fftlog[m2], clus2 ); // fprintf( stderr, "f=%d, clus1=%d, fftlog[m1]=%d, clus2=%d, fftlog[m2]=%d\n", ffttry, clus1, fftlog[m1], clus2, fftlog[m2] ); if( constraint == 2 ) { if( alg == 'M' ) { fprintf( stderr, "\n\nMemory saving mode is not supported.\n\n" ); exit( 1 ); } fprintf( stderr, "c" ); if( alg == 'A' ) { imp_match_init_strict( NULL, clus1, clus2, strlen( mseq1[0] ), strlen( mseq2[0] ), mseq1, mseq2, effarr1, effarr2, effarr1_kozo, effarr2_kozo, localhomshrink, 1 ); if( rnakozo ) imp_rna( clus1, clus2, mseq1, mseq2, effarr1, effarr2, grouprna1, grouprna2, NULL, NULL, NULL ); pscore = A__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, localhomshrink, &dumfl, NULL, NULL, NULL, NULL ); } else if( alg == 'H' ) { imp_match_init_strictH( NULL, clus1, clus2, strlen( mseq1[0] ), strlen( mseq2[0] ), mseq1, mseq2, effarr1, effarr2, localhomshrink, 1 ); pscore = H__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, localhomshrink, &dumfl, NULL, NULL, NULL, NULL ); } else if( alg == 'Q' ) { imp_match_init_strictQ( NULL, clus1, clus2, strlen( mseq1[0] ), strlen( mseq2[0] ), mseq1, mseq2, effarr1, effarr2, localhomshrink, 1 ); if( rnakozo ) imp_rnaQ( clus1, clus2, mseq1, mseq2, effarr1, effarr2, grouprna1, grouprna2, NULL, NULL, NULL ); pscore = Q__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, localhomshrink, &dumfl, NULL, NULL, NULL, NULL ); } else if( alg == 'R' ) { imp_match_init_strictR( NULL, clus1, clus2, strlen( mseq1[0] ), strlen( mseq2[0] ), mseq1, mseq2, effarr1, effarr2, localhomshrink, 1 ); pscore = R__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, localhomshrink, &dumfl, NULL, NULL, NULL, NULL ); } } else if( force_fft || ( use_fft && ffttry ) ) { fprintf( stderr, "f" ); if( alg == 'M' ) { fprintf( stderr, "m" ); pscore = Falign_udpari_long( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, fftlog+m1 ); } else pscore = Falign( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, fftlog+m1 ); } else { fprintf( stderr, "d" ); fftlog[m1] = 0; switch( alg ) { case( 'a' ): pscore = Aalign( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen ); break; case( 'M' ): fprintf( stderr, "m" ); pscore = MSalignmm( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, NULL, NULL, NULL ); break; case( 'A' ): pscore = A__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL ); break; case( 'Q' ): pscore = Q__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL ); break; case( 'R' ): pscore = R__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL ); break; case( 'H' ): pscore = H__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL ); break; case ( 'C' ): if( outgap && ( ( clus1 == 1 && clus2 != 1 ) || ( clus1 != 1 && clus2 == 1 ) ) ) { pscore = translate_and_Calign( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen ); } else { pscore = A__align( mseq1, mseq2, effarr1, effarr2, clus1, clus2, *alloclen, NULL, &dumfl, NULL, NULL, NULL, NULL ); } break; default: ErrorExit( "ERROR IN SOURCE FILE" ); } } nlen[m1] = 0.5 * ( nlen[m1] + nlen[m2] ); #if SCOREOUT fprintf( stderr, "score = %10.2f\n", pscore ); #endif tscore += pscore; /* 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" ); */ // writePre( njob, name, nlen, aseq, 0 ); if( disp ) display( aseq, njob ); } #if SCOREOUT fprintf( stderr, "totalscore = %10.2f\n\n", tscore ); #endif } static void WriteOptions( FILE *fp ) { if( dorp == 'd' ) fprintf( fp, "DNA\n" ); else { if ( scoremtx == 0 ) fprintf( fp, "JTT %dPAM\n", pamN ); else if( scoremtx == 1 ) fprintf( fp, "BLOSUM %d\n", nblosum ); else if( scoremtx == 2 ) fprintf( fp, "M-Y\n" ); } fprintf( stderr, "Gap Penalty = %+5.2f, %+5.2f, %+5.2f\n", (double)ppenalty/1000, (double)ppenalty_ex/1000, (double)poffset/1000 ); if( use_fft ) fprintf( fp, "FFT on\n" ); fprintf( fp, "tree-base method\n" ); if( tbrweight == 0 ) fprintf( fp, "unweighted\n" ); else if( tbrweight == 3 ) fprintf( fp, "clustalw-like weighting\n" ); if( tbitr || tbweight ) { fprintf( fp, "iterate at each step\n" ); if( tbitr && tbrweight == 0 ) fprintf( fp, " unweighted\n" ); if( tbitr && tbrweight == 3 ) fprintf( fp, " reversely weighted\n" ); if( tbweight ) fprintf( fp, " weighted\n" ); fprintf( fp, "\n" ); } fprintf( fp, "Gap Penalty = %+5.2f, %+5.2f, %+5.2f\n", (double)ppenalty/1000, (double)ppenalty_ex/1000, (double)poffset/1000 ); if( alg == 'a' ) fprintf( fp, "Algorithm A\n" ); else if( alg == 'A' ) fprintf( fp, "Algorithm A+\n" ); else if( alg == 'S' ) fprintf( fp, "Apgorithm S\n" ); else if( alg == 'C' ) fprintf( fp, "Apgorithm A+/C\n" ); else fprintf( fp, "Unknown algorithm\n" ); if( treemethod == 'X' ) fprintf( fp, "Tree = UPGMA (mix).\n" ); else if( treemethod == 'E' ) fprintf( fp, "Tree = UPGMA (average).\n" ); else if( treemethod == 'q' ) fprintf( fp, "Tree = Minimum linkage.\n" ); else fprintf( fp, "Unknown tree.\n" ); if( use_fft ) { fprintf( fp, "FFT on\n" ); if( dorp == 'd' ) fprintf( fp, "Basis : 4 nucleotides\n" ); else { if( fftscore ) fprintf( fp, "Basis : Polarity and Volume\n" ); else fprintf( fp, "Basis : 20 amino acids\n" ); } fprintf( fp, "Threshold of anchors = %d%%\n", fftThreshold ); fprintf( fp, "window size of anchors = %dsites\n", fftWinSize ); } else fprintf( fp, "FFT off\n" ); fflush( fp ); } int main( int argc, char *argv[] ) { static int *nlen; static float *selfscore; int nogaplen; static char **name, **seq; static char **mseq1, **mseq2; static char **bseq; static float **iscore, **iscore_kozo; static double *eff, *eff_kozo, *eff_kozo_mapped = NULL; int i, j, ien, ik, jk; static int ***topol, ***topol_kozo; static float **len, **len_kozo; FILE *prep; FILE *infp; FILE *orderfp; FILE *hat2p; char c; int alloclen; LocalHom **localhomtable; RNApair ***singlerna; float ssi, ssj, bunbo; static char *kozoarivec; int nkozo; 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 ); nkozo = 0; 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+1 ); mseq1 = AllocateCharMtx( njob, 0 ); mseq2 = AllocateCharMtx( njob, 0 ); name = AllocateCharMtx( njob, B+1 ); nlen = AllocateIntVec( njob ); selfscore = AllocateFloatVec( njob ); topol = AllocateIntCub( njob, 2, 0 ); len = AllocateFloatMtx( njob, 2 ); iscore = AllocateFloatHalfMtx( njob ); eff = AllocateDoubleVec( njob ); kozoarivec = AllocateCharVec( njob ); if( constraint ) { localhomtable = (LocalHom **)calloc( njob, sizeof( LocalHom *) ); for( i=0; inext ) { free( (void *)tmppt1 ); tmppt1 = tmppt2; } free( (void *)tmppt1 ); } free( (void *)(localhomtable[i]+j) ); } free( (void *)localhomtable ); } #endif fprintf( trap_g, "done.\n" ); fclose( trap_g ); writeData_pointer( prep_g, njob, name, nlen, bseq ); #if 0 writeData( stdout, njob, name, nlen, bseq ); writePre( njob, name, nlen, bseq, !contin ); writeData_pointer( prep_g, njob, name, nlen, aseq ); #endif #if IODEBUG fprintf( stderr, "OSHIMAI\n" ); #endif if( constraint ) FreeLocalHomTable( localhomtable, njob ); SHOWVERSION; return( 0 ); }