void Write(IRStmt* stmt, unsigned int indent, bool noBraces = false)
{
if (IRSeq* seq = UtCast<IRSeq*>(stmt))
WriteSeq(seq, indent, noBraces);
else
return Dispatch<void>(stmt, indent);
}
void
Finish(struct InstrSeq *Code)
{ struct InstrSeq *code;
struct SymEntry *entry;
struct Attr * attr;
struct StrLabels * strlabel;
code = GenInstr(NULL,".text",NULL,NULL,NULL);
AppendSeq(code,GenInstr(NULL,".globl","main",NULL,NULL));
AppendSeq(code, GenInstr("main",NULL,NULL,NULL,NULL));
AppendSeq(code, doFuncStmt("main"));
AppendSeq(code, GenInstr(NULL, "li", "$v0", "10", NULL));
AppendSeq(code, GenInstr(NULL,"syscall",NULL,NULL,NULL));
AppendSeq(code,Code);
AppendSeq(code,GenInstr(NULL,".data",NULL,NULL,NULL));
AppendSeq(code,GenInstr(NULL,".align","4",NULL,NULL));
AppendSeq(code,GenInstr("_nl",".asciiz","\"\\n\"",NULL,NULL));
AppendSeq(code,GenInstr("_tru",".asciiz","\"true\"",NULL,NULL));
AppendSeq(code,GenInstr("_fal",".asciiz","\"false\"",NULL,NULL));
AppendSeq(code,GenInstr("_sp", ".asciiz","\" \"",NULL,NULL));
entry = FirstEntry(table);
while (entry) {
struct Vtype * vtype = ((struct Vtype *)entry->Attributes);
AppendSeq(code, GenInstr(NULL, ".align", "4", NULL, NULL));
if(vtype->Type == TYPE_INTARR || vtype->Type == TYPE_BOOLARR){
char * buffer;
buffer = Imm(4 * vtype->Size);
AppendSeq(code, GenInstr( (char*)GetName(entry), ".space", buffer, NULL, NULL));
}
else {
AppendSeq(code,GenInstr((char *) GetName(entry),".word","0",NULL,NULL));
}
entry = NextEntry(table, entry);
}
strlabel = slabels;
while( strlabel ){
AppendSeq(code, GenInstr(NULL, ".align", "4", NULL, NULL));
AppendSeq(code, GenInstr(strlabel->SLabel, ".asciiz", strlabel->Str, NULL, NULL));
strlabel = strlabel->next;
}
WriteSeq(code);
return;
}
main (int argc, char ** argv )
{
char *seqfile; /* name of sequence file */
SQINFO sqinfo; /* extra info about sequence */
SQFILE *dbfp; /* open sequence file */
int fmt,ofmt=106; /* format of seqfile */
/* 106 is PHYLIP format in SQUID */
char *seq; /* sequence */
int type; /* kAmino, kDNA, kRNA, or kOtherSeq */
sequence * seqs, * cds_seqs;
sequence tmp_seqs[2], tmp_cds_seqs[2];
char *optname;
char *optarg, *t;
int optind;
int be_quiet;
int seqct = 0,cdsct = 0;
int min_aln_len = 0;
int do_oneline = 0;
char * output_filename = 0, *submat_file = 0;
int showaln = 1;
int showheader=1;
FILE *ofd, *fd;
alignment *cds_aln;
alignment * opt_alignment = NULL; /* place for pairwise alignment */
int len,i,j, k, jk,ik,aln_count, rc;
pairwise_distances pwMLdist, pwNGdist;
int firsttime = 1;
struct timeval tp;
pwMLdist.N = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.dN = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.S = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.dS = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.dNdS = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.SEdS = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.SEdN = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.t = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwMLdist.kappa= make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwNGdist.dN = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwNGdist.dS = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
pwNGdist.dNdS = make_double_matrix(NUM_PW_SEQS,NUM_PW_SEQS);
/*
pwMLdist.N = pwMLdist.dN = pwMLdist.S = 0;
pwMLdist.dS = pwMLdist.dNdS = pwMLdist.SEdS = 0;
pwMLdist.SEdN = pwMLdist.t = pwMLdist.kappa= 0;
pwNGdist.dN = pwNGdist.dS = pwNGdist.dNdS = 0;
*/
Alntype = default_aln_type;
/* Command line Parse */
fmt = SQFILE_UNKNOWN; /* default: autodetect format */
be_quiet = FALSE;
type = kOtherSeq;
/* for our purposes this is only pairwise alignments, but
* would rather do it correctly in case we move to MSA case
*/
while (Getopt(argc, argv, OPTIONS, NOPTIONS, usage,
&optind, &optname, &optarg))
{
if (strcmp(optname, "--matrix") == 0) submat_file = optarg;
else if (strcmp(optname, "--quiet") == 0) be_quiet = TRUE;
else if (strcmp(optname, "--gapopen") == 0) {
Gapopen = atoi(optarg);
if( Gapopen < 0 ) Gapopen *= -1;
} else if (strcmp(optname, "--gapext") == 0) {
Gapext = atoi(optarg);
if( Gapext < 0 ) Gapext *= -1;
} else if (strcmp(optname, "--informat") == 0) {
fmt = String2SeqfileFormat(optarg);
if (fmt == SQFILE_UNKNOWN)
Die("unrecognized sequence file format \"%s\"", optarg);
} else if (strcmp(optname, "--outformat") == 0) {
ofmt = String2SeqfileFormat(optarg);
if (ofmt == SQFILE_UNKNOWN)
Die("unrecognized sequence file format \"%s\"", optarg);
} else if( strcmp(optname, "--global") == 0 ) {
Alntype = global;
} else if (strcmp(optname, "-h") == 0) {
puts(usage);
puts(experts);
exit(EXIT_SUCCESS);
} else if ( strcmp(optname, "-v") == 0 ) {
Verbose = 1;
} else if ( strcmp(optname, "--gapchar") == 0 ) {
GapChar = optarg[0];
} else if( strcmp(optname, "--output") == 0 ) {
output_filename = optarg;
} else if( strcmp(optname, "--showtable" ) == 0 ) {
showaln = 0;
} else if( strcmp(optname, "--noheader" ) == 0 ) {
showheader = 0;
}
}
if (argc - optind < 1) Die("%s\n", usage);
if( ! submat_file ) {
if( (t = getenv("SUBOPTDIR")) != 0 ||
(t = getenv("SUBOPT_DIR")) != 0 ) {
submat_file = calloc(strlen(t) + 24, sizeof(char));
sprintf(submat_file, "%s/%s",t,Default_submat);
} else {
submat_file = calloc(strlen((void *)Default_submat) + 24, sizeof(char));
sprintf(submat_file, "../%s",Default_submat);
}
}
/* open matrix */
fd = fopen(submat_file, "r");
if( ! ParsePAMFile(fd,&ScoringMatrix, &MatrixScale) ) {
fprintf(stderr, "Cannot parse or open matrix file %s\n",submat_file);
free(submat_file);
exit(EXIT_SUCCESS);
}
if( output_filename && strlen(output_filename) != 1 &&
output_filename[0] != '-') {
ofd = fopen(output_filename,"w");
if( ! ofd ) {
fprintf(stderr, "could not open file %s",output_filename);
goto end;
}
} else
ofd = stdout;
while( optind < argc ) {
seqfile = argv[optind++];
/* Try to work around inability to autodetect from a pipe or .gz:
* assume FASTA format
*/
if (fmt == SQFILE_UNKNOWN &&
(Strparse("^.*\\.gz$", seqfile, 0) || strcmp(seqfile, "-") == 0))
fmt = SQFILE_FASTA;
if ((dbfp = SeqfileOpen(seqfile, fmt, NULL)) == NULL)
Die("Failed to open sequence file %s for reading", seqfile);
while (ReadSeq(dbfp, dbfp->format, &seq, &sqinfo))
{
FreeSequence(NULL, &sqinfo);
seqct++;
}
cds_seqs = (sequence *)calloc(seqct, sizeof(sequence));
seqs = (sequence *)calloc(seqct, sizeof(sequence));
SeqfileRewind(dbfp);
seqct=0;
while (ReadSeq(dbfp, dbfp->format, &seq, &sqinfo))
{
sqinfo.type = Seqtype(seq);
if( sqinfo.type == kDNA || sqinfo.type == kRNA ) {
seqs[seqct].seqstr = Translate(seq,stdcode1);
/* Let's remove the last codon if it is a stop codon */
len = strlen(seqs[seqct].seqstr);
if( Verbose )
fprintf(stderr,"seqct is %d length is %d\n",seqct,
len);
if( seqs[seqct].seqstr[len-1] == '*' ) {
seqs[seqct].seqstr[len-1] = '\0';
seq[strlen(seq) - 3] = '\0';
}
cds_seqs[cdsct].seqstr = seq;
seqs[seqct].seqname = calloc(strlen(sqinfo.name)+1,sizeof(char));
cds_seqs[cdsct].seqname = calloc(strlen(sqinfo.name)+1,sizeof(char));
strcpy(seqs[seqct].seqname,sqinfo.name );
strcpy(cds_seqs[cdsct].seqname,sqinfo.name);
cds_seqs[cdsct].length = sqinfo.len;
cds_seqs[cdsct].alphabet = ( sqinfo.type == kDNA ) ? dna : rna;
seqs[seqct].length = strlen(seqs[seqct].seqstr);
seqs[seqct].alphabet = protein;
cdsct++; seqct++;
} else {
fprintf(stderr,"Expect CDS sequences (DNA or RNA) not Protein\n");
goto end;
}
FreeSequence(NULL, &sqinfo);
if( Verbose && seqct > 3 )
break;
}
if( seqct < 2 ) {
fprintf(stderr,"Must have provided a valid file with at least 2 sequences in it");
goto end;
}
for( i=0; i < seqct; i++ ) {
for(k=i+1; k < seqct; k++ ) {
if( (opt_alignment = (alignment *)calloc(1,sizeof(alignment *))) == NULL) {
fprintf(stderr,"Could not allocate memory\n");
goto end;
}
opt_alignment->msa = NULL;
rc = optimal_align(&seqs[i],&seqs[k],opt_alignment);
if( rc != 1 ) {
fprintf(stderr,"Could not make an optimal alignment\n");
goto end;
} else {
tmp_cds_seqs[0] = cds_seqs[i];
tmp_cds_seqs[1] = cds_seqs[k];
rc = mrtrans(opt_alignment, tmp_cds_seqs, &cds_aln,0);
if( rc != 0 ) {
fprintf(stderr, "Could not map the coding sequence to the protein alignemnt for aln %d: %d\n",i,rc);
goto end;
}
if( showaln ) {
if( ofmt >= 100 ) {
MSAFileWrite(ofd,cds_aln->msa, ofmt,do_oneline);
} else {
for(j=0; j < cds_aln->msa->nseq; j++ ) {
WriteSeq(ofd, ofmt,
cds_aln->msa->aseq[j],
&(cds_aln->sqinfo[j]) );
}
}
} else {
if( showheader && firsttime ) {
fprintf(ofd,"SEQ1\tSEQ2\tSCORE\tdN\tdS\tOMEGA\tN\tS\tkappa\tt\tLENGTH\n");
firsttime = 0;
}
if( do_kaks_yn00(cds_aln->msa, &pwMLdist,&pwNGdist) < 0 ) {
fprintf(stderr, "warning: problem with align for %s %s\n",
cds_aln->msa->sqname[0], cds_aln->msa->sqname[1]);
continue;
}
for(ik = 0; ik < NUM_PW_SEQS; ik++ ) {
for( jk = ik+1; jk < NUM_PW_SEQS; jk++ ) {
fprintf(ofd,"%s\t%s\t%d\t%f\t%f\t%f\t%f\t%f\t%f\t%f\t%d\n",
cds_aln->sqinfo[ik].name,
cds_aln->sqinfo[jk].name,
opt_alignment->score,
pwMLdist.dN[ik][jk],pwMLdist.dS[ik][jk],
pwMLdist.dNdS[ik][jk],
pwMLdist.N[ik][jk],
pwMLdist.S[ik][jk],
pwMLdist.kappa[ik][jk],
pwMLdist.t[ik][jk],
opt_alignment->msa->alen);
}
}
}
}
cleanup_alignment(cds_aln);
cleanup_alignment(opt_alignment);
}
}
}
if( ofd && ofd != stdout )
fclose(ofd);
end:
free(submat_file);
Free2DArray((void **)ScoringMatrix,27);
for(i =0; i< seqct; i++ ) {
free(seqs[i].seqstr);
free(seqs[i].seqname);
seqs[i].seqstr = seqs[i].seqname = 0;
}
for(i = 0; i < cdsct; i++) {
free(cds_seqs[i].seqstr);
free(cds_seqs[i].seqname);
cds_seqs[i].seqstr = cds_seqs[i].seqname = 0;
}
cleanup_matrix((void **)pwMLdist.N,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.dN,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.S,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.dS,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.SEdS,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.SEdN,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.t,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.dNdS,NUM_PW_SEQS);
cleanup_matrix((void **)pwMLdist.kappa,NUM_PW_SEQS);
cleanup_matrix((void **)pwNGdist.dN,NUM_PW_SEQS);
cleanup_matrix((void **)pwNGdist.dS,NUM_PW_SEQS);
cleanup_matrix((void **)pwNGdist.dNdS,NUM_PW_SEQS);
free(pwNGdist.dNdS);
free(pwNGdist.dN);
free(pwNGdist.dS);
free(pwMLdist.dNdS);
free(pwMLdist.dN);
free(pwMLdist.dS);
free(pwMLdist.N);
free(pwMLdist.S);
free(pwMLdist.SEdS);
free(pwMLdist.SEdN);
free(pwMLdist.t);
free(pwMLdist.kappa);
return 0;
}
int main(int argc, char **argv)
{
const char *hmmfile; /* file to read HMMs from */
FILE *fp; /* output file handle */
HMMFILE *hmmfp; /* opened hmmfile for reading */
struct plan7_s *hmm; /* HMM to generate from */
int L; /* length of a sequence */
int i; /* counter over sequences */
char *ofile; /* output sequence file */
int nseq; /* number of seqs to sample */
int seed; /* random number generator seed */
int be_quiet; /* TRUE to silence header/footer */
int do_alignment; /* TRUE to output in aligned format */
int do_consensus; /* TRUE to do a single consensus seq */
AjBool ajselex;
AjBool ajcons;
AjPFile inf=NULL;
AjPFile outf=NULL;
AjPStr instr=NULL;
AjPStr outstr=NULL;
#ifdef MEMDEBUG
unsigned long histid1, histid2, orig_size, current_size;
orig_size = malloc_inuse(&histid1);
fprintf(stderr, "[... memory debugging is ON ...]\n");
#endif
/***********************************************
* Parse command line
***********************************************/
nseq = 10;
be_quiet = FALSE;
do_alignment = FALSE;
do_consensus = FALSE;
ofile = NULL;
embInitPV("ohmmemit",argc,argv,"HMMER",VERSION);
ajselex = ajAcdGetBoolean("selex");
ajcons = ajAcdGetBoolean("consensus");
nseq = ajAcdGetInt("number");
seed = ajAcdGetInt("seed");
inf = ajAcdGetInfile("infile");
outf = ajAcdGetOutfile("outfile");
if(!seed)
seed = time ((time_t *) NULL);
if(ajselex)
do_alignment=TRUE;
else
do_alignment=FALSE;
if(ajcons)
do_consensus=TRUE;
else
do_consensus=FALSE;
instr = ajStrNewC((char *)ajFileGetNameC(inf));
outstr = ajStrNewC((char *)ajFileGetNameC(outf));
hmmfile = ajStrGetPtr(instr);
sre_srandom(seed);
if (do_alignment && do_consensus)
ajFatal("Sorry, -selex and -consensus are incompatible.\n");
if (nseq != 10 && do_consensus)
ajWarn("-consensus overrides -number (# of sampled seqs)");
/***********************************************
* Open HMM file (might be in HMMERDB or current directory).
* Read a single HMM from it.
***********************************************/
if ((hmmfp = HMMFileOpen(hmmfile, "HMMERDB")) == NULL)
ajFatal("Failed to open HMM file %s\n", hmmfile);
if (!HMMFileRead(hmmfp, &hmm))
ajFatal("Failed to read any HMMs from %s\n", hmmfile);
HMMFileClose(hmmfp);
if (hmm == NULL)
ajFatal("HMM file %s corrupt or in incorrect format? Parse failed",
hmmfile);
/* Configure the HMM to shut off N,J,C emission: so we
* do a simple single pass through the model.
*/
Plan7NakedConfig(hmm);
Plan7Renormalize(hmm);
/***********************************************
* Open the output file, or stdout
***********************************************/
fp = ajFileGetFileptr(outf);
/***********************************************
* Show the options banner
***********************************************/
be_quiet=TRUE;
if (! be_quiet)
{
printf("HMM file: %s\n", hmmfile);
if (! do_consensus)
{
printf("Number of seqs: %d\n", nseq);
printf("Random seed: %d\n", seed);
}
printf("- - - - - - - - - - - - - - - - - - - - - - - - - "
"- - - - - - -\n\n");
}
/***********************************************
* Do the work.
* If we're generating an alignment, we have to collect
* all our traces, then output. If we're generating unaligned
* sequences, we can emit one at a time.
***********************************************/
if (do_consensus)
{
char *seq;
SQINFO sqinfo; /* info about sequence (name/desc) */
EmitConsensusSequence(hmm, &seq, NULL, &L, NULL);
strcpy(sqinfo.name, "consensus");
sqinfo.len = L;
sqinfo.flags = SQINFO_NAME | SQINFO_LEN;
WriteSeq(fp, kPearson, seq, &sqinfo);
free(seq);
}
else if (do_alignment)
{
struct p7trace_s **tr;
char **dsq;
SQINFO *sqinfo;
char **aseq;
AINFO ainfo;
float *wgt;
dsq = MallocOrDie(sizeof(char *) * nseq);
tr = MallocOrDie(sizeof(struct p7trace_s *) * nseq);
sqinfo = MallocOrDie(sizeof(SQINFO) * nseq);
wgt = MallocOrDie(sizeof(float) * nseq);
FSet(wgt, nseq, 1.0);
for (i = 0; i < nseq; i++)
{
EmitSequence(hmm, &(dsq[i]), &L, &(tr[i]));
sprintf(sqinfo[i].name, "seq%d", i+1);
sqinfo[i].len = L;
sqinfo[i].flags = SQINFO_NAME | SQINFO_LEN;
}
P7Traces2Alignment(dsq, sqinfo, wgt, nseq, hmm->M, tr, FALSE,
&aseq, &ainfo);
/* Output the alignment */
WriteSELEX(fp, aseq, &ainfo, 50);
if (ofile != NULL && !be_quiet)
printf("Alignment saved in file %s\n", ofile);
/* Free memory
*/
for (i = 0; i < nseq; i++)
{
P7FreeTrace(tr[i]);
free(dsq[i]);
}
FreeAlignment(aseq, &ainfo);
free(sqinfo);
free(dsq);
free(wgt);
free(tr);
}
else /* unaligned sequence output */
{
struct p7trace_s *tr;
char *dsq;
char *seq;
SQINFO sqinfo;
for (i = 0; i < nseq; i++)
{
EmitSequence(hmm, &dsq, &L, &tr);
sprintf(sqinfo.name, "seq%d", i+1);
sqinfo.len = L;
sqinfo.flags = SQINFO_NAME | SQINFO_LEN;
seq = DedigitizeSequence(dsq, L);
WriteSeq(fp, kPearson, seq, &sqinfo);
P7FreeTrace(tr);
free(dsq);
free(seq);
}
}
ajFileClose(&outf);
FreePlan7(hmm);
SqdClean();
#ifdef MEMDEBUG
current_size = malloc_inuse(&histid2);
if (current_size != orig_size)
malloc_list(2, histid1, histid2);
else
fprintf(stderr, "[No memory leaks.]\n");
#endif
ajStrDel(&instr);
ajStrDel(&outstr);
ajFileClose(&inf);
ajFileClose(&outf);
embExit();
return 0;
}
void Visit(IRSeq* seq, unsigned int indent)
{
WriteSeq(seq, indent, false);
}
