static void
onefetch_subseq(ESL_GETOPTS *go, FILE *ofp, ESL_SQFILE *sqfp, char *newname, char *key, uint32_t given_start, uint32_t given_end)
{
int start, end;
int do_revcomp;
ESL_SQ *sq = esl_sq_Create();
if (sqfp->data.ascii.ssi == NULL) esl_fatal("no ssi index");
/* reverse complement indicated by coords. */
/* -c 52: would be 52,0, so watch out for given_end = 0 case */
if (given_end != 0 && given_start > given_end)
{ start = given_end; end = given_start; do_revcomp = TRUE; }
else
{ start = given_start; end = given_end; do_revcomp = FALSE; }
if (esl_sqio_FetchSubseq(sqfp, key, start, end, sq) != eslOK) esl_fatal(esl_sqfile_GetErrorBuf(sqfp));
if (newname != NULL) esl_sq_SetName(sq, newname);
else esl_sq_FormatName(sq, "%s/%d-%d", key, given_start, (given_end == 0) ? sq->L : given_end);
/* Two ways we might have been asked to revcomp: by coord, or by -r option */
/* (If both happen, they'll cancel each other out) */
if (do_revcomp)
if (esl_sq_ReverseComplement(sq) != eslOK) esl_fatal("Failed to reverse complement %s; is it a protein?\n", sq->name);
if (esl_opt_GetBoolean(go, "-r"))
if (esl_sq_ReverseComplement(sq) != eslOK) esl_fatal("Failed to reverse complement %s; is it a protein?\n", sq->name);
esl_sqio_Write(ofp, sq, eslSQFILE_FASTA, FALSE);
esl_sq_Destroy(sq);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 1, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
char *hmmfile = esl_opt_GetArg(go, 1);
int L = esl_opt_GetInteger(go, "-L");
int N = esl_opt_GetInteger(go, "-N");
ESL_ALPHABET *abc = NULL;
P7_HMMFILE *hfp = NULL;
P7_HMM *hmm = NULL;
P7_BG *bg = NULL;
P7_PROFILE *gm = NULL;
P7_TRACE *tr = p7_trace_Create();
ESL_SQ *sq = NULL;
char errbuf[eslERRBUFSIZE];
int i;
int status;
status = p7_hmmfile_OpenE(hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, hmmfile, errbuf);
status = p7_hmmfile_Read(hfp, &abc, &hmm);
if (status == eslEFORMAT) p7_Fail("Bad file format in HMM file %s:\n%s\n", hfp->fname, hfp->errbuf);
else if (status == eslEINCOMPAT) p7_Fail("HMM in %s is not in the expected %s alphabet\n", hfp->fname, esl_abc_DecodeType(abc->type));
else if (status == eslEOF) p7_Fail("Empty HMM file %s? No HMM data found.\n", hfp->fname);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s\n", hfp->fname);
p7_hmmfile_Close(hfp);
bg = p7_bg_Create(abc); p7_bg_SetLength(bg, L);
gm = p7_profile_Create(hmm->M, abc); p7_ProfileConfig(hmm, bg, gm, L, p7_LOCAL);
sq = esl_sq_CreateDigital(abc);
for (i = 0; i < N; i++)
{
p7_ProfileEmit(rng, hmm, gm, bg, sq, tr);
esl_sq_FormatName(sq, "%s-sample%d", hmm->name, i);
esl_sqio_Write(stdout, sq, eslSQFILE_FASTA, FALSE);
if (p7_trace_Validate(tr, abc, sq->dsq, errbuf) != eslOK) esl_fatal(errbuf);
esl_sq_Reuse(sq);
p7_trace_Reuse(tr);
}
esl_sq_Destroy(sq);
p7_trace_Destroy(tr);
p7_profile_Destroy(gm);
p7_bg_Destroy(bg);
p7_hmm_Destroy(hmm);
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
return 0;
}
/* onefetch():
* Given one <key> (a seq name or accession), retrieve the corresponding sequence.
* In SSI mode, we can do this quickly by positioning the file, then regurgitating
* every line until the end-of-record marker; we don't even have to parse.
* Without an SSI index, we have to parse the file sequentially 'til we find
* the one we're after.
*/
static void
onefetch(ESL_GETOPTS *go, FILE *ofp, char *key, ESL_SQFILE *sqfp)
{
ESL_SQ *sq = esl_sq_Create();
int do_revcomp = esl_opt_GetBoolean(go, "-r");
char *newname = esl_opt_GetString(go, "-n");
int status;
/* Try to position the file at the desired sequence with SSI. */
if (sqfp->data.ascii.ssi != NULL)
{
status = esl_sqfile_PositionByKey(sqfp, key);
if (status == eslENOTFOUND) esl_fatal("seq %s not found in SSI index for file %s\n", key, sqfp->filename);
else if (status == eslEFORMAT) esl_fatal("Failed to parse SSI index for %s\n", sqfp->filename);
else if (status != eslOK) esl_fatal("Failed to look up location of seq %s in SSI index of file %s\n", key, sqfp->filename);
status = esl_sqio_Read(sqfp, sq);
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status == eslEOF) esl_fatal("Unexpected EOF reading sequence file %s",
status, sqfp->filename);
else if (status != eslOK) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
if (strcmp(key, sq->name) != 0 && strcmp(key, sq->acc) != 0)
esl_fatal("whoa, internal error; found the wrong sequence %s, not %s", sq->name, key);
}
else
{ /* Else, we have to read the whole damn file sequentially until we find the seq */
while ((status = esl_sqio_Read(sqfp, sq)) != eslEOF) {
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
if (strcmp(key, sq->name) == 0 || strcmp(key, sq->acc) == 0) break;
esl_sq_Reuse(sq);
}
if (status == eslEOF) esl_fatal("Failed to find sequence %s in file %s\n", key, sqfp->filename);
}
if (do_revcomp == FALSE && newname == NULL && ! esl_sqio_IsAlignment(sqfp->format))
{ /* If we're not manipulating the sequence in any way, and it's not from an alignment file, we can Echo() it. */
if (esl_sqio_Echo(sqfp, sq, ofp) != eslOK) esl_fatal("Echo failed: %s\n", esl_sqfile_GetErrorBuf(sqfp));
}
else
{ /* Otherwise we Write() the parsed version. */
if (do_revcomp && esl_sq_ReverseComplement(sq) != eslOK) esl_fatal("Failed to reverse complement %s; is it a protein?\n", sq->name);
if (newname != NULL) esl_sq_SetName(sq, newname);
esl_sqio_Write(ofp, sq, eslSQFILE_FASTA, FALSE);
}
esl_sq_Destroy(sq);
}
static void
emit_sequences(ESL_GETOPTS *go, FILE *ofp, int outfmt, ESL_RANDOMNESS *r, P7_HMM *hmm)
{
ESL_SQ *sq = NULL;
P7_TRACE *tr = NULL;
P7_BG *bg = NULL;
P7_PROFILE *gm = NULL;
int do_profile = esl_opt_GetBoolean(go, "-p");
int N = esl_opt_GetInteger(go, "-N");
int L = esl_opt_GetInteger(go, "-L");
int mode = p7_LOCAL;
int nseq;
int status;
if (esl_opt_GetBoolean(go, "--local")) mode = p7_LOCAL;
else if (esl_opt_GetBoolean(go, "--unilocal")) mode = p7_UNILOCAL;
else if (esl_opt_GetBoolean(go, "--glocal")) mode = p7_GLOCAL;
else if (esl_opt_GetBoolean(go, "--uniglocal")) mode = p7_UNIGLOCAL;
if ((sq = esl_sq_CreateDigital(hmm->abc)) == NULL) esl_fatal("failed to allocate sequence");
if ((tr = p7_trace_Create()) == NULL) esl_fatal("failed to allocate trace");
if ((bg = p7_bg_Create(hmm->abc)) == NULL) esl_fatal("failed to create null model");
if ((gm = p7_profile_Create(hmm->M, hmm->abc)) == NULL) esl_fatal("failed to create profile");
if (p7_ProfileConfig(hmm, bg, gm, L, mode) != eslOK) esl_fatal("failed to configure profile");
if (p7_bg_SetLength(bg, L) != eslOK) esl_fatal("failed to reconfig null model length");
if (p7_hmm_Validate (hmm, NULL, 0.0001) != eslOK) esl_fatal("whoops, HMM is bad!");
if (p7_profile_Validate(gm, NULL, 0.0001) != eslOK) esl_fatal("whoops, profile is bad!");
for (nseq = 1; nseq <= N; nseq++)
{
if (do_profile) status = p7_ProfileEmit(r, hmm, gm, bg, sq, tr);
else status = p7_CoreEmit (r, hmm, sq, tr);
if (status) esl_fatal("Failed to emit sequence\n");
status = esl_sq_FormatName(sq, "%s-sample%d", hmm->name, nseq);
if (status) esl_fatal("Failed to set sequence name\n");
status = esl_sqio_Write(ofp, sq, outfmt, FALSE);
if (status != eslOK) esl_fatal("Failed to write sequence\n");
p7_trace_Reuse(tr);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
p7_trace_Destroy(tr);
p7_bg_Destroy(bg);
p7_profile_Destroy(gm);
return;
}
/* seq_generation()
*
* Generating sequences.
*/
static int
seq_generation(ESL_GETOPTS *go, ESL_RANDOMNESS *r, FILE *ofp, int outfmt)
{
ESL_ALPHABET *abc = NULL;
ESL_SQ *sq = NULL;
double *fq = NULL;
int alphatype = eslUNKNOWN; // static checkers can't see that 1 of --rna, --dna, --amino must be true
int N = esl_opt_GetInteger(go, "-N");
int L = esl_opt_GetInteger(go, "-L");
int i;
int status;
if (L <= 0) esl_fatal("To generate sequences, set -L option (length of generated seqs) > 0 ");
if (esl_opt_GetBoolean(go, "--rna")) alphatype = eslRNA;
if (esl_opt_GetBoolean(go, "--dna")) alphatype = eslDNA;
if (esl_opt_GetBoolean(go, "--amino")) alphatype = eslAMINO;
abc = esl_alphabet_Create(alphatype);
sq = esl_sq_CreateDigital(abc);
esl_sq_GrowTo(sq, L);
/* Pick the iid frequency distribution to use */
ESL_ALLOC(fq, sizeof(double) * abc->K);
switch (alphatype) {
case eslRNA:
case eslDNA: esl_vec_DSet(fq, 4, 0.25); break;
case eslAMINO: esl_composition_SW34(fq); break;
default: esl_vec_DSet(fq, abc->K, 1.0 / (double) abc->K); break;
}
/* generate */
for (i = 0; i < N; i++)
{
esl_rsq_xIID(r, fq, abc->K, L, sq->dsq);
if (N > 1) esl_sq_FormatName(sq, "random%d", i);
else esl_sq_SetName(sq, "random");
sq->n = L;
esl_sqio_Write(ofp, sq, outfmt, FALSE);
}
free(fq);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
return eslOK;
ERROR:
if (fq != NULL) free(fq);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
return status;
}
static void
emit_consensus(ESL_GETOPTS *go, FILE *ofp, int outfmt, P7_HMM *hmm)
{
ESL_SQ *sq = NULL;
if ((sq = esl_sq_CreateDigital(hmm->abc)) == NULL) esl_fatal("failed to allocate sequence");
if (p7_emit_SimpleConsensus(hmm, sq) != eslOK) esl_fatal("failed to create simple consensus seq");
if (esl_sq_FormatName(sq, "%s-consensus", hmm->name) != eslOK) esl_fatal("failed to set sequence name");
if (esl_sqio_Write(ofp, sq, outfmt, FALSE) != eslOK) esl_fatal("failed to write sequence");
esl_sq_Destroy(sq);
return;
}
static void
emit_fancycons(ESL_GETOPTS *go, FILE *ofp, int outfmt, P7_HMM *hmm)
{
ESL_SQ *sq = NULL;
float minl = esl_opt_GetReal(go, "--minl");
float minu = esl_opt_GetReal(go, "--minu");
if ((sq = esl_sq_Create()) == NULL) esl_fatal("failed to allocate sequence");
if (p7_emit_FancyConsensus(hmm, minl, minu, sq) != eslOK) esl_fatal("failed to create consensus seq");
if (esl_sq_FormatName(sq, "%s-consensus", hmm->name) != eslOK) esl_fatal("failed to set sequence name");
if (esl_sqio_Write(ofp, sq, outfmt, FALSE) != eslOK) esl_fatal("failed to write sequence");
esl_sq_Destroy(sq);
return;
}
static int
synthesize_negatives(ESL_GETOPTS *go, struct cfg_s *cfg, int nneg)
{
ESL_SQ *sq = esl_sq_CreateDigital(cfg->abc);
int a;
int i;
int L1,L2,L3,d1n,d2n;
for (i = 0; i < nneg; i++)
{
/* Select a random test seq, to use its same segments */
a = esl_rnd_Roll(cfg->r, cfg->ntest);
L1 = cfg->test_lens[a].L1;
L2 = cfg->test_lens[a].L2;
L3 = cfg->test_lens[a].L3;
d1n = cfg->test_lens[a].d1n;
d2n = cfg->test_lens[a].d2n;
esl_sq_GrowTo(sq, cfg->test_lens[a].L);
esl_sq_FormatName(sq, "decoy%d", i+1);
esl_sq_FormatDesc(sq, "L=%d in segments: %d/%d/%d/%d/%d", cfg->test_lens[a].L, L1, d1n, L2, d2n, L3);
sq->n = cfg->test_lens[a].L;
fprintf(cfg->negsummfp, "%-15s %5d %5d %5d %5d %5d %5d",
sq->name, (int) sq->n,
L1, d1n, L2, d2n, L3);
sq->dsq[0] = sq->dsq[cfg->test_lens[a].L+1] = eslDSQ_SENTINEL;
set_random_segment(go, cfg, cfg->negsummfp, sq->dsq+1, L1);
set_random_segment(go, cfg, cfg->negsummfp, sq->dsq+1+L1, d1n);
set_random_segment(go, cfg, cfg->negsummfp, sq->dsq+1+L1+d1n, L2);
set_random_segment(go, cfg, cfg->negsummfp, sq->dsq+1+L1+d1n+L2, d2n);
set_random_segment(go, cfg, cfg->negsummfp, sq->dsq+1+L1+d1n+L2+d2n, L3);
fprintf(cfg->negsummfp, "\n");
esl_sqio_Write(cfg->out_seqfp, sq, eslSQFILE_FASTA, FALSE);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
return eslOK;
}
int main(int argc, char **argv)
{
ESL_SQFILE *sqfp = NULL;
ESL_SQ *sq = NULL;
ESL_SQ *dsq = NULL;
ESL_SQ **prot;
int c;
ESL_ALPHABET *abc, *prot_abc;
ESL_SQ *prot6[6];
int x;
abc = esl_alphabet_Create(eslDNA);
prot_abc = esl_alphabet_Create(eslAMINO);
if(argc != 2)
{
printf("You need to pass an argument for a filepath to a dna/rna fasta file\n");
exit(0);
}
if(eslOK != esl_sqfile_Open(argv[1], eslSQFILE_FASTA, NULL, &sqfp))
{
printf("Invalid filepath: %s\n", argv[1]);
exit(0);
}
sq = esl_sq_Create();
if(sq == NULL)
{
printf("could not allocate new sequence\n");
exit(0);
}
if(esl_sqio_Read(sqfp, sq) != eslOK)
{
printf("Not a valid fasta file %s\n", argv[1]);
exit(0);
}
dsq = esl_sq_Create();
if(dsq == NULL)
{
printf("could not allocate digital sequence\n");
exit(0);
}
if(esl_sq_Copy(sq, dsq) != eslOK)
{
printf("could not copy sequence\n");
exit(0);
}
if(esl_sq_Digitize(abc, dsq) != eslOK)
{
printf("could not digitize sequence\n");
exit(0);
}
esl_sqio_Write(stdout, sq, eslSQFILE_FASTA, 0);
if(esl_trans_6frame(sq, prot6) != eslOK)
{
printf("could not generate six frame translation\n");
exit(0);
}
for(x = 0; x < 6; x++)
{
esl_sqio_Write(stdout, prot6[x], eslSQFILE_FASTA, 0);
}
if(esl_trans_orf(dsq, &prot, &c, 10) != eslOK)
{
printf("could not translate open reading frames\n");
exit(0);
}
for(x = 0; x < c; x++)
{
esl_sqio_Write(stdout, prot[x], eslSQFILE_FASTA, 0);
}
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
char *maskfile = NULL; /* mask coordinate file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
int outfmt = eslSQFILE_FASTA; /* format code for output seqs */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
ESL_FILEPARSER *maskefp = NULL; /* open mask coord file */
FILE *ofp = NULL; /* output stream for masked seqs */
char *source = NULL; /* name of current seq to mask */
char *p1, *p2; /* pointers used in parsing */
int64_t start, end; /* start, end coord for masking */
int64_t i, j, pos; /* coords in a sequence */
int64_t overmask; /* # of extra residues to mask */
ESL_SQ *sq = esl_sq_Create(); /* current sequence */
int do_fetching;
int do_lowercase;
int maskchar;
int status; /* easel return code */
/****************************************************************************
* Parse command line
****************************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
do_fetching = esl_opt_GetBoolean(go, "-R");
do_lowercase = esl_opt_GetBoolean(go, "-l");
overmask = (esl_opt_IsOn(go, "-x") ? esl_opt_GetInteger(go, "-x") : 0);
maskchar = (esl_opt_IsOn(go, "-m") ? esl_opt_GetChar(go, "-m") : 'X');
seqfile = esl_opt_GetArg(go, 1);
maskfile = esl_opt_GetArg(go, 2);
/* Open the <seqfile>: text mode, not digital */
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
if(do_fetching)
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
/* Open the <maskfile> */
if (esl_fileparser_Open(maskfile, NULL, &maskefp) != eslOK)
cmdline_failure(argv[0], "Failed to open mask coordinate file %s\n", maskfile);
esl_fileparser_SetCommentChar(maskefp, '#');
/* Open the output file, if any */
if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/****************************************************************************
* Main loop over lines in <maskfile>
****************************************************************************/
/* Read one data line at a time from the <maskfile>;
* parse into data fields <seqname> <start> <end>
*/
while (esl_fileparser_NextLine(maskefp) == eslOK)
{
/* First field is sequence name */
if (esl_fileparser_GetTokenOnLine(maskefp, &source, NULL) != eslOK)
esl_fatal("Failed to read source seq name on line %d of file %s\n", maskefp->linenumber, maskfile);
/* Get the sequence */
if (do_fetching)
{ /* If the <seqfile> is SSI indexed, try to reposition it and read <source> seq by random access */
status = esl_sqio_Fetch(sqfp, source, sq);
if (status == eslENOTFOUND) esl_fatal("seq %s not found in SSI index for file %s\n", source, sqfp->filename);
else if (status == eslEINVAL) esl_fatal("No SSI index or can't reposition in file %s\n", sqfp->filename);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected failure in fetching %s from file %s\n", source, sqfp->filename);
}
else
{ /* else, assume we're reading sequentially; <sqfile> and <maskfile> have seqs in same order */
status = esl_sqio_Read(sqfp, sq);
if (status == eslEOF) esl_fatal("File %s ended prematurely; didn't find %s\n", sqfp->filename, source);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected error reading sequence file %s\n", sqfp->filename);
if ((strcmp(sq->name, source) != 0) && (strcmp(sq->acc, source) != 0))
esl_fatal("Sequences in <sqfile> and <maskfile> aren't in same order; try -R");
}
/* If we're masking by lowercase, first make sure everything's uppercase */
if (do_lowercase)
for (pos = 0; pos < sq->n; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = toupper(sq->seq[pos]);
/* Next two fields are <start>, <end> for the masking */
/* possible future extension: wrap loop around this, enable multiple masked regions */
if (esl_fileparser_GetTokenOnLine(maskefp, &p1, NULL) != eslOK)
esl_fatal("Failed to read start coord on line %d of file %s\n", maskefp->linenumber, maskfile);
start = strtoll(p1, &p2, 0) - 1;
if (esl_fileparser_GetTokenOnLine(maskefp, &p2, NULL) != eslOK)
esl_fatal("Failed to read end coord on line %d of file %s\n", maskefp->linenumber, maskfile);
end = strtoll(p2, &p1, 0) - 1;
/* Do the masking */
if (esl_opt_GetBoolean(go, "-r")) /* Reverse masking */
{ /* leave start..end unmasked; mask prefix 0..start-1, end+1..L-1 */
i = 0;
j = ESL_MIN(sq->n-1, start - 1 + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
i = ESL_MAX(0, end + 1 - overmask);
j = sq->n-1;
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
else
{ /* normal: mask start..end */
i = ESL_MAX(0, start - overmask);
j = ESL_MIN(sq->n-1, end + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
esl_sqio_Write(ofp, sq, outfmt, FALSE);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
esl_fileparser_Close(maskefp);
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
if (ofp != stdout) fclose(ofp);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 2, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = NULL;
char *ghmmfile = esl_opt_GetArg(go, 1); /* HMMs parameterized for sequence generation */
char *ahmmfile = esl_opt_GetArg(go, 2); /* HMMs parameterized for alignment */
int N = esl_opt_GetInteger(go, "-N");
P7_HMMFILE *ghfp = NULL;
P7_HMMFILE *ahfp = NULL;
P7_HMM *ghmm = NULL;
P7_HMM *ahmm = NULL;
P7_PROFILE *ggm = NULL;
P7_PROFILE *agm = NULL;
P7_OPROFILE *aom = NULL;
P7_BG *bg = NULL;
ESL_SQ *sq = NULL;
P7_TRACE *reftr = p7_trace_Create();
P7_TRACE *testtr = p7_trace_Create();
P7_TRACE_METRICS *tmetrics = p7_trace_metrics_Create();
P7_REFMX *rmx = p7_refmx_Create(100,100);
// P7_FILTERMX *ox = NULL;
P7_HARDWARE *hw;
if ((hw = p7_hardware_Create ()) == NULL) p7_Fail("Couldn't get HW information data structure");
P7_SPARSEMASK *sm = p7_sparsemask_Create(100, 100, hw->simd);
P7_SPARSEMX *sxv = p7_sparsemx_Create(NULL);
int idx;
char errbuf[eslERRBUFSIZE];
int status;
p7_Init();
/* open HMM file containing models parameterized for generation (sampling) of seqs */
status = p7_hmmfile_OpenE(ghmmfile, NULL, &ghfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", ghmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", ghmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, ghmmfile, errbuf);
/* open HMM file containing models parameterized for alignment (may be the same as ghmmfile) */
status = p7_hmmfile_OpenE(ahmmfile, NULL, &ahfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", ahmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", ahmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, ahmmfile, errbuf);
while ( (status = p7_hmmfile_Read(ghfp, &abc, &ghmm)) == eslOK) /* <abc> gets set on first read */
{
/* read the counterpart HMM from <ahfp> */
status = p7_hmmfile_Read(ahfp, &abc, &ahmm);
if (status == eslEFORMAT) p7_Fail("Bad file format in HMM file %s:\n%s\n", ahfp->fname, ahfp->errbuf);
else if (status == eslEINCOMPAT) p7_Fail("HMM in %s is not in the expected %s alphabet\n", ahfp->fname, esl_abc_DecodeType(abc->type));
else if (status == eslEOF) p7_Fail("Empty HMM file %s? No HMM data found.\n", ahfp->fname);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s\n", ahfp->fname);
/* try to validate that they're the "same" */
if (ahmm->M != ghmm->M || strcmp(ahmm->name, ghmm->name) != 0) p7_Fail("<gen-hmmfile>, <ali-hmmfile> contain different set or order of models");
/* deferred one-time creation of structures that need to know the alphabet */
if (!bg) bg = p7_bg_Create(abc);
if (!sq) sq = esl_sq_CreateDigital(abc);
ggm = p7_profile_Create(ghmm->M, abc);
agm = p7_profile_Create(ahmm->M, abc);
aom = p7_oprofile_Create(ahmm->M, abc, hw->simd);
p7_profile_ConfigCustom(ggm, ghmm, bg, esl_opt_GetInteger(go, "--gL"), esl_opt_GetReal(go, "--gnj"), esl_opt_GetReal(go, "--gpglocal"));
p7_profile_ConfigCustom(agm, ahmm, bg, 100, esl_opt_GetReal(go, "--anj"), esl_opt_GetReal(go, "--apglocal"));
p7_oprofile_Convert(agm, aom);
for (idx = 1; idx <= N; idx++)
{
p7_ProfileEmit(rng, ghmm, ggm, bg, sq, reftr);
if (esl_opt_GetBoolean(go, "--dumpseqs")) {
esl_sq_FormatName(sq, "seq%d", idx);
esl_sqio_Write(stdout, sq, eslSQFILE_FASTA, FALSE);
}
p7_bg_SetLength(bg, sq->n);
p7_profile_SetLength(agm, sq->n);
p7_sparsemask_Reinit(sm, agm->M, sq->n);
p7_sparsemask_AddAll(sm);
if (esl_opt_GetBoolean(go, "--vit")) p7_ReferenceViterbi(sq->dsq, sq->n, agm, rmx, testtr, /*opt_vsc=*/NULL);
else p7_SparseViterbi (sq->dsq, sq->n, agm, sm, sxv, testtr, /*opt_vsc=*/NULL);
p7_trace_metrics(reftr, testtr, tmetrics);
p7_sparsemask_Reuse(sm);
p7_sparsemx_Reuse(sxv);
//p7_filtermx_Reuse(ox);
p7_refmx_Reuse(rmx);
esl_sq_Reuse(sq);
p7_trace_Reuse(reftr);
p7_trace_Reuse(testtr);
}
p7_oprofile_Destroy(aom);
p7_profile_Destroy(ggm);
p7_profile_Destroy(agm);
p7_hmm_Destroy(ghmm);
p7_hmm_Destroy(ahmm);
}
/* we leave the loop with <status> set by a p7_hmmfile_Read() on ghfp; if all is well, status=eslEOF */
if (status == eslEFORMAT) p7_Fail("Bad file format in HMM file %s:\n%s\n", ghfp->fname, ghfp->errbuf);
else if (status == eslEINCOMPAT) p7_Fail("HMM in %s is not in the expected %s alphabet\n", ghfp->fname, esl_abc_DecodeType(abc->type));
else if (status != eslEOF) p7_Fail("Unexpected error in reading HMMs from %s\n", ghfp->fname);
p7_trace_metrics_Dump(stdout, tmetrics);
p7_hmmfile_Close(ghfp);
p7_hmmfile_Close(ahfp);
// p7_filtermx_Destroy(ox);
p7_sparsemask_Destroy(sm);
p7_sparsemx_Destroy(sxv);
p7_refmx_Destroy(rmx);
p7_trace_metrics_Destroy(tmetrics);
p7_trace_Destroy(testtr);
p7_trace_Destroy(reftr);
p7_bg_Destroy(bg);
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
}
/* seq_shuffling()
* SRE, Tue Jan 22 08:35:51 2008 [Market Street Cafe, Leesburg]
*
* Shuffling of input sequences.
*
* Fixed-length (L>0) vs. full-length (L=0) modes handled differently.
* In fixed-length mode:
* <shuff->seq> only needs to be allocated once, for L
* <targ> is an allocated copy of a random subseq of length L
* sequences < L residues long can't be shuffled
* In full-length mode:
* <shuff->seq> is grown to length <sq->n> for each input seq
* <targ> just points to <sq->seq>
*/
static int
seq_shuffling(ESL_GETOPTS *go, ESL_RANDOMNESS *r, FILE *ofp, int outfmt)
{
char *seqfile = esl_opt_GetArg(go, 1);
int infmt = eslSQFILE_UNKNOWN;
ESL_SQFILE *sqfp = NULL;
ESL_SQ *sq = esl_sq_Create();
ESL_SQ *shuff = esl_sq_Create();
char *targ = NULL;
int N = esl_opt_GetInteger(go, "-N");
int L = esl_opt_GetInteger(go, "-L"); /* L>0 means select random fixed-len subseqs */
int kmers = 0;
int i;
int status;
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
if (esl_opt_IsOn(go, "-k")) kmers = esl_opt_GetInteger(go, "-k");
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", seqfile);
else if (status == eslEINVAL) esl_fatal("Can't autodetect stdin or .gz.");
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
if (L>0) {
esl_sq_GrowTo(shuff, L);
shuff->n = L;
ESL_ALLOC(targ, sizeof(char) * (L+1));
}
while ((status = esl_sqio_Read(sqfp, sq)) == eslOK)
{
if (L == 0) { /* shuffling entire sequence */
esl_sq_GrowTo(shuff, sq->n); /* make sure shuff can hold sq */
shuff->n = sq->n;
targ = sq->seq;
} else {
if (sq->n < L) continue; /* reject seqs < L long */
}
for (i = 0; i < N; i++)
{
if (L > 0) { /* fixed-len mode: copy a random subseq */
int pos = esl_rnd_Roll(r, sq->n - L + 1);
strncpy(targ, sq->seq + pos, L);
targ[L] = '\0';
}
/* Do the requested kind of shuffling */
if (esl_opt_GetBoolean(go, "-m")) esl_rsq_CShuffle (r, targ, shuff->seq); /* monoresidue shuffling */
else if (esl_opt_GetBoolean(go, "-d")) esl_rsq_CShuffleDP (r, targ, shuff->seq); /* diresidue shuffling */
else if (esl_opt_IsOn (go, "-k")) esl_rsq_CShuffleKmers(r, targ, kmers, shuff->seq); /* diresidue shuffling */
else if (esl_opt_GetBoolean(go, "-0")) esl_rsq_CMarkov0 (r, targ, shuff->seq); /* 0th order Markov */
else if (esl_opt_GetBoolean(go, "-1")) esl_rsq_CMarkov1 (r, targ, shuff->seq); /* 1st order Markov */
else if (esl_opt_GetBoolean(go, "-r")) esl_rsq_CReverse ( targ, shuff->seq); /* reverse */
else if (esl_opt_IsOn (go, "-w")) { /* regionally shuffle */
int W= esl_opt_GetInteger(go, "-w"); esl_rsq_CShuffleWindows(r, targ, W, shuff->seq);
}
/* Set the name of the shuffled sequence */
if (N > 1) esl_sq_FormatName(shuff, "%s-shuffled-%d", sq->name, i);
else esl_sq_FormatName(shuff, "%s-shuffled", sq->name);
/* Output the resulting sequence */
esl_sqio_Write(ofp, shuff, outfmt, FALSE);
/* don't need to reuse the shuffled sequence: we will use exactly the same memory */
}
esl_sq_Reuse(sq);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
if (L>0) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return eslOK;
ERROR:
if (targ != NULL) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return status;
}
/* Each test sequence will contain one or two domains, depending on whether --single is set.
*/
static int
synthesize_positives(ESL_GETOPTS *go, struct cfg_s *cfg, char *testname, ESL_STACK *teststack, int *ret_ntest)
{
ESL_SQ *domain1, *domain2;
ESL_SQ *sq;
void *p;
int64_t L; /* total length of synthetic test seq */
int d1n, d2n; /* lengths of two domains */
int L1,L2,L3; /* lengths of three random regions */
int i,j;
int ntest = 0;
int ndomains = ( (esl_opt_GetBoolean(go, "--single") == TRUE) ? 1 : 2);
int status;
while (esl_stack_ObjectCount(teststack) >= ndomains)
{
ESL_RALLOC(cfg->test_lens, p, (cfg->ntest+1) * sizeof(struct testseq_s));
/* Pop our one or two test domains off the stack */
esl_stack_PPop(teststack, &p);
domain1 = p;
d1n = domain1->n;
if (ndomains == 2)
{
esl_stack_PPop(teststack, &p);
domain2 = p;
d2n = domain2->n;
}
else
{
domain2 = NULL;
d2n = 0;
}
/* Select a random total sequence length */
if (d1n+d2n > cfg->db_maxL) esl_fatal("can't construct test seq; no db seq >= %d residues\n", d1n+d2n);
do {
if (esl_ssi_FindNumber(cfg->dbfp->data.ascii.ssi, esl_rnd_Roll(cfg->r, cfg->db_nseq), NULL, NULL, NULL, &L, NULL) != eslOK)
esl_fatal("failed to look up a random seq");
} while (L < d1n+d2n);
/* Now figure out the embedding */
if (ndomains == 2)
{
/* Select random lengths of three flanking domains;
* Imagine picking two "insert after" points i,j in sequence 1..L', for
* L' = L-d1n-d2n (the total length of nonhomologous test seq)
*/
do {
i = esl_rnd_Roll(cfg->r, L - d1n - d2n + 1 ); /* i = 0..L' */
j = esl_rnd_Roll(cfg->r, L - d1n - d2n + 1 ); /* j = 0..L' */
} while (i > j);
/* now 1 .. i = random region 1 (if i==0, there's none);
* i+1 .. i+d1n = domain 1
* i+d1n+1 .. j+d1n = random region 2 (if i==j, there's none);
* j+d1n+1 .. j+d1n+d2n = domain 2
* j+d1n+d2n+1 .. L = random region 3 (if j == L-d1n-d2n, there's none);
*/
L1 = i;
L2 = j-i;
L3 = L - d1n - d2n - j;
}
else
{ /* embedding one domain */
i = esl_rnd_Roll(cfg->r, L - d1n + 1 ); /* i = 0..L' */
/* now 1 .. i = random region 1 (if i==0, there's none);
* i+1 .. i+d1n = domain 1
* i+d1n+1 .. L = random region 2 (if i==j, there's none);
*/
L1 = i;
L2 = L - d1n - L1;
L3 = 0;
}
sq = esl_sq_CreateDigital(cfg->abc);
esl_sq_GrowTo(sq, L);
sq->n = L;
if (ndomains == 2)
{
esl_sq_FormatName(sq, "%s/%d/%d-%d/%d-%d", testname, cfg->ntest, i+1, i+d1n, j+d1n+1, j+d1n+d2n);
esl_sq_FormatDesc(sq, "domains: %s %s", domain1->name, domain2->name);
}
else
{
esl_sq_FormatName(sq, "%s/%d/%d-%d", testname, cfg->ntest, i+1, i+d1n);
esl_sq_FormatDesc(sq, "domain: %s", domain1->name);
}
fprintf(cfg->possummfp, "%-35s %5d %5d %5d %5d %5d %5d", sq->name, (int) sq->n, L1, d1n, L2, d2n, L3);
sq->dsq[0] = sq->dsq[L+1] = eslDSQ_SENTINEL;
set_random_segment(go, cfg, cfg->possummfp, sq->dsq+1, L1);
memcpy(sq->dsq+i+1, domain1->dsq+1, sizeof(ESL_DSQ) * d1n);
fprintf(cfg->possummfp, " %-24s %5d %5d", domain1->name, 1, d1n);
set_random_segment(go, cfg, cfg->possummfp, sq->dsq+i+d1n+1, L2);
if (ndomains == 2)
{
memcpy(sq->dsq+j+d1n+1, domain2->dsq+1, sizeof(ESL_DSQ) * d2n);
fprintf(cfg->possummfp, " %-24s %5d %5d", domain2->name, 1, d2n);
set_random_segment(go, cfg, cfg->possummfp, sq->dsq+j+d1n+d2n+1, L3);
}
fprintf(cfg->possummfp, "\n");
cfg->test_lens[cfg->ntest].L = L;
cfg->test_lens[cfg->ntest].L1 = L1;
cfg->test_lens[cfg->ntest].d1n = d1n;
cfg->test_lens[cfg->ntest].L2 = L2;
cfg->test_lens[cfg->ntest].d2n = d2n;
cfg->test_lens[cfg->ntest].L3 = L3;
cfg->ntest++;
ntest++;
esl_sqio_Write(cfg->out_seqfp, sq, eslSQFILE_FASTA, FALSE);
esl_sq_Destroy(domain1);
if (ndomains == 2) esl_sq_Destroy(domain2);
esl_sq_Destroy(sq);
}
*ret_ntest = ntest;
return eslOK;
ERROR:
esl_fatal("Failure in synthesize_positives");
return status;
}
/* multifetch:
* given a file containing lines with one name or key per line;
* parse the file line-by-line;
* if we have an SSI index available, retrieve the seqs by key
* as we see each line;
* else, without an SSI index, store the keys in a hash, then
* read the entire seq file in a single pass, outputting seqs
* that are in our keylist.
*
* Note that with an SSI index, you get the seqs in the order they
* appear in the <keyfile>, but without an SSI index, you get seqs in
* the order they occur in the seq file.
*/
static void
multifetch(ESL_GETOPTS *go, FILE *ofp, char *keyfile, ESL_SQFILE *sqfp)
{
ESL_KEYHASH *keys = esl_keyhash_Create();
ESL_FILEPARSER *efp = NULL;
int nseq = 0;
int nkeys = 0;
char *key;
int keylen;
int keyidx;
int status;
if (esl_fileparser_Open(keyfile, NULL, &efp) != eslOK) esl_fatal("Failed to open key file %s\n", keyfile);
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
if (esl_fileparser_GetTokenOnLine(efp, &key, &keylen) != eslOK)
esl_fatal("Failed to read seq name on line %d of file %s\n", efp->linenumber, keyfile);
status = esl_keyhash_Store(keys, key, keylen, &keyidx);
if (status == eslEDUP) esl_fatal("seq key %s occurs more than once in file %s\n", key, keyfile);
/* if we have an SSI index, just fetch them as we go. */
if (sqfp->data.ascii.ssi != NULL) { onefetch(go, ofp, key, sqfp); nseq++; }
nkeys++;
}
/* If we don't have an SSI index, we haven't fetched anything yet; do it now. */
if (sqfp->data.ascii.ssi == NULL)
{
ESL_SQ *sq = esl_sq_Create();
while ((status = esl_sqio_Read(sqfp, sq)) == eslOK)
{
if ( (sq->name[0] != '\0' && esl_keyhash_Lookup(keys, sq->name, -1, NULL) == eslOK) ||
(sq->acc[0] != '\0' && esl_keyhash_Lookup(keys, sq->acc, -1, NULL) == eslOK))
{
if (esl_opt_GetBoolean(go, "-r") )
if (esl_sq_ReverseComplement(sq) != eslOK)
esl_fatal("Failed to reverse complement %s\n", sq->name);
esl_sqio_Write(ofp, sq, eslSQFILE_FASTA, FALSE);
nseq++;
}
esl_sq_Reuse(sq);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
esl_sq_Destroy(sq);
}
if (nkeys != nseq) esl_fatal("Tried to retrieve %d keys, but only retrieved %d sequences\n", nkeys, nseq);
if (ofp != stdout) printf("\nRetrieved %d sequences.\n", nseq);
esl_keyhash_Destroy(keys);
esl_fileparser_Close(efp);
return;
}
