static void
utest_goodfile(char *filename, int testnumber, int expected_alphatype, int expected_nseq, int expected_alen)
{
ESL_ALPHABET *abc = NULL;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa1 = NULL;
ESL_MSA *msa2 = NULL;
char tmpfile1[32] = "esltmpXXXXXX";
char tmpfile2[32] = "esltmpXXXXXX";
FILE *ofp = NULL;
int status;
/* guessing both the format and the alphabet should work: this is a digital open */
/* PSIBLAST format is autodetected as SELEX, which is fine - selex parser is more general */
if ( (status = eslx_msafile_Open(&abc, filename, NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK) esl_fatal("psiblast good file test %d failed: digital open", testnumber);
if (afp->format != eslMSAFILE_SELEX) esl_fatal("psiblast good file test %d failed: format autodetection", testnumber);
if (abc->type != expected_alphatype) esl_fatal("psiblast good file test %d failed: alphabet autodetection", testnumber);
afp->format = eslMSAFILE_PSIBLAST;
/* This is a digital read, using <abc>. */
if ( (status = esl_msafile_psiblast_Read(afp, &msa1)) != eslOK) esl_fatal("psiblast good file test %d failed: msa read, digital", testnumber);
if (msa1->nseq != expected_nseq || msa1->alen != expected_alen) esl_fatal("psiblast good file test %d failed: nseq/alen", testnumber);
if (esl_msa_Validate(msa1, NULL) != eslOK) esl_fatal("psiblast good file test %d failed: msa1 invalid", testnumber);
eslx_msafile_Close(afp);
/* write it back out to a new tmpfile (digital write) */
if ( (status = esl_tmpfile_named(tmpfile1, &ofp)) != eslOK) esl_fatal("psiblast good file test %d failed: tmpfile creation", testnumber);
if ( (status = esl_msafile_psiblast_Write(ofp, msa1)) != eslOK) esl_fatal("psiblast good file test %d failed: msa write, digital", testnumber);
fclose(ofp);
/* now open and read it as text mode, in known format. (We have to pass fmtd now, to deal with the possibility of a nonstandard name width) */
if ( (status = eslx_msafile_Open(NULL, tmpfile1, NULL, eslMSAFILE_PSIBLAST, NULL, &afp)) != eslOK) esl_fatal("psiblast good file test %d failed: text mode open", testnumber);
if ( (status = esl_msafile_psiblast_Read(afp, &msa2)) != eslOK) esl_fatal("psiblast good file test %d failed: msa read, text", testnumber);
if (msa2->nseq != expected_nseq || msa2->alen != expected_alen) esl_fatal("psiblast good file test %d failed: nseq/alen", testnumber);
if (esl_msa_Validate(msa2, NULL) != eslOK) esl_fatal("psiblast good file test %d failed: msa2 invalid", testnumber);
eslx_msafile_Close(afp);
/* write it back out to a new tmpfile (text write) */
if ( (status = esl_tmpfile_named(tmpfile2, &ofp)) != eslOK) esl_fatal("psiblast good file test %d failed: tmpfile creation", testnumber);
if ( (status = esl_msafile_psiblast_Write(ofp, msa2)) != eslOK) esl_fatal("psiblast good file test %d failed: msa write, text", testnumber);
fclose(ofp);
esl_msa_Destroy(msa2);
/* open and read it in digital mode */
if ( (status = eslx_msafile_Open(&abc, tmpfile1, NULL, eslMSAFILE_PSIBLAST, NULL, &afp)) != eslOK) esl_fatal("psiblast good file test %d failed: 2nd digital mode open", testnumber);
if ( (status = esl_msafile_psiblast_Read(afp, &msa2)) != eslOK) esl_fatal("psiblast good file test %d failed: 2nd digital msa read", testnumber);
if (esl_msa_Validate(msa2, NULL) != eslOK) esl_fatal("psiblast good file test %d failed: msa2 invalid", testnumber);
eslx_msafile_Close(afp);
/* this msa <msa2> should be identical to <msa1> */
if (esl_msa_Compare(msa1, msa2) != eslOK) esl_fatal("psiblast good file test %d failed: msa compare", testnumber);
remove(tmpfile1);
remove(tmpfile2);
esl_msa_Destroy(msa1);
esl_msa_Destroy(msa2);
esl_alphabet_Destroy(abc);
}
static void
read_test_msas_digital(char *pbfile, char *stkfile)
{
char msg[] = "PSIBLAST msa digital read unit test failed";
ESL_ALPHABET *abc = NULL;
ESLX_MSAFILE *afp1 = NULL;
ESLX_MSAFILE *afp2 = NULL;
ESL_MSA *msa1, *msa2, *msa3, *msa4;
FILE *pbfp, *stkfp;
char pbfile2[32] = "esltmppb2XXXXXX";
char stkfile2[32] = "esltmpstk2XXXXXX";
if ( eslx_msafile_Open(&abc, pbfile, NULL, eslMSAFILE_PSIBLAST, NULL, &afp1) != eslOK) esl_fatal(msg);
if ( !abc || abc->type != eslAMINO) esl_fatal(msg);
if ( eslx_msafile_Open(&abc, stkfile, NULL, eslMSAFILE_STOCKHOLM, NULL, &afp2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa1) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa2) != eslOK) esl_fatal(msg);
if ( esl_msa_Compare(msa1, msa2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa3) != eslEOF) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa3) != eslEOF) esl_fatal(msg);
eslx_msafile_Close(afp2);
eslx_msafile_Close(afp1);
/* Now write stk to psiblast file, and vice versa; then retest */
if ( esl_tmpfile_named(pbfile2, &pbfp) != eslOK) esl_fatal(msg);
if ( esl_tmpfile_named(stkfile2, &stkfp) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Write (pbfp, msa2) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Write(stkfp, msa1, eslMSAFILE_STOCKHOLM) != eslOK) esl_fatal(msg);
fclose(pbfp);
fclose(stkfp);
if ( eslx_msafile_Open(&abc, pbfile2, NULL, eslMSAFILE_PSIBLAST, NULL, &afp1) != eslOK) esl_fatal(msg);
if ( eslx_msafile_Open(&abc, stkfile2, NULL, eslMSAFILE_STOCKHOLM, NULL, &afp2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa3) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa4) != eslOK) esl_fatal(msg);
if ( esl_msa_Compare(msa3, msa4) != eslOK) esl_fatal(msg);
remove(pbfile2);
remove(stkfile2);
eslx_msafile_Close(afp2);
eslx_msafile_Close(afp1);
esl_msa_Destroy(msa1);
esl_msa_Destroy(msa2);
esl_msa_Destroy(msa3);
esl_msa_Destroy(msa4);
esl_alphabet_Destroy(abc);
}
static void
read_test_msas_text(char *pbfile, char *stkfile)
{
char msg[] = "PSIBLAST msa text-mode read unit test failed";
ESLX_MSAFILE *afp1 = NULL;
ESLX_MSAFILE *afp2 = NULL;
ESL_MSA *msa1, *msa2, *msa3, *msa4;
FILE *pbfp, *stkfp;
char pbfile2[32] = "esltmppb2XXXXXX";
char stkfile2[32] = "esltmpstk2XXXXXX";
/* vvvv-- everything's the same as the digital utest except these NULLs */
if ( eslx_msafile_Open(NULL, pbfile, NULL, eslMSAFILE_PSIBLAST, NULL, &afp1) != eslOK) esl_fatal(msg);
if ( eslx_msafile_Open(NULL, stkfile, NULL, eslMSAFILE_STOCKHOLM, NULL, &afp2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa1) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa2) != eslOK) esl_fatal(msg);
if ( esl_msa_Compare(msa1, msa2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa3) != eslEOF) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa3) != eslEOF) esl_fatal(msg);
eslx_msafile_Close(afp2);
eslx_msafile_Close(afp1);
if ( esl_tmpfile_named(pbfile2, &pbfp) != eslOK) esl_fatal(msg);
if ( esl_tmpfile_named(stkfile2, &stkfp) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Write (pbfp, msa2) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Write(stkfp, msa1, eslMSAFILE_STOCKHOLM) != eslOK) esl_fatal(msg);
fclose(pbfp);
fclose(stkfp);
if ( eslx_msafile_Open(NULL, pbfile2, NULL, eslMSAFILE_PSIBLAST, NULL, &afp1) != eslOK) esl_fatal(msg);
if ( eslx_msafile_Open(NULL, stkfile2, NULL, eslMSAFILE_STOCKHOLM, NULL, &afp2) != eslOK) esl_fatal(msg);
if ( esl_msafile_psiblast_Read (afp1, &msa3) != eslOK) esl_fatal(msg);
if ( esl_msafile_stockholm_Read(afp2, &msa4) != eslOK) esl_fatal(msg);
if ( esl_msa_Compare(msa3, msa4) != eslOK) esl_fatal(msg);
remove(pbfile2);
remove(stkfile2);
eslx_msafile_Close(afp2);
eslx_msafile_Close(afp1);
esl_msa_Destroy(msa1);
esl_msa_Destroy(msa2);
esl_msa_Destroy(msa3);
esl_msa_Destroy(msa4);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
char *msafile = esl_opt_GetArg(go, 1);
ESL_ALPHABET *abc = NULL;
int infmt = eslMSAFILE_UNKNOWN;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
FILE *ofp = stdout;
int nali = 0;
int namewidth;
double pid;
int nid, n;
int i,j;
int status;
/* allow user to assert the input MSA alphabet */
if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
/* allow user to assert the input MSA format */
if (esl_opt_IsOn(go, "--informat") &&
(infmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"))) == eslMSAFILE_UNKNOWN)
esl_fatal("%s is not a valid MSA file format for --informat", esl_opt_GetString(go, "--informat"));
/* digital open */
if ( ( status = eslx_msafile_Open(&abc, msafile, NULL, infmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
while ((status = eslx_msafile_Read(afp, &msa)) == eslOK)
{
nali++;
namewidth = esl_str_GetMaxWidth(msa->sqname, msa->nseq);
for (i = 0; i < msa->nseq; i++)
for (j = i+1; j < msa->nseq; j++)
{
esl_dst_XPairId(abc, msa->ax[i], msa->ax[j], &pid, &nid, &n);
fprintf(ofp, "%-*s %-*s %6.2f %6d %6d\n", namewidth, msa->sqname[i], namewidth, msa->sqname[j], pid*100.0, nid, n);
}
esl_msa_Destroy(msa);
}
if (nali == 0 || status != eslEOF) eslx_msafile_ReadFailure(afp, status);
eslx_msafile_Close(afp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_Create(0);
char *msafile = esl_opt_GetArg(go, 1);
int fmt = eslMSAFILE_UNKNOWN;
ESL_ALPHABET *abc = NULL;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
int textmode = esl_opt_GetBoolean(go, "--text");
int nali = 0;
int status;
/* If you know the alphabet you want, create it - you'll pass it to eslx_msafile_Open() */
if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
/* Open in text or digital mode.
* To let the Open() function autoguess the format, you pass <infmt=eslMSAFILE_UNKNOWN>.
* To let it autoguess the alphabet, you set <abc=NULL> and pass <&abc>.
* To open in text mode instead of digital, you pass <NULL> for the alphabet argument.
* eslx_msafile_OpenFailure() is a convenience, printing various diagnostics of any
* open failure to <stderr>. You can of course handle your own diagnostics instead.
*/
if (textmode) status = eslx_msafile_Open(NULL, msafile, NULL, fmt, NULL, &afp);
else status = eslx_msafile_Open(&abc, msafile, NULL, fmt, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
fmt = afp->format;
while ((status = eslx_msafile_Read(afp, &msa)) == eslOK)
{
/* if digital MSA: msa->ax[idx=0..nseq-1][acol=1..alen] is the alignment data;
* if text MSA: msa->aseq[idx=0..nseq-1][acol=0..alen-1] */
nali++;
/* permute it */
esl_msashuffle_PermuteSequenceOrder(rng, msa);
eslx_msafile_Write(stdout, msa, fmt);
esl_msa_Destroy(msa);
}
if (nali == 0 || status != eslEOF) eslx_msafile_ReadFailure(afp, status); /* a convenience, like eslx_msafile_OpenFailure() */
esl_alphabet_Destroy(abc);
eslx_msafile_Close(afp);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
exit(0);
}
int main(int argc, char **argv)
{
ESLX_MSAFILE *afp;
ESL_MSA *msa;
int i;
int status;
if ( (status = eslx_msafile_Open(NULL, argv[1], NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
if ( (status = eslx_msafile_Read(afp, &msa)) != eslOK)
eslx_msafile_ReadFailure(afp, status);
eslx_msafile_Close(afp);
esl_msaweight_GSC(msa);
for (i = 0; i < msa->nseq; i++)
printf("%20s %f\n", msa->sqname[i], msa->wgt[i]);
return 0;
}
int
main(int argc, char **argv)
{
char *filename = argv[1];
int fmt = eslMSAFILE_PSIBLAST;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
int status;
if ( (status = eslx_msafile_Open(NULL, filename, NULL, fmt, NULL, &afp)) != eslOK) eslx_msafile_OpenFailure(afp, status);
if ( (status = esl_msafile_psiblast_Read(afp, &msa)) != eslOK) eslx_msafile_ReadFailure(afp, status);
printf("%6d seqs, %5d columns\n", msa->nseq, (int) msa->alen);
esl_msafile_psiblast_Write(stdout, msa);
esl_msa_Destroy(msa);
eslx_msafile_Close(afp);
exit(0);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
char *filename = esl_opt_GetArg(go, 1);
int infmt = eslMSAFILE_UNKNOWN;
ESL_ALPHABET *abc = NULL;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
int status;
if (esl_opt_GetBoolean(go, "-1")) infmt = eslMSAFILE_PSIBLAST; /* override format autodetection */
if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
/* Text mode: pass NULL for alphabet.
* Digital mode: pass ptr to expected ESL_ALPHABET; and if abc=NULL, alphabet is guessed
*/
if (esl_opt_GetBoolean(go, "-t")) status = eslx_msafile_Open(NULL, filename, NULL, infmt, NULL, &afp);
else status = eslx_msafile_Open(&abc, filename, NULL, infmt, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
if ((status = esl_msafile_psiblast_Read(afp, &msa)) != eslOK)
eslx_msafile_ReadFailure(afp, status);
printf("alphabet: %s\n", (abc ? esl_abc_DecodeType(abc->type) : "none (text mode)"));
printf("# of seqs: %d\n", msa->nseq);
printf("# of cols: %d\n", (int) msa->alen);
printf("\n");
if (! esl_opt_GetBoolean(go, "-q"))
esl_msafile_psiblast_Write(stdout, msa);
esl_msa_Destroy(msa);
eslx_msafile_Close(afp);
if (abc) esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
exit(0);
}
/**
* <pre>
* Function: eslx_msafile_Open()
* Synopsis: Open a multiple sequence alignment file for input.
*
* Purpose: Open a multiple sequence alignment file <msafile> for input.
* Return an open <ESLX_MSAFILE> handle in <*ret_afp>.
*
* <msafile> is usually the name of a file. Alignments may
* also be read from standard input, or from
* gzip-compressed files. If <msafile> is ``-'', alignment
* input is taken from the standard input stream. If
* <msafile> ends in ``.gz'', alignment input is read
* through a pipe from <gzip -dc>.
*
* <byp_abc>, <env>, <format>, and <fmtd> support a variety
* of optional/advanced operations, as described
* below. Minimally, a caller can set <byp_abc> to <NULL>,
* <format> to <eslMSAFILE_UNKNOWN>, and <fmtd> to <NULL>,
* and <msafile> will be opened in text mode; in the
* current working directory; and its format will be
* autodetected.
*
* The <byp_abc> argument controls whether data are to be
* read in text or digital mode. In digital mode, alignment
* data are immediately digitized into an Easel internal
* alphabet (which among other things, allows various
* things to operate on sequence data more efficiently) and
* because an expected alphabet is known, parsers are able
* to detect invalid characters. The caller may either
* provide an alphabet (thus asserting what it's expected
* to be), or have <eslx_msafile_Open()> look at the file
* and guess what alphabet it appears to be (DNA or amino
* acid code, usually). In text mode, alignment data are
* read verbatim. It might be advantageous for an
* application to read in text mode -- for example, if a
* variant alignment format is using characters in some
* special way, and you need to deal with them specially.
* All this goes through the setting of the passed-by-reference
* alphabet pointer <byp_abc>. If caller passes NULL for
* the <byp_abc> argument, input is in text mode. If caller
* provides a valid non-NULL <byp_abc> pointer but
* <*byp_abc> is NULL (that is, caller has declared
* <ESL_ALPHABET *abc = NULL> and passed <&abc> as an
* argument), then we attempt to guess the digital alphabet
* using <eslx_msafile_GuessAlphabet()>, based on the first
* alignment in the input. In this case, the new alphabet
* is allocated here and returned to the caller. If caller
* provides a digital alphabet (that is, <ESL_ALPHABET *abc
* = esl_alphabet_Create...()> and passed <&abc>), that's
* the alphabet we use.
*
* The <env> argument controls where we search for the
* <msafile>. If <env> is <NULL>, only the current working
* directory is checked. Optionally, caller can provide in
* <env> the name of an environment variable ("PFAMDB",
* perhaps), in which the routine can find a
* colon-delimited list of directories. Then, if <msafile>
* is not found in the current working directory, we look
* for it in these directories, in the order they're
* listed.
*
* The <format> argument allows the caller to either allow
* <eslx_msafile_Open()> to autodetect the file format of
* <msafile>, or to assert that it knows the file is in a
* particular format. If <format> is <eslMSAFILE_UNKNOWN>,
* format autodetection is performed. Other valid codes include:
* | <eslMSAFILE_STOCKHOLM> | Stockholm format |
* | <eslMSAFILE_AFA> | Aligned FASTA format |
* | <eslMSAFILE_CLUSTAL> | Clustal format (strict) |
* | <eslMSAFILE_CLUSTALLIKE> | Clustal-like (MUSCLE, PROBCONS...) |
* | <eslMSAFILE_PHYLIP> | PHYLIP interleaved format |
* | <eslMSAFILE_PHYLIPS> | PHYLIP sequential format |
* | <eslMSAFILE_A2M> | UCSC SAM A2M (dotless or dotful) |
* | <eslMSAFILE_PSIBLAST> | NCBI PSI-BLAST |
* | <eslMSAFILE_SELEX> | a general alignment block format |
*
* The <fmtd> argument is an optional pointer to a
* <ESLX_MSAFILE_FMTDATA> structure that the caller may
* initialize and provide, in order to assert any
* additional unusual constraints on the input format --
* for example, to dictate that a PHYLIP format file has
* some nonstandard name field width. Generally, though,
* <fmtd> will be <NULL>.
*
* Args: byp_abc - digital alphabet to use, or NULL for text mode
* if <*byp_abc> is NULL, guess the digital alphabet,
* create it, and return it in <*byp_abc>.
* If <*byp_abc> is a digital alphabet, use it.
* msafile - name of alignment input to open;
* if "-", read standard input;
* if "*.gz", read through a <gzip -dc> pipe.
* env - <NULL>, or the name of an environment variable
* containing colon-delimited list of directories
* in which to search for <msafile> (e.g. "PFAMDB").
* format - format code, such as <eslMSAFILE_STOCKHOLM>;
* or <eslMSAFILE_UNKNOWN> to autodetect format.
* fmtd - <NULL>, or a pointer to an initialized
* <ESLX_MSAFILE_FMTDATA> structure, containing
* any additional unusual constraints to apply
* to the input format.
* *ret_afp - RETURN: open MSA input stream.
*
* Returns: <eslOK> on success, and <*ret_afp> is the newly opened msa file.
*
* <eslENOTFOUND> if <msafile> doesn't exist or can't be
* opened for reading; or (in the case of a <.gz> file) if
* a <gzip> executable doesn't exist in user's <PATH> or
* can't be executed. <afp->errmsg> is something like
* "couldn't open %s for reading", with <%s> being the
* name of the msafile.
*
* <eslENOFORMAT> if we tried to autodetect the file format
* (caller provided <format=eslMSAFILE_UNKNOWN>), and
* failed. <afp->errmsg> is something like "couldn't
* determine alignment input format".
*
* <eslENOALPHABET> if we tried to autodetect the alphabet
* (caller provided <&abc>, <abc=NULL> to request digital
* mode w/ alphabet autodetection) but the alphabet could
* not be reliably guessed.
*
* <eslFAIL> in the case of a <.gz> file and the <gzip -dc>
* command fails on it.
*
* On any of these normal errors, <*ret_afp> is returned in
* an error state, containing a user-directed error message
* in <afp->errmsg> and (if relevant) the full path to
* <msafile> that we attempted to open in
* <afp->bf->filename>. See <eslx_msafile_OpenFailure()> for
* a function that gives a standard way of reporting these
* diagnostics to <stderr>.
*
* Throws: <eslEMEM> on allocation failure.
* <eslESYS> on a system call failure, such as <fread()>.
* <eslEINVAL> if we tried to use <stdin> but the <stdin> stream was
* invalid (in an error state, <NULL>, or at <EOF>).
* On thrown exceptions, <*ret_afp> is <NULL>.
* </pre>
*/
int
profillic_eslx_msafile_Open(ESL_ALPHABET **byp_abc, const char *msafile, const char *env, int format, ESLX_MSAFILE_FMTDATA *fmtd, ESLX_MSAFILE **ret_afp)
{
ESLX_MSAFILE *afp = NULL;
int status;
if ( (status = profillic_msafile_Create(&afp)) != eslOK) goto ERROR;
if ((status = esl_buffer_Open(msafile, env, &(afp->bf))) != eslOK)
ESL_XFAIL(status, afp->errmsg, "%s", afp->bf->errmsg); /* ENOTFOUND; FAIL are normal here */
if ( (status = profillic_msafile_OpenBuffer(byp_abc, afp->bf, format, fmtd, afp)) != eslOK) goto ERROR;
*ret_afp = afp;
return eslOK;
ERROR: /* on normal errors, afp is returned in an error state */
if (status == eslENOTFOUND || status == eslFAIL || status == eslEFORMAT || status == eslENODATA || eslENOALPHABET)
{ afp->abc = NULL; *ret_afp = afp;}
else
{ if (afp) eslx_msafile_Close(afp); *ret_afp = NULL; }
return status;
}
int main(int argc, char **argv)
{
ESLX_MSAFILE *afp;
ESL_MSA *msa;
ESL_DMATRIX *P;
int i,j;
double min, avg, max;
int status;
if ((status = eslx_msafile_Open(NULL, argv[1], NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
if ((status = eslx_msafile_Read(afp, &msa)) != eslOK)
eslx_msafile_ReadFailure(afp, status);
esl_dst_CPairIdMx(msa->aseq, msa->nseq, &P);
min = 1.0;
max = 0.0;
avg = 0.0;
for (i = 0; i < msa->nseq; i++)
for (j = i+1; j < msa->nseq; j++)
{
avg += P->mx[i][j];
if (P->mx[i][j] < min) min = P->mx[i][j];
if (P->mx[i][j] > max) max = P->mx[i][j];
}
avg /= (double) (msa->nseq * (msa->nseq-1) / 2);
printf("Average pairwise %% id: %.1f%%\n", avg * 100.);
printf("Minimum pairwise %% id: %.1f%%\n", min * 100.);
printf("Maximum pairwise %% id: %.1f%%\n", max * 100.);
esl_dmatrix_Destroy(P);
esl_msa_Destroy(msa);
eslx_msafile_Close(afp);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
ESL_ALPHABET *abc = NULL; /* biological alphabet */
char *alifile = NULL; /* alignment file name */
int fmt; /* format code for alifiles */
ESLX_MSAFILE *afp = NULL; /* open alignment file */
ESL_MSA *msa = NULL; /* multiple sequence alignment */
int status; /* easel return code */
int do_info = TRUE; /* TRUE if -i */
int do_max = FALSE; /* TRUE if -x */
int do_ffreq = FALSE; /* TRUE if --ffreq */
int do_fmin = FALSE; /* TRUE if --fmin */
float fthresh = 0.; /* <x> from -f <x> */
int do_remove_bps = FALSE; /* TRUE if -r */
int do_consistent = FALSE; /* TRUE if -c */
int do_indi2cons = FALSE; /* TRUE if --indi <x> */
int have_cons; /* TRUE if first alignment has consensus sequence */
int do_newcons = FALSE; /* TRUE if we're creating a new consensus structure
* and outputing a new alignment (if -x -f -c or --indi)
*/
int do_a = FALSE; /* TRUE if -a */
char *indi; /* for <x> from --indi <x> */
int nindi_read; /* number of individual sequence SS lines we've read for current alignment */
int a; /* counter over seqs */
int i, i2; /* counter over residues */
int j, j2; /* counter over residues */
int nali; /* counter over alignments */
int **bp = NULL; /* bp[i][j] is number of individual bps exist between aln cols i and j */
int *cur_ct = NULL; /* ct array of basepairs for current sequence */
int *cons_ct = NULL; /* ct array of basepairs for SS_cons being created */
int *xcons_ct = NULL; /* ct array of basepairs for existing SS_cons */
int *ngaps = NULL; /* number of gaps in each alignment position */
FILE *ofp; /* output file (default is stdout) */
int be_verbose = FALSE; /* TRUE to print extra info */
int seqthresh; /* sequence number threshold for defining a bp as consensus (int) ((fthresh * nseq) + 0.5)*/
char *sscons = NULL; /* the new SS_cons line */
FILE *lfp = NULL; /* file to list sequences with conflicting bps to */
int nlist = 0; /* number of sequences listed to list file */
int *nconflictsA; /* number of conflicting bps in seq a's individual structure annotation */
int nconflicts_total = 0; /* total number of conflicts */
int nconflicts_list = 0; /* total number of conflicts in sequences listed to file <x> from -l <x> */
int noverlaps_total = 0; /* total number of overlaps */
int nconsistent_total = 0; /* total number of consistent bps */
int nbps_total = 0; /* total number of bps */
int *nconsistentA; /* number of consistent bps in seq a's individual structure annotation */
int *noverlapsA; /* number of bps in seq a's indi structure that overlap with consensus structure */
int *nbpsA; /* number of bps in seq a's indi structure that overlap with consensus structure */
int ncons_bps = 0; /* number of bps in consensus structure */
int max_noverlaps_aidx;
int max_nconsistent_aidx;
int max_nbps_aidx;
int *removebp; /* removebp[i] is TRUE remove consensus bp [i]:xcons_ct[i] */
int *has_conflict;
int *nmates_l2r; /* half matrix, nmate_l2r[i] = <x>, i < nmate_l2r[i], there are <x> different right mates j for i */
int *nmates_r2l; /* half matrix, nmate_r2l[j] = <x>, j < nmate_r2l[j], there are <x> different left mates i for j */
int lmax; /* with -l, maximum number of conflicts to allow */
int namewidth = 18; /* length of 'SS_cons(consensus)' */
char *namedashes = NULL; /* to store underline for seq name */
/* --fmin related variables */
int nbps = 0;
int prev_nbps = -1;
float fmin;
int inconsistent_flag;
int pknot_flag;
int k,l;
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK ||
esl_opt_VerifyConfig(go) != eslOK)
{
printf("Failed to parse command line: %s\n", go->errbuf);
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
if (esl_opt_GetBoolean(go, "-h") )
{
esl_banner(stdout, argv[0], banner);
esl_usage (stdout, argv[0], usage);
puts("\nwhere basic options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\noptions for defining a new consensus structure (all of these require -o):");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\noptions for listing sequences based on structure:");
esl_opt_DisplayHelp(stdout, go, 3, 2, 80);
exit(0);
}
if (esl_opt_ArgNumber(go) != 1)
{
printf("Incorrect number of command line arguments.\n");
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
alifile = esl_opt_GetArg(go, 1);
fmt = eslMSAFILE_STOCKHOLM;
/***********************************************
* Open the MSA file; determine alphabet; set for digital input
***********************************************/
if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
if ( (status = eslx_msafile_Open(&abc, alifile, NULL, fmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
/* open output file */
if (esl_opt_GetString(go, "-o") != NULL) {
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
esl_fatal("Failed to open -o output file %s\n", esl_opt_GetString(go, "-o"));
} else ofp = NULL;
if (esl_opt_GetString(go, "-l") != NULL) {
if ((lfp = fopen(esl_opt_GetString(go, "-l"), "w")) == NULL)
esl_fatal("Failed to open -l output file %s\n", esl_opt_GetString(go, "-l"));
}
/* determine if we're creating a structure */
do_max = esl_opt_GetBoolean(go, "-x");
if(!(esl_opt_IsDefault(go, "--ffreq"))) {
do_ffreq = TRUE;
fthresh = esl_opt_GetReal(go, "--ffreq");
}
if(!(esl_opt_IsDefault(go, "--fmin"))) {
do_fmin = TRUE;
}
do_remove_bps = esl_opt_GetBoolean(go, "-r");
do_consistent = esl_opt_GetBoolean(go, "-c");
if(!(esl_opt_IsDefault(go, "--indi"))) {
do_indi2cons = TRUE;
}
if(do_max || do_ffreq || do_fmin || do_remove_bps || do_consistent || do_indi2cons) {
do_newcons = TRUE;
}
do_a = esl_opt_GetBoolean(go, "-a");
if(do_a || do_max || do_ffreq || do_fmin || do_remove_bps || do_consistent || do_indi2cons) {
do_info = FALSE;
}
/***********************************************
* Read MSAs one at a time.
***********************************************/
nali = 0;
have_cons = FALSE;
lmax = esl_opt_GetInteger(go, "--lmax");
if(esl_opt_GetBoolean(go, "-v")) be_verbose = TRUE;
while ((status = eslx_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
nali++;
/* determine max length name */
namewidth = 18; /* length of 'SS_cons(consensus)' */
for(i = 0; i < msa->nseq; i++) namewidth = ESL_MAX(namewidth, strlen(msa->sqname[i]));
if(namedashes != NULL) { free(namedashes); }
ESL_ALLOC(namedashes, sizeof(char) * namewidth+1);
namedashes[namewidth] = '\0';
for(i = 0; i < namewidth; i++) namedashes[i] = '-';
ESL_ALLOC(sscons, sizeof(char) * (msa->alen+1));
ESL_ALLOC(cur_ct, sizeof(int) * (msa->alen+1));
ESL_ALLOC(cons_ct, sizeof(int) * (msa->alen+1));
ESL_ALLOC(xcons_ct, sizeof(int) * (msa->alen+1));
ESL_ALLOC(bp, sizeof(int *) * (msa->alen+1));
ESL_ALLOC(removebp, sizeof(int) * (msa->alen+1));
ESL_ALLOC(has_conflict, sizeof(int) * (msa->alen+1));
ESL_ALLOC(nmates_l2r, sizeof(int) * (msa->alen+1));
ESL_ALLOC(nmates_r2l, sizeof(int) * (msa->alen+1));
esl_vec_ISet(cur_ct, (msa->alen+1), 0);
esl_vec_ISet(cons_ct, (msa->alen+1), 0);
esl_vec_ISet(xcons_ct, (msa->alen+1), 0);
esl_vec_ISet(removebp, (msa->alen+1), FALSE);
esl_vec_ISet(has_conflict, (msa->alen+1), FALSE);
esl_vec_ISet(nmates_l2r, (msa->alen+1), 0);
esl_vec_ISet(nmates_r2l, (msa->alen+1), 0);
ESL_ALLOC(nconflictsA, sizeof(int) * msa->nseq);
ESL_ALLOC(noverlapsA, sizeof(int) * msa->nseq);
ESL_ALLOC(nconsistentA, sizeof(int) * msa->nseq);
ESL_ALLOC(nbpsA, sizeof(int) * msa->nseq);
esl_vec_ISet(nconflictsA, msa->nseq, 0);
esl_vec_ISet(noverlapsA, msa->nseq, 0);
esl_vec_ISet(nconsistentA, msa->nseq, 0);
esl_vec_ISet(nbpsA, msa->nseq, 0);
max_noverlaps_aidx = max_nconsistent_aidx = max_nbps_aidx = 0;
nconsistent_total = nbps_total = noverlaps_total = nconflicts_total = nconflicts_list = 0;
for(i = 1; i <= msa->alen; i++) {
ESL_ALLOC(bp[i], sizeof(int) * (msa->alen+1));
esl_vec_ISet(bp[i], (msa->alen+1), 0);
}
/* make sure we have ss_cons and indi ss if we need it */
if(msa->ss_cons == NULL && do_remove_bps) esl_fatal("-r requires all alignments have SS_cons annotation, alignment %d does not.", nali);
if(msa->ss == NULL && do_max) esl_fatal("-x requires all alignments have individual SS annotation, alignment %d does not.", nali);
if(msa->ss == NULL && do_consistent) esl_fatal("-c requires all alignments have individual SS annotation, alignment %d does not.", nali);
if(msa->ss == NULL && do_indi2cons) esl_fatal("--indi requires all alignments have individual SS annotation, alignment %d does not.", nali);
if(msa->ss == NULL && do_ffreq) esl_fatal("--ffreq requires all alignments have individual SS annotation, alignment %d does not.", nali);
if(msa->ss == NULL && do_fmin) esl_fatal("--fmin requires all alignments have individual SS annotation, alignment %d does not.", nali);
if(msa->ss_cons != NULL) {
if((status = esl_wuss2ct(msa->ss_cons, msa->alen, xcons_ct)) != eslOK) {
esl_fatal("Existing SS_cons for alignment %d is invalid.", nali);
}
ncons_bps = 0;
for(i = 1; i <= msa->alen; i++)
if(xcons_ct[i] != 0 && i < xcons_ct[i])
ncons_bps++;
if(nali > 1 && !have_cons)
esl_fatal("the first aln has SS_cons but aln %d lacks it, if one has it, they all must.", nali);
if(nali == 1) have_cons = TRUE;
}
else if (lfp != NULL) {
esl_fatal("the -l option requires existing SS_cons annotation, aln %d lacks it.", nali);
}
else if (do_remove_bps) {
esl_fatal("the -r option requires existing SS_cons annotation, aln %d lacks it.", nali);
}
else if (do_consistent) {
esl_fatal("the -c option requires existing SS_cons annotation, aln %d lacks it.", nali);
}
else {
if(nali > 1 && have_cons)
esl_fatal("the first aln does not have SS_cons but aln %d does, if one has it, they all must.", nali);
}
if(do_info) {
printf("# Per-sequence basepair information:\n");
printf("# Alignment file: %s\n", alifile);
printf("# Alignment idx: %d\n", nali);
if(msa->name != NULL) { printf("# Alignment name: %s\n", msa->name); }
if(have_cons) {
printf("#\n");
printf("# indibp: number of basepairs in the individual sequence SS annotation\n");
printf("# ovrlap: number of indibp basepairs that also exist as consensus basepairs\n");
printf("# cnsist: number of indibp basepairs that do not conflict with any consensus basepairs\n");
printf("# cnflct: number of indibp basepairs that conflict with >= 1 consensus basepairs\n");
printf("#\n");
printf("# A conflict exists between two basepairs in different structures, one between columns i and j\n");
printf("# and the other between columns k and l, if (i == k and j != l) or (j == l and i != k).\n");
printf("#\n");
printf("# %-*s %6s %6s %6s %6s\n", namewidth, "seqname", "indibp", "ovrlap", "cnsist", "cnflct");
printf("# %-*s %6s %6s %6s %6s\n", namewidth, namedashes, "------", "------", "-----", "------");
}
else {
printf("# %-*s %6s\n", namewidth, "seqname", "nbp");
printf("# %-*s %6s\n", namewidth, namedashes, "------");
}
}
nindi_read = 0;
for (a = 0; a < msa->nseq; a++) {
if(msa->ss != NULL && msa->ss[a] != NULL) {
if((status = esl_wuss2ct(msa->ss[a], msa->alen, cur_ct)) != eslOK) {
esl_fatal("SS annotation for sequence %d, aln %d is invalid.\n", (a+1), nali);
}
nindi_read++;
for(i = 1; i <= msa->alen; i++) {
if(i < cur_ct[i]) {
bp[i][cur_ct[i]]++;
if(bp[i][cur_ct[i]] == 1) {
nmates_l2r[i]++;
nmates_r2l[cur_ct[i]]++;
}
}
}
for(i = 1; i <= msa->alen; i++) {
if(cur_ct[i] != 0 && i < cur_ct[i]) {
if(xcons_ct[i] == cur_ct[i]) noverlapsA[a]++;
if((xcons_ct[i] != 0) && (xcons_ct[i] != cur_ct[i])) {
if(be_verbose) { printf("ali: %2d seq %3d (%s) bp %4d:%4d conflicts with consensus bp %4d:%4d\n", nali, a, msa->sqname[a], i, cur_ct[i], i, xcons_ct[i]); }
nconflictsA[a]++;
/* indi bp i:cur_ct[i] conflicts with i:xcons_ct[i] */
removebp[i] = TRUE;
removebp[xcons_ct[i]] = TRUE;
}
else if((xcons_ct[cur_ct[i]] != 0) && (xcons_ct[cur_ct[i]] != i) && (cur_ct[xcons_ct[cur_ct[i]]] == 0)) {
if(be_verbose) { printf("ali: %2d seq %3d (%s) bp %4d:%4d conflicts with consensus bp %4d:%4d\n", nali, a, msa->sqname[a], xcons_ct[i], cur_ct[xcons_ct[i]], xcons_ct[cur_ct[i]], cur_ct[i]); }
nconflictsA[a]++;
/* indi bp i:cur_ct[i] conflicts with xcons_ct[cur_ct[i]]:cur_ct[i] */
removebp[cur_ct[i]] = TRUE;
removebp[xcons_ct[cur_ct[i]]] = TRUE;
}
else nconsistentA[a]++;
}
}
if(nconflictsA[a] > lmax) {
if(lfp != NULL) fprintf(lfp, "%s\n", msa->sqname[a]);
nconflicts_list += nconflictsA[a];
nlist++;
}
nbpsA[a] = nconflictsA[a] + nconsistentA[a];
nconflicts_total += nconflictsA[a];
nconsistent_total += nconsistentA[a];
noverlaps_total += noverlapsA[a];
nbps_total += nbpsA[a];
if(do_info && have_cons) printf(" %-*s %6d %6d %6d %6d\n", namewidth, msa->sqname[a], nbpsA[a], noverlapsA[a], nconsistentA[a], nconflictsA[a]);
if(do_info && !have_cons) printf(" %-*s %6d\n", namewidth, msa->sqname[a], nbpsA[a]);
if(nbpsA[a] > nbpsA[max_nbps_aidx]) max_nbps_aidx = a;
if((noverlapsA[a] > noverlapsA[max_noverlaps_aidx]) || ((noverlapsA[a] == noverlapsA[max_noverlaps_aidx]) && (nbpsA[a] > nbpsA[max_noverlaps_aidx]))) max_noverlaps_aidx = a;
if((nconsistentA[a] > nconsistentA[max_nconsistent_aidx]) || ((nconsistentA[a] == nconsistentA[max_nconsistent_aidx]) && (nbpsA[a] > nbpsA[max_nconsistent_aidx]))) max_nconsistent_aidx = a;
}
else if(do_newcons || esl_opt_GetBoolean(go, "-a")) { esl_fatal("No SS annotation for sequence %d, aln %d.\n", (a+1), nali); }
}
if(do_info && have_cons) {
if(nindi_read > 0) printf("\n");
printf(" %-*s %6d %6d %6d %6d\n", namewidth, "SS_cons(consensus)", ncons_bps, ncons_bps, ncons_bps, 0);
if(nindi_read > 0) {
printf("\n# %6d/%6d (%.3f) overlap\n", noverlaps_total, nbps_total, nbps_total > 0 ? (float) noverlaps_total / (float) nbps_total : 0.);
printf("# %6d/%6d (%.3f) consistent\n", nconsistent_total, nbps_total, nbps_total > 0 ? (float) nconsistent_total / (float) nbps_total: 0.);
printf("# %6d/%6d (%.3f) conflict\n", nconflicts_total, nbps_total, nbps_total > 0 ? (float) nconflicts_total / (float) nbps_total: 0.);
}
else {
printf("# No sequences in the alignment have GR SS annotation.\n");
}
}
if(lfp != NULL) {
printf("# %d/%d sequences with %.3f individual bps on avg that conflict with SS_cons written to %s\n", nlist, msa->nseq, (float) nconflicts_list / (float) nlist, esl_opt_GetString(go, "-l"));
}
/* determine number of gaps per alignment column */
if((status = get_gaps_per_column(msa, &ngaps)) != eslOK) goto ERROR;
/* -x: determine max bp structure OR
* -a: list all conflicts in individual structures */
if(do_max || do_a) {
for(i = 1; i <= msa->alen; i++) {
if(nmates_l2r[i] > 1) {/* list the conflicts */
has_conflict[i] = TRUE;
for(j = 1; j <= msa->alen; j++) {
if(bp[i][j] > 0) {
if(do_a) printf("More than 1 right mates for left mate %4d %4d:%4d bp exists in %4d/%4d seqs (%.3f)\n", i, i, j, bp[i][j], msa->nseq - ngaps[i], (float) bp[i][j] / (float) (msa->nseq - ngaps[i]));
has_conflict[j] = TRUE;
}
}
}
}
for(i = 1; i <= msa->alen; i++) {
if(nmates_r2l[i] > 1) {/* list the conflicts */
has_conflict[i] = TRUE;
for(j = 1; j <= msa->alen; j++) {
if(bp[j][i] > 0) {
if(do_a) printf("More than 1 left mates for right mate %4d %4d:%4d bp exists in %4d/%4d seqs (%.3f)\n", i, j, i, bp[j][i], msa->nseq - ngaps[i], (float) bp[j][i] / (float) (msa->nseq - ngaps[i]));
has_conflict[j] = TRUE;
}
}
}
}
for(i = 1; i <= msa->alen; i++) {
/*printf("conflict[%4d]: %d\n", i, has_conflict[i]);*/
if(nmates_l2r[i] == 1 && (!(has_conflict[i]))) {
j = i+1;
while(bp[i][j] == 0) j++;
cons_ct[i] = j;
cons_ct[j] = i;
}
}
/* remove pseudoknotted bps greedily */
for(i = 1; i <= msa->alen; i++) {
j = cons_ct[i];
if(j != 0 && i < j) {
for(i2 = i+1; i2 <= msa->alen; i2++) {
j2 = cons_ct[i2];
if(j2 != 0 && i2 < j2) {
if((i2 < j) && (j < j2)) {
/*printf("KNOT %4d:%4d (%4d) %4d:%4d (%4d)\n", i, j, bp[i][j], i2, j2, bp[i2][j2]);*/
/* note: remove both if they have equal number of sequences */
if(bp[i][j] <= bp[i2][j2]) {
/*printf("rm %4d:%4d\n", i, j);*/
cons_ct[cons_ct[i]] = 0;
cons_ct[i] = 0;
}
if(bp[i][j] >= bp[i2][j2]) {
/*printf("rm %4d:%4d\n", i2, j2);*/
cons_ct[cons_ct[i2]] = 0;
cons_ct[i2] = 0;
}
}
}
}
}
}
}
/***************************************/
/*PARANOID, second check for knots
for(i = 1; i <= msa->alen; i++) {
j = cons_ct[i];
if(j != 0 && i < j) {
printf("BP: %4d:%4d\n", i, j);
for(i2 = 1; i2 <= msa->alen; i2++) {
j2 = cons_ct[i2];
if(j2 != 0 && i2 < j2) {
if((i2 < j) && (j < j2)) {
if((i < i2)) {
printf("KNOT %4d:%4d (%4d) %4d:%4d (%4d)\n", i, j, bp[i][j], i2, j2, bp[i2][j2]);
}
}
}
}
}
}
******************************************/
/***************************************/
/*PARANOID, check cons_ct for consistency
for(i = 1; i <= msa->alen; i++) {
if(cons_ct[i] != 0) {
if(cons_ct[cons_ct[i]] != i) { printf("ERROR: i: %4d cons_ct[i]: %4d cons_ct[cons_ct[i]]: %4d\n", i, cons_ct[i], cons_ct[cons_ct[i]]); }
}
}
*/
/*PARANOID, write out SS_cons
for(i = 1; i <= msa->alen; i++) {
if(i < cons_ct[i]) printf("<");
else if(cons_ct[i] != 0) { printf(">"); }
else printf(".");
}
printf("\n");
*/
/***************************************/
/* textize alignment */
if((status = esl_msa_Textize(msa)) != eslOK) esl_fatal("ERROR textizing alignment %d\n", nali);
/* --fmin */
if(do_fmin) {
/* define ss_cons */
prev_nbps = -1;
fthresh = 0.99;
inconsistent_flag = pknot_flag = FALSE;
printf("# Defining consensus structure:\n");
printf("# indi SS basepair aln columns i:j (from at least 1 indi SS) will become consensus basepair\n");
printf("# if > <x> individual SS contain i:j as a pair\n");
printf("# We'll search for minimal <x> that gives a consistent consensus structure.\n");
printf("# A consistent structure has each position involved in 0 or 1 basepairs.\n");
printf("#\n");
printf("# Alignment file: %s\n", alifile);
printf("# Alignment idx: %d\n", nali);
printf("# Number of seqs: %d\n", msa->nseq);
printf("#\n");
printf("# %5s %23s %6s\n", "<x>", "nseq-required-with-bp", "numbps");
printf("# %5s %23s %6s\n", "-----", "-----------------------", "------");
while(fthresh >= 0.00 && (inconsistent_flag == FALSE) && (pknot_flag == FALSE)) {
nbps = 0;
seqthresh = (int) (fthresh * msa->nseq);
/*printf("fthresh: %f seqthresh: %d nseq: %d\n", fthresh, seqthresh, msa->nseq);*/
esl_vec_ISet(cons_ct, msa->alen+1, 0);
for(i = 1; i <= msa->alen; i++) {
for(j = i+1; j <= msa->alen; j++) {
if(bp[i][j] > seqthresh) {
if(cons_ct[i] != 0 || cons_ct[j] != 0) {
inconsistent_flag = TRUE;
}
/* check for pseudoknots */
for(k = i+1; k < j; k++) {
l = cons_ct[k];
if((k < l) && (l > j)) {
pknot_flag = TRUE;
}
if((k > l) && (l != 0) && (l < i)) {
pknot_flag = TRUE;
}
}
cons_ct[i] = j;
cons_ct[j] = i;
nbps++;
}
}
}
if(inconsistent_flag)
printf(" %.3f %23d %s\n", fthresh, seqthresh+1, "inconsistent");
else if(pknot_flag)
printf(" %.3f %23d %s\n", fthresh, seqthresh+1, "pseudoknotted");
else {
if(nbps != prev_nbps) {
printf(" %.3f %23d %6d\n", fthresh, seqthresh+1, nbps);
}
fmin = fthresh;
}
fthresh -= 0.01;
prev_nbps = nbps;
}
fthresh = fmin;
esl_vec_ISet(cons_ct, msa->alen+1, 0);
}
/* --ffreq: determine structure by defining consensus bps that occur in <x> fraction of indi structures */
if(do_ffreq || do_fmin) {
if(do_fmin) { printf("#\n# <x> determined to be %.3f\n", fthresh); }
if(do_ffreq) {
printf("# Defining consensus structure:\n");
printf("# indi SS basepair aln columns i:j (from at least 1 indi SS) will become consensus basepair\n");
printf("# if > %f individual SS contain i:j as a pair\n", fthresh);
}
esl_vec_ISet(cons_ct, msa->alen+1, 0);
/* define ss_cons */
seqthresh = (int) (fthresh * msa->nseq);
/*printf("fthresh: %f seqthresh: %d nseq: %d\n", fthresh, seqthresh, msa->nseq);*/
for(i = 1; i <= msa->alen; i++) {
for(j = i+1; j <= msa->alen; j++) {
if(bp[i][j] > seqthresh) {
if(cons_ct[i] != 0) {
esl_fatal("ERROR, two base pairs including position %d satisfy threshold (%d:%d and %d:%d)!\n", i, i, cons_ct[i], i, j);
}
if(cons_ct[j] != 0) {
esl_fatal("ERROR, two base pairs including position %d satisfy threshold (%d:%d and %d:%d)!\n", j, j, cons_ct[j], i, j);
}
cons_ct[i] = j;
cons_ct[j] = i;
}
}
}
}
/* -r: redefine consensus struct by removing any bps that conflict with individual structures */
if(do_remove_bps) {
for(i = 1; i <= msa->alen; i++) {
if(!(removebp[i])) {
cons_ct[i] = xcons_ct[i];
cons_ct[cons_ct[i]] = i;
}
else {
printf("# Removing consensus bp: %d:%d\n", i, xcons_ct[i]);
cons_ct[xcons_ct[i]] = 0;
cons_ct[i] = 0;
}
}
}
/* -c: define consensus structure as indi sequence with highest number of consistent bps with structure OR */
/* --indi: define consensus structure as indi sequence <x> from --indi <x> */
if(do_consistent || do_indi2cons) {
if(do_indi2cons) {
indi = esl_opt_GetString(go, "--indi");
for(a = 0; a < msa->nseq; a++) {
if(strcmp(indi, msa->sqname[a]) == 0) break;
}
if(a == msa->nseq) esl_fatal("ERROR, could not find a sequence named %s in the alignment.\n", indi);
}
else { /* do_consistent */
a = max_nconsistent_aidx;
}
if(msa->ss == NULL || msa->ss[a] == NULL) esl_fatal("ERROR, no individual SS annotation for %s in the alignment.\n", msa->sqname[a]);
if((status = esl_wuss2ct(msa->ss[a], msa->alen, cons_ct)) != eslOK) {
esl_fatal("Second pass... SS annotation for sequence %d, aln %d is invalid.\n", (a), nali);
}
printf("# Defined new SS_cons as SS annotation for %s (%d basepairs)\n", msa->sqname[a], nbpsA[a]);
if(esl_opt_GetBoolean(go, "--rfc") || esl_opt_GetBoolean(go, "--rfindi")) {
if(msa->rf != NULL) { free(msa->rf); msa->rf = NULL; }
if((status = esl_strcat(&(msa->rf), -1, msa->aseq[a], msa->alen)) != eslOK) goto ERROR;
printf("# Defined new RF as %s sequence\n", msa->sqname[a]);
}
}
/* write out alignment with new SS_cons */
if(do_newcons) {
if((status = esl_ct2wuss(cons_ct, msa->alen, sscons)) != eslOK) goto ERROR;
if(msa->ss_cons != NULL) { free(msa->ss_cons); msa->ss_cons = NULL; }
if((status = esl_strcat(&(msa->ss_cons), -1, sscons, msa->alen)) != eslOK) goto ERROR;
status = eslx_msafile_Write(ofp, msa, (esl_opt_GetBoolean(go, "--pfam") ? eslMSAFILE_PFAM : eslMSAFILE_STOCKHOLM));
if (status == eslEMEM) esl_fatal("Memory error when outputting alignment\n");
else if (status != eslOK) esl_fatal("Writing alignment file failed with error %d\n", status);
}
free(sscons);
free(cur_ct);
free(cons_ct);
free(xcons_ct);
for(i = 1; i <= msa->alen; i++) free(bp[i]);
free(bp);
esl_msa_Destroy(msa);
}
if (nali == 0) esl_fatal("No alignments found in file %s\n", alifile);
/* Cleanup, normal return
*/
if(lfp != NULL) fclose(lfp);
if(ofp != NULL) {
printf("# Alignment(s) saved to file %s\n", esl_opt_GetString(go, "-o"));
fclose(ofp);
}
eslx_msafile_Close(afp);
esl_getopts_Destroy(go);
return 0;
ERROR:
if(afp) eslx_msafile_Close(afp);
if(go) esl_getopts_Destroy(go);
if(msa) esl_msa_Destroy(msa);
if(lfp) fclose(lfp);
if(ofp) fclose(ofp);
esl_fatal("ERROR\n");
return 1;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *alifile = NULL; /* alignment file name */
int infmt = eslMSAFILE_UNKNOWN; /* format code for alifile */
int outfmt = eslMSAFILE_UNKNOWN; /* output format for fetched msa's */
ESLX_MSAFILE *afp = NULL; /* open alignment file */
FILE *ofp = NULL; /* output stream for alignments */
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) < 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
if (esl_opt_IsOn(go, "--informat")) {
infmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslMSAFILE_UNKNOWN) esl_fatal("%s is not a valid input alignment file format for --informat", esl_opt_GetString(go, "--informat"));
}
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) esl_fatal("%s is not a valid output alignment file format for --outformat", esl_opt_GetString(go, "--outformat"));
alifile = esl_opt_GetArg(go, 1);
/* Open the alignment file. */
if ( (status = eslx_msafile_Open(NULL, alifile, NULL, infmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
/* Open the SSI index, if any */
if (! esl_opt_GetBoolean(go, "--index"))
{
if (afp->bf->mode_is == eslBUFFER_FILE ||
afp->bf->mode_is == eslBUFFER_ALLFILE ||
afp->bf->mode_is == eslBUFFER_MMAP)
{
char *ssifile = NULL;
esl_sprintf(&ssifile, "%s.ssi", afp->bf->filename);
status = esl_ssi_Open(ssifile, &(afp->ssi));
if (status == eslERANGE ) esl_fatal("SSI index %s has 64-bit offsets; this system doesn't support them", ssifile);
else if (status == eslEFORMAT) esl_fatal("SSI index %s has an unrecognized format. Try recreating, w/ esl-afetch --index", ssifile);
else if (status == eslENOTFOUND) afp->ssi = NULL;
else if (status != eslOK) esl_fatal("SSI index %s: open failed, error code %d\n", ssifile, status);
free(ssifile);
}
}
/* Open the output file, if any
*/
if (esl_opt_GetBoolean(go, "-O"))
{
if ((ofp = fopen(esl_opt_GetArg(go, 2), "w")) == NULL)
esl_fatal("Failed to open output file %s\n", esl_opt_GetArg(go, 2));
}
else if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
esl_fatal("Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Hand off control flow as appropriate */
if (esl_opt_GetBoolean(go, "--index"))
{
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
create_ssi_index(go, afp);
}
else if (esl_opt_GetBoolean(go, "-f"))
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
multifetch(go, ofp, outfmt, esl_opt_GetArg(go, 2), afp);
}
else
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
onefetch(go, ofp, outfmt, esl_opt_GetArg(go, 2), afp);
if (ofp != stdout) printf("\n\nRetrieved alignment %s.\n", esl_opt_GetArg(go, 2));
}
eslx_msafile_Close(afp);
esl_getopts_Destroy(go);
exit(0);
}
int
main(int argc, char **argv)
{
int i,j;
ESL_GETOPTS *go = NULL; /* command line processing */
ESL_STOPWATCH *w = esl_stopwatch_Create();
int status;
ESL_MSA *msa = NULL;
FILE *ofp = NULL; /* output file (default is stdout) */
ESL_ALPHABET *abc = NULL; /* digital alphabet */
char *alifile; /* name of the alignment file we're building HMMs from */
ESLX_MSAFILE *afp = NULL; /* open alifile */
int infmt = eslMSAFILE_UNKNOWN; /* autodetect alignment format by default. */
int outfmt = eslMSAFILE_STOCKHOLM;
char *postmsafile; /* optional file to resave annotated, modified MSAs to */
FILE *postmsafp = NULL; /* open <postmsafile>, or NULL */
int mask_range_cnt = 0;
uint32_t mask_starts[100]; // over-the-top allocation.
uint32_t mask_ends[100];
char *rangestr;
char *range;
int *map = NULL; /* map[i]=j, means model position i comes from column j of the alignment; 1..alen */
int keep_mm;
/* Set processor specific flags */
impl_Init();
alifile = NULL;
postmsafile = NULL;
/* Parse the command line
*/
process_commandline(argc, argv, &go, &alifile, &postmsafile);
keep_mm = esl_opt_IsUsed(go, "--apendmask");
/* Initialize what we can in the config structure (without knowing the alphabet yet).
* Fields controlled by masters are set up in usual_master() or mpi_master()
* Fields used by workers are set up in mpi_worker()
*/
ofp = NULL;
infmt = eslMSAFILE_UNKNOWN;
afp = NULL;
abc = NULL;
if (esl_opt_IsOn(go, "--informat")) {
infmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslMSAFILE_UNKNOWN) p7_Fail("%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--informat"));
}
/* Determine output alignment file format */
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) p7_Fail(argv[0], "%s is not a recognized output MSA file format\n", esl_opt_GetString(go, "--outformat"));
/* Parse the ranges */
if (esl_opt_IsUsed(go, "--alirange")) {
esl_strdup(esl_opt_GetString(go, "--alirange"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--modelrange")) {
esl_strdup(esl_opt_GetString(go, "--modelrange"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--model2ali")) {
esl_strdup(esl_opt_GetString(go, "--model2ali"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--ali2model")) {
esl_strdup(esl_opt_GetString(go, "--ali2model"), -1, &rangestr) ;
} else {
if (puts("Must specify mask range with --modelrange, --alirange, --model2ali, or --ali2model\n") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto ERROR;
}
while ( (status = esl_strtok(&rangestr, ",", &range) ) == eslOK) {
status = esl_regexp_ParseCoordString(range, mask_starts + mask_range_cnt, mask_ends + mask_range_cnt );
if (status == eslESYNTAX) esl_fatal("range flags take coords <from>..<to>; %s not recognized", range);
if (status == eslFAIL) esl_fatal("Failed to find <from> or <to> coord in %s", range);
mask_range_cnt++;
}
/* Start timing. */
esl_stopwatch_Start(w);
/* Open files, set alphabet.
* afp - open alignment file for input
* abc - alphabet expected or guessed in ali file
* postmsafp - open MSA output file
* ofp - optional open output file, or stdout
*/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else abc = NULL;
status = eslx_msafile_Open(&abc, alifile, NULL, infmt, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
postmsafp = fopen(postmsafile, "w");
if (postmsafp == NULL) p7_Fail("Failed to MSA output file %s for writing", postmsafile);
}
if (esl_opt_IsUsed(go, "-o"))
{
ofp = fopen(esl_opt_GetString(go, "-o"), "w");
if (ofp == NULL) p7_Fail("Failed to open -o output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Looks like the i/o is set up successfully...
* Initial output to the user
*/
output_header(go, ofp, alifile, postmsafile); /* cheery output header */
/* read the alignment */
if ((status = eslx_msafile_Read(afp, &msa)) != eslOK) eslx_msafile_ReadFailure(afp, status);
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
/* add/modify mmline for the mask */
if (msa->mm == NULL) {
ESL_ALLOC(msa->mm, msa->alen);
keep_mm = FALSE;
}
if (!keep_mm)
for (i=0; i<msa->alen; i++) msa->mm[i] = '.';
}
// convert model coordinates to alignment coordinates, if necessary
if (esl_opt_IsUsed(go, "--modelrange") || esl_opt_IsUsed(go, "--model2ali") || esl_opt_IsUsed(go, "--ali2model") ) {
float symfrac = esl_opt_GetReal(go, "--symfrac");
int do_hand = esl_opt_IsOn(go, "--hand");
int L;
//same as p7_builder relative_weights
if (esl_opt_IsOn(go, "--wnone") ) { esl_vec_DSet(msa->wgt, msa->nseq, 1.); }
else if (esl_opt_IsOn(go, "--wgiven") ) ;
else if (esl_opt_IsOn(go, "--wpb") ) status = esl_msaweight_PB(msa);
else if (esl_opt_IsOn(go, "--wgsc") ) status = esl_msaweight_GSC(msa);
else if (esl_opt_IsOn(go, "--wblosum")) status = esl_msaweight_BLOSUM(msa, esl_opt_GetReal(go, "--wid"));
if ((status = esl_msa_MarkFragments(msa, esl_opt_GetReal(go, "--fragthresh"))) != eslOK) goto ERROR;
//build a map of model mask coordinates to alignment coords
ESL_ALLOC(map, sizeof(int) * (msa->alen+1));
L = p7_Alimask_MakeModel2AliMap(msa, do_hand, symfrac, map );
if ( esl_opt_IsUsed(go, "--model2ali") ) {
//print mapping
printf ("model coordinates alignment coordinates\n");
for (i=0; i<mask_range_cnt; i++)
printf ("%8d..%-8d -> %8d..%-8d\n", mask_starts[i], mask_ends[i], map[mask_starts[i]-1], map[mask_ends[i]-1]);
/* If I wanted to, I could print all the map values independently:
printf("\n\n-----------\n");
printf("Map\n");
printf("---\n");
for (i=0; i<L; i++)
printf("%d -> %d\n", i+1, map[i]);
*/
} else if ( esl_opt_IsUsed(go, "--ali2model") ) {
//print mapping (requires scanning the inverted map
int alistart = 0;
int aliend = 0;
printf ("alignment coordinates model coordinates\n");
for (i=0; i<mask_range_cnt; i++) {
//find j for ali positions
while (map[alistart] < mask_starts[i] )
alistart++;
aliend = alistart;
while (map[aliend] < mask_ends[i] )
aliend++;
printf (" %8d..%-8d -> %8d..%-8d\n", map[alistart], map[aliend], alistart+1, aliend+1);
}
} else {
//convert the mask coords based on map
for (i=0; i<mask_range_cnt; i++) {
mask_starts[i] = map[mask_starts[i]-1]; //-1 because mmline is offset by one relative to the 1-base alignment
mask_ends[i] = map[mask_ends[i]-1];
}
}
}
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
//overwrite '.' with 'm' everywhere the range says to do it
for (i=0; i<mask_range_cnt; i++)
for (j=mask_starts[i]; j<=mask_ends[i]; j++)
msa->mm[j-1] = 'm';
if ((status = eslx_msafile_Write(postmsafp, msa, outfmt)) != eslOK) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed");
}
esl_stopwatch_Stop(w);
if (esl_opt_IsOn(go, "-o")) fclose(ofp);
if (postmsafp) fclose(postmsafp);
if (afp) eslx_msafile_Close(afp);
if (abc) esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
esl_stopwatch_Destroy(w);
return 0;
ERROR:
return eslFAIL;
}
/* regurgitate_pfam_as_afa()
*
* Given an open Pfam formatted msafile, read the next alignment and
* regurgitate it in aligned FASTA (AFA) format without storing
* it in a esl_msa data structure.
*
* We need to do two passes through the file because in Pfam
* sequence accessions (#=GS <seqname> AC) and sequence descriptions
* (#=GS <seqname> DE) appear altogether before any aligned sequence
* data, while in AFA they appear on the same line as the sequence
* name (accession, then description).
*
* Example:
* # STOCKHOLM 1.0
* #=GS tRNA1 AC RF00005-1
* #=GS tRNA2 AC RF00005-2
* #=GS tRNA1 DE first tRNA
* #=GS tRNA2 DE second tRNA
*
* tRNA1 GCGGAUUUAGCUCAGUUGGG.AGAGCGCCAGACUGAAGAUCUGGAGGUCCUGUGUUCGAUCCACAGAAUUCGCA
* tRNA2 UCCGAUAUAGUGUAAC.GGCUAUCACAUCACGCUUUCACCGUGGAGA.CCGGGGUUCGACUCCCCGUAUCGGAG
*
* converts to AFA:
* >tRNA1 RF00005-1 first tRNA
* GCGGAUUUAGCUCAGUUGGG.AGAGCGCCAGACUGAAGAUCUGGAGGUCCUGUGUUCGAU
* CCACAGAAUUCGCA
* >tRNA2 RF00005-2 second tRNA
* UCCGAUAUAGUGUAAC.GGCUAUCACAUCACGCUUUCACCGUGGAGA.CCGGGGUUCGAC
* UCCCCGUAUCGGAG
*
* In the first pass, output the sequence names and accessions we find
* as '#=GS <seqname> AC' lines in the Pfam alignment to an accession
* tmpfile, and output sequence names and descriptions we find as
* as '#=GS <seqname> DE' lines in the Pfam alignment to a description
* tmpfile.
*
* In the second pass, rewind all (up to 3) files: <ac_tmpfile>,
* <de_tmpfile> and the Pfam alignment file and start reading them
* again. As we're reading them, output the accessions, descriptions
* and aligned sequence data in the proper order to an aligned FASTA
* file.
*
* Set <ret_reached_eof> as TRUE if the alignment read and reformatted
* appears to be the only one remaining in afp. Set <ret_reached_eof>
* as FALSE if afp appears to include at least one more alignment.
*
* Returns void. Dies upon any input error.
*/
static void
regurgitate_pfam_as_afa(ESLX_MSAFILE *afp, FILE *ofp, char *alifile, char *gapsym, int force_lower, int force_upper, int force_rna, int force_dna, int iupac_to_n, int x_is_bad, char *rename, char *rfrom, char *rto, int *ret_reached_eof)
{
char *p = NULL;
esl_pos_t n = 0;
esl_pos_t gslen, seqnamelen, taglen;
char *seqname = NULL;
char *first_seqname = NULL;
char *tag = NULL;
char *gs = NULL;
int nseq_read = 0;
int reached_eof;
/* variables related to reading accessions */
char ac_tmpfile[16] = "esltmpXXXXXX";
FILE *ac_fp = NULL; /* file ptr for accession tmpfile */
char *ac_buf = NULL; /* buffer for line input w/ sre_fgets() */
int ac_buflen = 0; /* current allocated length for buf */
char *ac_s = NULL;
char *ac_seqname = NULL;
char *ac = NULL;
int have_ac = FALSE;
/* variables related to reading descriptions */
char de_tmpfile[16] = "esltmpXXXXXX";
FILE *de_fp = NULL; /* file ptr for description tmpfile */
char *de_buf = NULL; /* buffer for line input w/ sre_fgets() */
int de_buflen = 0; /* current allocated length for buf */
char *de_s = NULL;
char *de_seqname = NULL;
char *de = NULL;
int have_de = FALSE;
/* variables related to printing out sequences */
char *aseq = NULL;
esl_pos_t aseqlen = 0;
int64_t apos;
char aseqbuf[61];
int cpl = 60; /* number of residues per afa seq line */
int acpl; /* actual number of character per line */
int status;
afp->errmsg[0] = '\0';
/**************************************************************************************************
* First pass, go through each line of the Pfam file and output all GS DE and AC annotation to tmpfiles
**************************************************************************************************/
/* Check the magic Stockholm header line, allowing blank lines */
do {
status = eslx_msafile_GetLine(afp, &p, &n);
if (status == eslEOF) return;
else if (status != eslOK) esl_fatal("small mem parse error. problem reading line %d of msafile", (int) afp->linenumber);
} while (esl_memspn(afp->line, afp->n, " \t") == afp->n || /* skip blank lines */
(esl_memstrpfx(afp->line, afp->n, "#") /* and skip comment lines */
&& ! esl_memstrpfx(afp->line, afp->n, "# STOCKHOLM"))); /* but stop on Stockholm header */
if (! esl_memstrpfx(afp->line, afp->n, "# STOCKHOLM 1.")) esl_fatal("small mem parse failed (line %d): missing \"# STOCKHOLM\" header", (int) afp->linenumber);
while ((status = eslx_msafile_GetLine(afp, &p, &n)) == eslOK)
{
while (n && ( *p == ' ' || *p == '\t')) { p++; n--; } /* skip leading whitespace */
if (esl_memstrpfx(p, n, "#=GS"))
{ /* only lines we need to check are AC and DE lines, we don't even check other lines for validity */
if (esl_memtok(&p, &n, " \t", &gs, &gslen) != eslOK) esl_fatal("small mem parse failed (line %d) in a way that can't happen", (int) afp->linenumber);
if (esl_memtok(&p, &n, " \t", &seqname, &seqnamelen) != eslOK) esl_fatal("small mem parse failed (line %d): #=GS line missing <seqname>, <tag>, annotation", (int) afp->linenumber);
if (esl_memtok(&p, &n, " \t", &tag, &taglen) != eslOK) esl_fatal("small mem parse failed (line %d): #=GS line missing <tag>, annotation", (int) afp->linenumber);
if (! esl_memstrcmp(gs, gslen, "#=GS")) esl_fatal("small mem parse failed (line %d): faux #=GS line?", (int) afp->linenumber);
if (esl_memstrcmp(tag, taglen, "AC"))
{
if (! ac_fp && esl_tmpfile(ac_tmpfile, &ac_fp) != eslOK) esl_fatal("small mem parse failed, unable to open accession tmpfile");
fprintf(ac_fp, "%.*s %.*s\n", (int) seqnamelen, seqname, (int) n, p);
}
if (esl_memstrcmp(tag, taglen, "DE"))
{
if (! de_fp && esl_tmpfile(de_tmpfile, &de_fp) != eslOK) esl_fatal("small mem parse failed, unable to open description tmpfile");
fprintf(de_fp, "%.*s %.*s\n", (int) seqnamelen, seqname, (int) n, p);
}
}
else if (esl_memstrpfx(p, n, "//")) break;
}
if (status == eslEOF) esl_fatal("small mem parse failed (line %d): missing // terminator", (int) afp->linenumber);
else if (status != eslOK) esl_fatal("small mem parse failed (line %d) with code %d", (int) afp->linenumber, status);
/* The regurgitate_*() functions are limited, and only deal with single-record Pfam files.
* If there appears to be more data in the file, drop the reached_eof flag.
*/
while ((status = eslx_msafile_GetLine(afp, &p, &n)) == eslOK)
{
while (n && ( *p == ' ' || *p == '\t')) { p++; n--; } /* skip leading whitespace */
if (esl_memstrpfx(p, n, "# STOCKHOLM 1.")) break;
if (n && ! esl_memstrpfx(p, n, "#")) esl_fatal("small mem parse failed (line %d): unexpected data", (int) afp->linenumber);
}
if (status == eslOK) reached_eof = FALSE;
else if (status == eslEOF) reached_eof = TRUE;
else esl_fatal("--small parse error. problem reading line %d of msafile", (int) afp->linenumber);
/*****************************************************************
* Pass 1 complete; rewind (close/reopen) all files
*****************************************************************/
eslx_msafile_Close(afp);
if ((status = eslx_msafile_Open(NULL, alifile, NULL, eslMSAFILE_PFAM, NULL, &afp)) != eslOK)
esl_fatal("--small, second pass, unable to open file %s for reading", alifile);
if (ac_fp) { /* open the tmpfile with the seq accessions */
rewind(ac_fp);
if((status = esl_fgets(&(ac_buf), &(ac_buflen), ac_fp)) != eslOK) esl_fatal("--small accession tmpfile parse failed");
ac_s = ac_buf;
if (esl_strtok_adv(&ac_s, " \t\n\r", &ac_seqname, NULL, NULL) != eslOK) esl_fatal("--small accession tmpfile parse failed");
if (esl_strtok_adv(&ac_s, "\n\r", &ac, NULL, NULL) != eslOK) esl_fatal("--small accession tmpfile parse failed");
}
if (de_fp) { /* open the tmpfile with the seq descriptions */
rewind(de_fp);
if((status = esl_fgets(&(de_buf), &(de_buflen), de_fp)) != eslOK) esl_fatal("--small description tmpfile parse failed");
de_s = de_buf;
if (esl_strtok_adv(&de_s, " \t\n\r", &de_seqname, NULL, NULL) != eslOK) esl_fatal("--small description tmpfile parse failed");
if (esl_strtok_adv(&de_s, "\n\r", &de, NULL, NULL) != eslOK) esl_fatal("--small description tmpfile parse failed");
}
/******************************************************************************************
* Pass 2, step through files, outputting appropriately
******************************************************************************************/
do {
status = eslx_msafile_GetLine(afp, &p, &n);
if (status == eslEOF) return;
else if (status != eslOK) esl_fatal("small mem parse pass 2 error. problem reading line %d of msafile", (int) afp->linenumber);
} while (esl_memspn(afp->line, afp->n, " \t") == afp->n || /* skip blank lines */
(esl_memstrpfx(afp->line, afp->n, "#") /* and skip comment lines */
&& ! esl_memstrpfx(afp->line, afp->n, "# STOCKHOLM"))); /* but stop on Stockholm header */
if (! esl_memstrpfx(afp->line, afp->n, "# STOCKHOLM 1.")) esl_fatal("small mem parse pass 2 failed (line %d): missing \"# STOCKHOLM\" header", (int) afp->linenumber);
while ((status = eslx_msafile_GetLine(afp, &p, &n)) == eslOK)
{
while (n && ( *p == ' ' || *p == '\t')) { p++; n--; } /* skip leading whitespace */
if (!n || *p == '#') continue; /* skip blank lines, comments */
else if (esl_memstrpfx(p, n, "//")) break; /* end of alignment: end of record */
else
{ /* sequence line. parse line into temporary strings */
if (esl_memtok(&p, &n, " \t", &seqname, &seqnamelen) != eslOK) esl_fatal("small mem parse pass 2 failed (line %d): no seq name", (int) afp->linenumber);
if (esl_memtok(&p, &n, " \t", &aseq, &aseqlen) != eslOK) esl_fatal("small mem parse pass 2 failed (line %d): no aseq", (int) afp->linenumber);
/* make sure we haven't just read a second line of the first sequence in file (we must be in Pfam 1 line/seq file) */
if (nseq_read == 0) { if ((status = esl_memstrdup(seqname, seqnamelen, &(first_seqname))) != eslOK) esl_fatal("small mem parse failed: unable to copy seqname"); }
else if (esl_memstrcmp(seqname, seqnamelen, first_seqname)) esl_fatal("--small parse pass 2 failed (line %d): two seqs named %s. Alignment appears to be in interleaved Stockholm (not Pfam) format.", (int) afp->linenumber, seqname);
nseq_read++;
/* determine if we have an accession and/or description for this sequence */
have_de = have_ac = FALSE;
if (ac_seqname && (esl_memstrcmp(seqname, seqnamelen, ac_seqname))) have_ac = TRUE;
if (de_seqname && (esl_memstrcmp(seqname, seqnamelen, de_seqname))) have_de = TRUE;
if (rename) fprintf(ofp, ">%s.%d%s%s%s%s\n", rename, nseq_read, (have_ac ? " " : "") , (have_ac ? ac : ""), (have_de ? " " : "") , (have_de ? de : ""));
else fprintf(ofp, ">%.*s%s%s%s%s\n", (int) seqnamelen, seqname, (have_ac ? " " : "") , (have_ac ? ac : ""), (have_de ? " " : "") , (have_de ? de : ""));
/* load next ac, de */
if (have_ac) {
status = esl_fgets(&(ac_buf), &(ac_buflen), ac_fp);
if (status == eslEOF) ac_seqname = NULL;
else if (status == eslOK) {
ac_s = ac_buf;
if (esl_strtok_adv(&ac_s, " \t\n\r", &ac_seqname, NULL, NULL) != eslOK) esl_fatal("--small accession tmpfile parse failed");
if (esl_strtok_adv(&ac_s, "\n\r", &ac, NULL, NULL) != eslOK) esl_fatal("--small accession tmpfile parse failed");
}
}
if (have_de) {
status = esl_fgets(&(de_buf), &(de_buflen), de_fp);
if(status == eslEOF) de_seqname = NULL;
else if (status == eslOK) {
de_s = de_buf;
if (esl_strtok_adv(&de_s, " \t\n\r", &de_seqname, NULL, NULL) != eslOK) esl_fatal("--small description tmpfile parse failed");
if (esl_strtok_adv(&de_s, "\n\r", &de, NULL, NULL) != eslOK) esl_fatal("--small description tmpfile parse failed");
}
}
/* now print sequence, after converting symbols as nec */
/* remember, aseq itself is part of an ESL_BUFFER and you
can't write to it, so symconverts have to be on the
copy */
for (apos = 0; apos < aseqlen; apos += cpl)
{
acpl = (aseqlen - apos > cpl ? cpl : aseqlen - apos);
strncpy(aseqbuf, aseq + apos, acpl);
aseqbuf[acpl] = '\0';
if (rfrom) symconvert(aseqbuf, rfrom, rto);
if (gapsym) symconvert(aseqbuf, "-_.", gapsym);
if (force_lower) symconvert(aseqbuf,
"ABCDEFGHIJKLMNOPQRSTUVWXYZ",
"abcdefghijklmnopqrstuvwxyz");
if (force_upper) symconvert(aseqbuf,
"abcdefghijklmnopqrstuvwxyz",
"ABCDEFGHIJKLMNOPQRSTUVWXYZ");
if (force_rna) symconvert(aseqbuf, "Tt", "Uu");
if (force_dna) symconvert(aseqbuf, "Uu", "Tt");
if (iupac_to_n) symconvert(aseqbuf,
"RYMKSWHBVDrymkswhbvd",
"NNNNNNNNNNnnnnnnnnnn");
if (x_is_bad) symconvert(aseqbuf, "Xx", "Nn");
fprintf(ofp, "%s\n", aseqbuf);
}
}
}
/* If we saw a normal // end, we would've successfully read a line,
* so when we get here, status (from the line read) should be eslOK.
*/
if (status != eslOK) esl_fatal("--small parse pass 2 failed (line %d): didn't find // at end of alignment", (int) afp->linenumber);
if (ac_seqname) esl_fatal("--small parse pass 2 failed, sequence %s with #=GS AC line does not exist in alignment or is in different order.", ac_seqname);
if (de_seqname) esl_fatal("--small parse pass 2 failed, sequence %s with #=GS DE line does not exist in alignment or is in different order.", de_seqname);
if (ac_fp) fclose(ac_fp);
if (de_fp) fclose(de_fp);
eslx_msafile_Close(afp);
if (first_seqname) free(first_seqname);
if (ac_buf) free(ac_buf);
if (de_buf) free(de_buf);
*ret_reached_eof = reached_eof;
return;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go;
char *msafile;
ESLX_MSAFILE *afp;
ESL_MSA *msa;
int do_gsc;
int do_pb;
int do_blosum;
int maxN;
double maxid;
int nsmall, nbig;
int i;
int status;
/* Process command line */
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) esl_fatal("%s", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) esl_fatal("%s", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") == TRUE){
puts(usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, 0, 2, 80); /* 0=all docgroups; 2=indentation; 80=width */
return 0;
}
do_blosum = esl_opt_GetBoolean(go, "--blosum");
do_gsc = esl_opt_GetBoolean(go, "--gsc");
do_pb = esl_opt_GetBoolean(go, "--pb");
maxid = esl_opt_GetReal (go, "--id");
maxN = esl_opt_GetInteger(go, "--maxN");
if (esl_opt_ArgNumber(go) != 1) {
puts("Incorrect number of command line arguments.");
puts(usage);
return 1;
}
if ((msafile = esl_opt_GetArg(go, 1)) == NULL) esl_fatal("%s", go->errbuf);
esl_getopts_Destroy(go);
/* Weight one or more alignments from input file
*/
if ((status = eslx_msafile_Open(NULL, msafile, NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
while ( (status = eslx_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
if (maxN > 0 && msa->nseq > maxN) { esl_msa_Destroy(msa); continue; }
if (do_gsc) esl_msaweight_GSC(msa);
else if (do_pb) esl_msaweight_PB(msa);
else if (do_blosum) esl_msaweight_BLOSUM(msa, maxid);
for (nsmall = 0, nbig = 0, i = 0; i < msa->nseq; i++) {
if (msa->wgt[i] < 0.2) nsmall++;
if (msa->wgt[i] > 5.0) nbig++;
}
printf("%-20s %5d %5d %8.4f %8.4f %5d %5d\n",
msa->name,
msa->nseq,
msa->alen,
esl_vec_DMin(msa->wgt, msa->nseq),
esl_vec_DMax(msa->wgt, msa->nseq),
nsmall,
nbig);
esl_msa_Destroy(msa);
}
eslx_msafile_Close(afp);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_STOPWATCH *w;
ESL_GETOPTS *go;
char *msafile;
ESLX_MSAFILE *afp;
ESL_MSA *msa;
int do_gsc;
int do_pb;
int do_blosum;
int maxN;
double maxid;
double cpu;
int status;
/* Process command line
*/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) esl_fatal("failed to parse cmd line: %s", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) esl_fatal("failed to parse cmd line: %s", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") == TRUE) {
puts(usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, 0, 2, 80); /* 0=all docgroups; 2=indentation; 80=width */
return 0;
}
do_blosum = esl_opt_GetBoolean(go, "--blosum");
do_gsc = esl_opt_GetBoolean(go, "--gsc");
do_pb = esl_opt_GetBoolean(go, "--pb");
maxid = esl_opt_GetReal (go, "--id");
maxN = esl_opt_GetInteger(go, "--maxN");
if (esl_opt_ArgNumber(go) != 1) {
puts("Incorrect number of command line arguments.");
puts(usage);
return 1;
}
if ((msafile = esl_opt_GetArg(go, 1)) == NULL) esl_fatal("failed to parse cmd line: %s", go->errbuf);
esl_getopts_Destroy(go);
w = esl_stopwatch_Create();
/* Weight one or more alignments from input file
*/
if ((status = eslx_msafile_Open(NULL, msafile, NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
while ( (status = eslx_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
if (maxN > 0 && msa->nseq > maxN) { esl_msa_Destroy(msa); continue; }
esl_stopwatch_Start(w);
if (do_gsc) esl_msaweight_GSC(msa);
else if (do_pb) esl_msaweight_PB(msa);
else if (do_blosum) esl_msaweight_BLOSUM(msa, maxid);
esl_stopwatch_Stop(w);
cpu = w->user;
printf("%-20s %6d %6d %.3f\n", msa->name, msa->alen, msa->nseq, cpu);
esl_msa_Destroy(msa);
}
eslx_msafile_Close(afp);
esl_stopwatch_Destroy(w);
return eslOK;
}
static int
map_alignment(const char *msafile, const P7_HMM *hmm, ESL_SQ ***ret_sq, P7_TRACE ***ret_tr, int *ret_ntot)
{
ESL_SQ **sq = NULL;
P7_TRACE **tr = NULL;
ESLX_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
ESL_ALPHABET *abc = (ESL_ALPHABET *) hmm->abc; /* removing const'ness to make compiler happy. Safe. */
int *matassign = NULL;
uint32_t chksum = 0;
int i,k;
int status;
status = eslx_msafile_Open(&abc, msafile, NULL, eslMSAFILE_UNKNOWN, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
status = eslx_msafile_Read(afp, &msa);
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
if (! (hmm->flags & p7H_CHKSUM) ) esl_fatal("HMM has no checksum. --mapali unreliable without it.");
if (! (hmm->flags & p7H_MAP) ) esl_fatal("HMM has no map. --mapali can't work without it.");
esl_msa_Checksum(msa, &chksum);
if (hmm->checksum != chksum) esl_fatal("--mapali MSA %s isn't same as the one HMM came from (checksum mismatch)", msafile);
ESL_ALLOC(sq, sizeof(ESL_SQ *) * msa->nseq);
ESL_ALLOC(tr, sizeof(P7_TRACE *) * msa->nseq);
ESL_ALLOC(matassign, sizeof(int) * (msa->alen + 1));
esl_vec_ISet(matassign, msa->alen+1, 0);
for (k = 1; k <= hmm->M; k++) matassign[hmm->map[k]] = 1;
p7_trace_FauxFromMSA(msa, matassign, p7_DEFAULT, tr);
/* The 'faux' core traces constructed by FauxFromMSA() may contain
* D->I and I->D transitions. They may *only* now be passed to
* p7_tracealign_Seqs(), which can deal with these 'illegal'
* transitions, in order to exactly reproduce the input --mapali
* alignment.
*/
for (i = 0; i < msa->nseq; i++)
esl_sq_FetchFromMSA(msa, i, &(sq[i]));
*ret_ntot = msa->nseq;
*ret_tr = tr;
*ret_sq = sq;
eslx_msafile_Close(afp);
esl_msa_Destroy(msa);
free(matassign);
return eslOK;
ERROR:
*ret_ntot = 0;
*ret_tr = NULL;
*ret_sq = NULL;
if (afp != NULL) eslx_msafile_Close(afp);
if (msa != NULL) esl_msa_Destroy(msa);
if (matassign != NULL) free(matassign);
return status;
}
SV* isaMSA (const char *input, int is_msa, int ali_hmm, int dna_ok, int frag){
ESLX_MSAFILE *mfp = NULL;
ESL_MSA *msa = NULL;
ESL_MSA *msa_clone = NULL;
ESL_ALPHABET *abc = NULL;
P7_HMM *ret_hmm = NULL;
char *ascii_hmm = NULL;
int alpha;
int status;
char errbuf[eslERRBUFSIZE];
HV* hash = newHV();
hv_store(hash, "type", strlen("type"), newSVpv("UNK", 3), 0);
if ((status = eslx_msafile_OpenMem( &abc, input, -1, NULL, NULL, &mfp)) != eslOK){
hv_store(hash, "error", strlen("error"), newSViv(status), 0);
}else{
status = eslx_msafile_Read(mfp, &msa);
if(status != eslOK){
hv_store(hash, "error", strlen("error"), newSVpv(mfp->errmsg, strlen(mfp->errmsg)), 0);
hv_store(hash, "guess", strlen("guess"), newSViv(mfp->format), 0);
hv_store(hash, "type", strlen("type"), newSVpv("MSA", 3), 0);
hv_store(hash, "position", strlen("position"), newSViv(mfp->linenumber), 0);
}else{
esl_msa_Textize(msa);
esl_msa_GuessAlphabet(msa, &alpha);
hv_store(hash, "alpha", strlen("alpha"), newSViv(alpha), 0);
if(alpha == eslAMINO || (dna_ok == 1 && (alpha == eslDNA || alpha == eslRNA))){
abc = esl_alphabet_Create(alpha);
esl_msa_Digitize( abc, msa, errbuf);
hv_store(hash, "count", strlen("count"), newSViv(msa->nseq), 0);
if(msa->nseq == 1 && mfp->format == eslMSAFILE_AFA){
hv_store(hash, "type", strlen("type"), newSVpv("SS", 2), 0);
hv_store(hash, "hmmpgmd", strlen("hmmpgmd"), newSVpv(input, strlen(input)), 0);
}else{
// We have been told it is an MSA or it is any other format other than AFA
hv_store(hash, "type", strlen("type"), newSVpv("MSA", 3), 0);
status = constructHMM( msa, abc, ali_hmm, frag, &ret_hmm, errbuf);
if (status != eslOK) {
hv_store(hash, "error", strlen("error"), newSVpv(errbuf, strlen(errbuf)), 0);
hv_store(hash, "guess", strlen("guess"), newSViv(mfp->format), 0);
} else {
p7_hmmfile_WriteToString(&ascii_hmm, -1, ret_hmm);
hv_store(hash, "hmmpgmd", strlen("hmmpgmd"), newSVpv(ascii_hmm, strlen(ascii_hmm)), 0);
}
}
}else{
if(alpha == 0 ){
hv_store(hash, "error", strlen("error"), newSVpv("Could not determine alphabet", 28), 0);
}else{
hv_store(hash, "error", strlen("error"), newSVpv("Bad alphabet", 12), 0);
}
if (msa->nseq > 1) {
if( mfp->format == eslMSAFILE_AFA){
hv_store(hash, "type", strlen("type"), newSVpv("MS?", 3), 0);
}else{
hv_store(hash, "type", strlen("type"), newSVpv("MSA", 3), 0);
}
}
}
}
}
if (mfp != NULL) eslx_msafile_Close(mfp);
if (abc != NULL) esl_alphabet_Destroy(abc);
if (msa != NULL) esl_msa_Destroy(msa);
if (ret_hmm != NULL) p7_hmm_Destroy(ret_hmm);
if (ascii_hmm != NULL ) free( ascii_hmm );
return newRV_noinc((SV*) hash);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
ESL_ALPHABET *abc = NULL; /* biological alphabet */
char *alifile1= NULL; /* alignment 1 file name */
char *alifile2= NULL; /* alignment 2 file name */
int fmt; /* format code for alifiles */
ESLX_MSAFILE *afp1 = NULL; /* open alignment file 1 */
ESLX_MSAFILE *afp2 = NULL; /* open alignment file 2 */
ESL_MSA *msa1 = NULL; /* multiple sequence alignment 1 */
ESL_MSA *msa2 = NULL; /* multiple sequence alignment 2 */
int status; /* easel return code */
char errbuf[eslERRBUFSIZE*4];
int *msa1_to_msa2_map; /* map from <msafile1> to <msafile2> */
char *sub_msa1_to_msa2_mask; /* with --sub the map from <msafile1> to <msafile2> in mask form */
FILE *subfp = NULL;
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK ||
esl_opt_VerifyConfig(go) != eslOK)
{
printf("Failed to parse command line: %s\n", go->errbuf);
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
if (esl_opt_GetBoolean(go, "-h") )
{
esl_banner(stdout, argv[0], banner);
esl_usage (stdout, argv[0], usage);
puts("\nwhere basic options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
exit(0);
}
if (esl_opt_ArgNumber(go) != 2)
{
printf("Incorrect number of command line arguments.\n");
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
alifile1 = esl_opt_GetArg(go, 1);
alifile2 = esl_opt_GetArg(go, 2);
fmt = eslMSAFILE_STOCKHOLM;
/***********************************************
* Open the MSA files
***********************************************/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
if ( (status = eslx_msafile_Open(&abc, alifile1, NULL, fmt, NULL, &afp1)) != eslOK) eslx_msafile_OpenFailure(afp1, status);
if ( (status = eslx_msafile_Open(&abc, alifile2, NULL, fmt, NULL, &afp2)) != eslOK) eslx_msafile_OpenFailure(afp2, status);
/******************************************************************
* Read first alignment from each file, we only use the first one
******************************************************************/
if ((status = eslx_msafile_Read(afp1, &msa1)) != eslOK) eslx_msafile_ReadFailure(afp1, status);
if ((status = eslx_msafile_Read(afp2, &msa2)) != eslOK) eslx_msafile_ReadFailure(afp2, status);
/* map the alignments in msa1 and msa2 */
if(! esl_opt_IsOn(go, "--submap")) {
if((status = map_msas(go, errbuf, msa1, msa2, &msa1_to_msa2_map)) != eslOK) goto ERROR;
free(msa1_to_msa2_map);
}
/* --submap: if nec, map <msafile1> to a subset of it's own columns in <msafile2> */
else { /* --submap was enabled */
if ((subfp = fopen(esl_opt_GetString(go, "--submap"), "w")) == NULL)
ESL_FAIL(eslFAIL, errbuf, "Failed to open --submap output file %s\n", esl_opt_GetString(go, "--submap"));
if((status = map_sub_msas(go, errbuf, msa1, msa2, &sub_msa1_to_msa2_mask)) != eslOK) goto ERROR;
fprintf(subfp, "%s\n", sub_msa1_to_msa2_mask);
fclose(subfp);
subfp = NULL;
printf("# Mask of 1/0s with 1 indicating aln column in %s maps to a column in %s saved to file %s.\n", alifile1, alifile2, esl_opt_GetString(go, "--submap"));
free(sub_msa1_to_msa2_mask);
}
/* Cleanup, normal return
*/
eslx_msafile_Close(afp1);
eslx_msafile_Close(afp2);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
esl_msa_Destroy(msa1);
esl_msa_Destroy(msa2);
return 0;
ERROR:
if (afp1) eslx_msafile_Close(afp1);
if (afp2) eslx_msafile_Close(afp2);
if (go) esl_getopts_Destroy(go);
if (msa1) esl_msa_Destroy(msa1);
if (msa2) esl_msa_Destroy(msa2);
if (subfp) fclose(subfp);
esl_fatal(errbuf);
return 1; /* never reached */
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
ESL_ALPHABET *abc = NULL; /* biological alphabet */
char *alifile = NULL; /* alignment file name */
int fmt = eslMSAFILE_UNKNOWN; /* format code for alifile */
ESLX_MSAFILE *afp = NULL; /* open msa file */
ESL_MSAFILE2 *old_afp = NULL; /* open msa file, legacy (--small) */
ESL_MSA *msa = NULL; /* one multiple sequence alignment */
int nali; /* number of alignments read */
int i; /* counter over seqs */
int64_t alen; /* alignment length */
int nseq; /* number of sequences in the msa */
int64_t rlen; /* a raw (unaligned) seq length */
int64_t small, large; /* smallest, largest sequence */
int64_t nres; /* total # of residues in msa */
double avgid; /* average fractional pair id */
int max_comparisons; /* maximum # comparisons for avg id */
int do_stall; /* used to stall when debugging */
double **abc_ct = NULL; /* [0..msa->alen-1][0..abc->K] number of each residue at each position (abc->K is gap) */
double ***bp_ct = NULL; /* [0..msa->alen-1][0..abc->Kp-1][0..abc->Kp-1] per (non-pknotted) consensus basepair *
* count of each possible basepair over all seqs basepairs are indexed by 'i' the minimum *
* of 'i:j' for a pair between i and j, where i < j. */
double **pp_ct = NULL; /* [0..msa->alen-1][0..11], count of each posterior probability (PP) code, over all sequences, gap is 11 */
int *i_am_rf = NULL; /* [0..i..msa->alen-1]: TRUE if pos i is non-gap RF posn, if msa->rf == NULL remains NULL */
int *rf2a_map = NULL; /* [0..rfpos..rflen-1] = apos,
* apos is the alignment position (0..msa->alen-1) that
* is non-gap RF position rfpos+1 (for rfpos in 0..rflen-1) */
int rflen = -1; /* nongap RF length */
char errbuf[eslERRBUFSIZE];
int status; /* easel return code */
/* optional output files */
FILE *iinfofp = NULL; /* output file for --iinfo */
FILE *pcinfofp = NULL; /* output file for --pcinfo */
FILE *psinfofp = NULL; /* output file for --psinfo */
FILE *rinfofp = NULL; /* output file for --rinfo */
FILE *icinfofp = NULL; /* output file for --icinfo */
FILE *listfp = NULL; /* output file for --list */
FILE *cinfofp = NULL; /* output file for --cinfo */
FILE *bpinfofp = NULL; /* output file for --bpinfo */
int use_weights; /* TRUE if --weight, reported weighted counts (using msa->wgt) to all output files */
int weights_exist; /* TRUE if at least one msa->wgt value differs from 1.0, FALSE if not (or if msa->wgt==NULL) */
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK ||
esl_opt_VerifyConfig(go) != eslOK)
{
printf("Failed to parse command line: %s\n", go->errbuf);
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
if (esl_opt_GetBoolean(go, "-h") )
{
esl_banner(stdout, argv[0], banner);
esl_usage (stdout, argv[0], usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\n small memory mode, requires --amino,--dna, or --rna and --informat pfam:");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\n optional output files:");
esl_opt_DisplayHelp(stdout, go, 3, 2, 80);
exit(0);
}
if (esl_opt_ArgNumber(go) != 1)
{
printf("Incorrect number of command line arguments.\n");
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
alifile = esl_opt_GetArg(go, 1);
if (esl_opt_IsOn(go, "--informat") &&
(fmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"))) == eslMSAFILE_UNKNOWN)
esl_fatal("%s is not a valid input sequence file format for --informat", esl_opt_GetString(go, "--informat"));
if (esl_opt_GetBoolean(go, "--small") && fmt != eslMSAFILE_PFAM) esl_fatal("--small requires --informat pfam\n");
max_comparisons = 1000;
do_stall = esl_opt_GetBoolean(go, "--stall"); /* a stall point for attaching gdb */
while (do_stall);
/***********************************************
* Open the MSA file; determine alphabet; set for digital input
***********************************************/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
/* We'd like to get rid of the legacy msafile interface, but it
* includes small memory functionality for Pfam format which we have
* to replace first. For now, use both interfaces, new and legacy
*/
if ( esl_opt_GetBoolean(go, "--small") )
{
if (! abc) esl_fatal("--small requires one of --amino, --dna, --rna be specified.");
status = esl_msafile2_OpenDigital(abc, alifile, NULL, &old_afp);
if (status == eslENOTFOUND) esl_fatal("Alignment file %s doesn't exist or is not readable\n", alifile);
else if (status == eslEFORMAT) esl_fatal("Couldn't determine format of alignment %s\n", alifile);
else if (status != eslOK) esl_fatal("Alignment file open failed with error %d\n", status);
}
else
{
if ( (status = eslx_msafile_Open(&abc, alifile, NULL, fmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
}
/**************************************
* Open optional output files, as nec *
**************************************/
/* determine name for first list file, if nec */
if( esl_opt_IsOn(go, "--list")) {
if ((listfp = fopen(esl_opt_GetString(go, "--list"), "w")) == NULL)
esl_fatal("Failed to open --list output file %s\n", esl_opt_GetString(go, "--list"));
}
if( esl_opt_IsOn(go, "--icinfo")) {
if ((icinfofp = fopen(esl_opt_GetString(go, "--icinfo"), "w")) == NULL)
esl_fatal("Failed to open --icinfo output file %s\n", esl_opt_GetString(go, "--icinfo"));
}
if( esl_opt_IsOn(go, "--rinfo")) {
if ((rinfofp = fopen(esl_opt_GetString(go, "--rinfo"), "w")) == NULL)
esl_fatal("Failed to open --rinfo output file %s\n", esl_opt_GetString(go, "--rinfo"));
}
if( esl_opt_IsOn(go, "--pcinfo")) {
if ((pcinfofp = fopen(esl_opt_GetString(go, "--pcinfo"), "w")) == NULL)
esl_fatal("Failed to open --pcinfo output file %s\n", esl_opt_GetString(go, "--pcinfo"));
}
if( esl_opt_IsOn(go, "--psinfo")) {
if ((psinfofp = fopen(esl_opt_GetString(go, "--psinfo"), "w")) == NULL)
esl_fatal("Failed to open --psinfo output file %s\n", esl_opt_GetString(go, "--psinfo"));
}
if( esl_opt_IsOn(go, "--iinfo")) {
if ((iinfofp = fopen(esl_opt_GetString(go, "--iinfo"), "w")) == NULL)
esl_fatal("Failed to open --iinfo output file %s\n", esl_opt_GetString(go, "--iinfo"));
}
if( esl_opt_IsOn(go, "--cinfo")) {
if ((cinfofp = fopen(esl_opt_GetString(go, "--cinfo"), "w")) == NULL)
esl_fatal("Failed to open --cinfo output file %s\n", esl_opt_GetString(go, "--cinfo"));
}
if( esl_opt_IsOn(go, "--bpinfo")) {
if ((bpinfofp = fopen(esl_opt_GetString(go, "--bpinfo"), "w")) == NULL)
esl_fatal("Failed to open --bpinfo output file %s\n", esl_opt_GetString(go, "--bpinfo"));
}
/***********************************************
* Read MSAs one at a time.
***********************************************/
if (esl_opt_GetBoolean(go, "-1")) {
puts("#");
if(! esl_opt_GetBoolean(go, "--small")) {
printf("# %-4s %-20s %10s %7s %7s %12s %6s %6s %10s %3s\n", "idx", "name", "format", "nseq", "alen", "nres", "small", "large", "avlen", "%id");
printf("# %-4s %-20s %10s %7s %7s %12s %6s %6s %10s %3s\n", "----", "--------------------", "----------", "-------", "-------", "------------", "------", "------", "----------", "---");
}
else {
printf("# %-4s %-20s %10s %7s %7s %12s %10s\n", "idx", "name", "format", "nseq", "alen", "nres", "avlen");
printf("# %-4s %-20s %10s %7s %7s %12s %10s\n", "----", "--------------------", "----------", "-------", "-------", "------------", "----------");
}
}
nali = 0;
fmt = (esl_opt_GetBoolean(go, "--small") ? old_afp->format : afp->format);
while ( (status = ( esl_opt_GetBoolean(go, "--small") ?
esl_msafile2_ReadInfoPfam(old_afp, listfp, abc, -1, NULL, NULL, &msa, &nseq, &alen, NULL, NULL, NULL, NULL, NULL, &abc_ct, &pp_ct, NULL, NULL, NULL) :
eslx_msafile_Read (afp, &msa))) == eslOK)
{
nali++;
nres = 0;
if (! esl_opt_GetBoolean(go, "--small")) {
nseq = msa->nseq;
alen = msa->alen;
small = large = -1;
for (i = 0; i < msa->nseq; i++)
{
rlen = esl_abc_dsqrlen(msa->abc, msa->ax[i]);
nres += rlen;
if (small == -1 || rlen < small) small = rlen;
if (large == -1 || rlen > large) large = rlen;
}
esl_dst_XAverageId(abc, msa->ax, msa->nseq, max_comparisons, &avgid);
}
else { /* --small invoked */
for(i = 0; i < alen; i++) nres += (int) esl_vec_DSum(abc_ct[i], abc->K);
}
if (esl_opt_GetBoolean(go, "-1"))
{
printf("%-6d %-20s %10s %7d %7" PRId64 " %12" PRId64,
nali,
msa->name,
eslx_msafile_DecodeFormat(fmt),
nseq,
alen,
nres);
if (! esl_opt_GetBoolean(go, "--small")) {
printf(" %6" PRId64 " %6" PRId64 " %10.1f %3.0f\n",
small,
large,
(double) nres / (double) msa->nseq,
100.*avgid);
}
else {
printf(" %10.1f\n", (double) nres / (double) nseq);
}
}
else
{
printf("Alignment number: %d\n", nali);
if (msa->name != NULL)
printf("Alignment name: %s\n", msa->name);
printf("Format: %s\n", eslx_msafile_DecodeFormat(fmt));
printf("Number of sequences: %d\n", nseq);
printf("Alignment length: %" PRId64 "\n", alen);
printf("Total # residues: %" PRId64 "\n", nres);
if(! esl_opt_GetBoolean(go, "--small")) {
printf("Smallest: %" PRId64 "\n", small);
printf("Largest: %" PRId64 "\n", large);
}
printf("Average length: %.1f\n", (double) nres / (double) nseq);
if(! esl_opt_GetBoolean(go, "--small")) {
printf("Average identity: %.0f%%\n", 100.*avgid);
}
printf("//\n");
}
/* Dump data to optional output files, if nec */
if(esl_opt_IsOn(go, "--list")) {
if(! esl_opt_GetBoolean(go, "--small")) {
/* only print sequence name to list file if ! --small, else we already have in esl_msafile2_ReadInfoPfam() */
for(i = 0; i < msa->nseq; i++) fprintf(listfp, "%s\n", msa->sqname[i]);
}
}
/* if RF exists, get i_am_rf array[0..alen] which tells us which positions are non-gap RF positions
* and rf2a_map, a map of non-gap RF positions to overall alignment positions */
if(msa->rf != NULL) {
if((status = map_rfpos_to_apos(msa, abc, errbuf, alen, &i_am_rf, &rf2a_map, &rflen)) != eslOK) esl_fatal(errbuf);
}
else i_am_rf = NULL;
weights_exist = check_msa_weights(msa);
use_weights = (weights_exist && esl_opt_GetBoolean(go, "--weight")) ? TRUE : FALSE;
if( (! esl_opt_GetBoolean(go, "--small")) &&
(esl_opt_IsOn(go, "--icinfo") || esl_opt_IsOn(go, "--rinfo") || esl_opt_IsOn(go, "--pcinfo") ||
esl_opt_IsOn(go, "--cinfo") || esl_opt_IsOn(go, "--bpinfo"))) {
/* collect counts of each residue and PPs (if they exist) from the msa */
if(esl_opt_GetBoolean(go, "--weight") && msa->wgt == NULL) esl_fatal("--weight requires all alignments have #=GS WT annotation, but aln %d does not", nali);
if((status = count_msa(msa, errbuf, nali,
esl_opt_GetBoolean(go, "--noambig"), /* ignore ambiguous residues? */
esl_opt_GetBoolean(go, "--weight"), /* use msa->wgt sequence weights? */
&abc_ct,
((bpinfofp != NULL && msa->ss_cons != NULL) ? &bp_ct : NULL), /* get basepair counts? */
(msa->pp != NULL ? &pp_ct : NULL))) /* get PP counts? */
!= eslOK) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--icinfo")) {
if((status = dump_infocontent_info(icinfofp, abc, abc_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--rinfo")) {
if((status = dump_residue_info(rinfofp, abc, abc_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if(esl_opt_IsOn(go, "--pcinfo")) {
if(pp_ct == NULL) esl_fatal("Error: --pcinfo requires all alignments have #=GR PP annotation, but alignment %d does not", nali);
if((status = dump_posterior_column_info(pcinfofp, pp_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if(esl_opt_IsOn(go, "--psinfo")) {
if(msa->pp == NULL) esl_fatal("Error: --psinfo requires all alignments have #=GR PP annotation, but alignment %d does not", nali);
if((status = dump_posterior_sequence_info(psinfofp, msa, nali, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--iinfo")) {
if(msa->rf == NULL) esl_fatal("--iinfo requires all alignments have #=GC RF annotation, but alignment %d does not", nali);
if(esl_opt_GetBoolean(go, "--weight") && msa->wgt == NULL) esl_fatal("--weight requires all alignments have #=GS WT annotation, but aln %d does not", nali);
if((status = dump_insert_info(iinfofp, msa, use_weights, nali, i_am_rf, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--cinfo")) {
if((status = dump_column_residue_counts(cinfofp, abc, abc_ct, esl_opt_GetBoolean(go, "--noambig"), use_weights, nali, alen, nseq, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--bpinfo")) {
if(msa->ss_cons == NULL) esl_fatal("--bpinfo requires all alignments have #=GC SS_cons annotation, but alignment %d does not", nali);
if((status = dump_basepair_counts(bpinfofp, msa, abc, bp_ct, use_weights, nali, nseq, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
esl_msa_Destroy(msa);
if(abc_ct != NULL) { esl_Free2D((void **) abc_ct, alen); abc_ct = NULL; }
if(bp_ct != NULL) { esl_Free3D((void ***) bp_ct, alen, abc->Kp); bp_ct = NULL; }
if(pp_ct != NULL) { esl_Free2D((void **) pp_ct, alen); pp_ct = NULL; }
if(i_am_rf != NULL) { free(i_am_rf); i_am_rf = NULL; }
if(rf2a_map != NULL) { free(rf2a_map); rf2a_map = NULL; }
}
/* If an msa read failed, we've dropped out to here with an informative status code.
* we have to handle failures from new vs. legacy msa parsing differently
*/
if (esl_opt_GetBoolean(go, "--small"))
{
if (status == eslEFORMAT) esl_fatal("Alignment file parse error, line %d of file %s:\n%s\nOffending line is:\n%s\n", old_afp->linenumber, old_afp->fname, old_afp->errbuf, old_afp->buf);
else if (status != eslEOF) esl_fatal("Alignment file read failed with error code %d\n", status);
else if (nali == 0) esl_fatal("No alignments found in file %s\n", alifile);
}
else
{
if (nali == 0 || status != eslEOF) eslx_msafile_ReadFailure(afp, status);
}
/* Cleanup, normal return
*/
if(listfp != NULL) {
fclose(listfp);
printf("# List of sequences in %d alignment(s) saved to file %s\n", nali, esl_opt_GetString(go, "--list"));
}
if(icinfofp != NULL) {
fclose(icinfofp);
printf("# Information content data saved to file %s.\n", esl_opt_GetString(go, "--icinfo"));
}
if(rinfofp != NULL) {
fclose(rinfofp);
printf("# Residue data saved to file %s.\n", esl_opt_GetString(go, "--rinfo"));
}
if(pcinfofp != NULL) {
fclose(pcinfofp);
printf("# Per-column posterior probability data saved to file %s.\n", esl_opt_GetString(go, "--pcinfo"));
}
if(psinfofp != NULL) {
fclose(psinfofp);
printf("# Per-sequence posterior probability data saved to file %s.\n", esl_opt_GetString(go, "--psinfo"));
}
if(iinfofp != NULL) {
printf("# Insert data saved to file %s.\n", esl_opt_GetString(go, "--iinfo"));
fclose(iinfofp);
}
if(cinfofp != NULL) {
printf("# Per-column counts data saved to file %s.\n", esl_opt_GetString(go, "--cinfo"));
fclose(cinfofp);
}
if(bpinfofp != NULL) {
printf("# Per-column basepair counts data saved to file %s.\n", esl_opt_GetString(go, "--bpinfo"));
fclose(bpinfofp);
}
if (afp) eslx_msafile_Close(afp);
if (old_afp) esl_msafile2_Close(old_afp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go;
char *msafile;
ESLX_MSAFILE *afp;
ESL_MSA *msa;
float *sqd;
int status;
int nbad;
int nali = 0;
int nbadali = 0;
int nwgt = 0;
int nbadwgt = 0;
int i;
int be_quiet;
int do_gsc;
int do_pb;
int do_blosum;
double maxid;
double tol;
int maxN;
/* Process command line
*/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) esl_fatal("failed to parse cmd line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) esl_fatal("failed to parse cmd line: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") == TRUE) {
puts(usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, 0, 2, 80); /* 0=all docgroups; 2=indentation; 80=width */
return 0;
}
be_quiet = esl_opt_GetBoolean(go, "-q");
do_blosum = esl_opt_GetBoolean(go, "--blosum");
do_gsc = esl_opt_GetBoolean(go, "--gsc");
do_pb = esl_opt_GetBoolean(go, "--pb");
maxid = esl_opt_GetReal (go, "--id");
tol = esl_opt_GetReal (go, "--tol");
maxN = esl_opt_GetInteger(go, "--maxN");
if (esl_opt_ArgNumber(go) != 1) {
puts("Incorrect number of command line arguments.");
puts(usage);
return 1;
}
msafile = esl_opt_GetArg(go, 1);
esl_getopts_Destroy(go);
/* Weight one or more alignments from input file
*/
if ((status = eslx_msafile_Open(NULL, msafile, NULL, eslMSAFILE_UNKNOWN, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
while ( (status = eslx_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
if (maxN > 0 && msa->nseq > maxN) { esl_msa_Destroy(msa); continue; }
nali++;
nwgt += msa->nseq;
ESL_ALLOC(sqd, sizeof(float) * msa->nseq);
if (do_gsc) {
esl_msaweight_GSC(msa);
GSCWeights(msa->aseq, msa->nseq, msa->alen, sqd);
} else if (do_pb) {
esl_msaweight_PB(msa);
PositionBasedWeights(msa->aseq, msa->nseq, msa->alen, sqd);
} else if (do_blosum) {
esl_msaweight_BLOSUM(msa, maxid);
BlosumWeights(msa->aseq, msa->nseq, msa->alen, maxid, sqd);
/* workaround SQUID bug: BLOSUM weights weren't renormalized to sum to nseq. */
esl_vec_FNorm (sqd, msa->nseq);
esl_vec_FScale(sqd, msa->nseq, (float) msa->nseq);
}
if (! be_quiet) {
for (i = 0; i < msa->nseq; i++)
fprintf(stdout, "%-20s %.3f %.3f\n",
msa->sqname[i], msa->wgt[i], sqd[i]);
}
nbad = 0;
for (i = 0; i < msa->nseq; i++)
if (esl_DCompare((double) sqd[i], msa->wgt[i], tol) != eslOK)
nbad++;
if (nbad > 0) nbadali++;
nbadwgt += nbad;
if (nbad > 0) printf("%-20s :: alignment shows %d weights that differ (out of %d) \n",
msa->name, nbad, msa->nseq);
esl_msa_Destroy(msa);
free(sqd);
}
eslx_msafile_Close(afp);
if (nbadali == 0)
printf("OK: all weights identical between squid and Easel in %d alignment(s)\n", nali);
else {
printf("%d of %d weights mismatched at (> %f fractional difference)\n",
nbadwgt, nwgt, tol);
printf("involving %d of %d total alignments\n", nbadali, nali);
}
return eslOK;
ERROR:
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* command line configuration */
struct cfg_s cfg; /* application configuration */
char *basename= NULL; /* base of the output file names */
char *alifile = NULL; /* alignment file name */
char *dbfile = NULL; /* name of seq db file */
char outfile[256]; /* name of an output file */
int alifmt; /* format code for alifile */
int dbfmt; /* format code for dbfile */
ESLX_MSAFILE *afp = NULL; /* open alignment file */
ESL_MSA *origmsa = NULL; /* one multiple sequence alignment */
ESL_MSA *msa = NULL; /* MSA after frags are removed */
ESL_MSA *trainmsa= NULL; /* training set, aligned */
ESL_STACK *teststack=NULL; /* test set: stack of ESL_SQ ptrs */
int status; /* easel return code */
int nfrags; /* # of fragments removed */
int ntestdom; /* # of test domains */
int ntest; /* # of test sequences created */
int nali; /* number of alignments read */
double avgid;
/* 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 app configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h")) cmdline_help(argv[0], go);
if (esl_opt_ArgNumber(go) != 3) cmdline_failure(argv[0], "Incorrect number of command line arguments\n");
basename = esl_opt_GetArg(go, 1);
alifile = esl_opt_GetArg(go, 2);
dbfile = esl_opt_GetArg(go, 3);
alifmt = eslMSAFILE_STOCKHOLM;
dbfmt = eslSQFILE_FASTA;
/* Set up the configuration structure shared amongst functions here */
if (esl_opt_IsDefault(go, "--seed")) cfg.r = esl_randomness_CreateTimeseeded();
else cfg.r = esl_randomness_Create(esl_opt_GetInteger(go, "--seed"));
cfg.abc = NULL; /* until we open the MSA file, below */
cfg.fragfrac = esl_opt_GetReal(go, "-F");
cfg.idthresh1 = esl_opt_GetReal(go, "-1");
cfg.idthresh2 = esl_opt_GetReal(go, "-2");
cfg.test_lens = NULL;
cfg.ntest = 0;
cfg.max_ntest = (esl_opt_IsOn(go, "--maxtest") ? esl_opt_GetInteger(go, "--maxtest") : 0);
cfg.max_ntrain = (esl_opt_IsOn(go, "--maxtrain") ? esl_opt_GetInteger(go, "--maxtrain") : 0);
/* Open the output files */
if (snprintf(outfile, 256, "%s.msa", basename) >= 256) esl_fatal("Failed to construct output MSA file name");
if ((cfg.out_msafp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open MSA output file %s\n", outfile);
if (snprintf(outfile, 256, "%s.fa", basename) >= 256) esl_fatal("Failed to construct output FASTA file name");
if ((cfg.out_seqfp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open FASTA output file %s\n", outfile);
if (snprintf(outfile, 256, "%s.pos", basename) >= 256) esl_fatal("Failed to construct pos test set summary file name");
if ((cfg.possummfp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open pos test set summary file %s\n", outfile);
if (snprintf(outfile, 256, "%s.neg", basename) >= 256) esl_fatal("Failed to construct neg test set summary file name");
if ((cfg.negsummfp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open neg test set summary file %s\n", outfile);
if (snprintf(outfile, 256, "%s.tbl", basename) >= 256) esl_fatal("Failed to construct benchmark table file name");
if ((cfg.tblfp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open benchmark table file %s\n", outfile);
if (esl_opt_GetBoolean(go, "--pid")) {
if (snprintf(outfile, 256, "%s.pid", basename) >= 256) esl_fatal("Failed to construct %%id table file name");
if ((cfg.pidfp = fopen(outfile, "w")) == NULL) esl_fatal("Failed to open %%id table file %s\n", outfile);
} else cfg.pidfp = NULL;
/* Open the MSA file, digital mode; determine alphabet */
if (esl_opt_GetBoolean(go, "--amino")) cfg.abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) cfg.abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) cfg.abc = esl_alphabet_Create(eslRNA);
status = eslx_msafile_Open(&(cfg.abc), alifile, NULL, alifmt, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
if (cfg.abc->type == eslAMINO) esl_composition_SW34(cfg.fq);
else esl_vec_DSet(cfg.fq, cfg.abc->K, 1.0 / (double) cfg.abc->K);
/* Open and process the dbfile; make sure it's in the same alphabet */
process_dbfile(&cfg, dbfile, dbfmt);
/* Read and process MSAs one at a time */
nali = 0;
while ((status = eslx_msafile_Read(afp, &origmsa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
esl_msa_ConvertDegen2X(origmsa);
esl_msa_Hash(origmsa);
remove_fragments(&cfg, origmsa, &msa, &nfrags);
separate_sets (&cfg, msa, &trainmsa, &teststack);
if ( esl_stack_ObjectCount(teststack) >= 2)
{
/* randomize test domain order, and apply size limit if any */
esl_stack_Shuffle(cfg.r, teststack);
if (cfg.max_ntest) pstack_select_topn(&teststack, cfg.max_ntest);
ntestdom = esl_stack_ObjectCount(teststack);
/* randomize training set alignment order, and apply size limit if any */
esl_msashuffle_PermuteSequenceOrder(cfg.r, trainmsa);
if (cfg.max_ntrain) msa_select_topn(&trainmsa, cfg.max_ntrain);
esl_msa_MinimGaps(trainmsa, NULL, NULL, FALSE);
if (esl_opt_GetBoolean(go, "--pid")) write_pids(cfg.pidfp, origmsa, trainmsa, teststack);
synthesize_positives(go, &cfg, msa->name, teststack, &ntest);
eslx_msafile_Write(cfg.out_msafp, trainmsa, eslMSAFILE_STOCKHOLM);
esl_dst_XAverageId(cfg.abc, trainmsa->ax, trainmsa->nseq, 10000, &avgid); /* 10000 is max_comparisons, before sampling kicks in */
fprintf(cfg.tblfp, "%-20s %3.0f%% %6d %6d %6d %6d %6d %6d\n", msa->name, 100.*avgid, (int) trainmsa->alen, msa->nseq, nfrags, trainmsa->nseq, ntestdom, ntest);
nali++;
}
esl_msa_Destroy(trainmsa);
esl_msa_Destroy(origmsa);
esl_msa_Destroy(msa);
}
if (nali == 0) esl_fatal("No alignments found in file %s\n", alifile);
synthesize_negatives(go, &cfg, esl_opt_GetInteger(go, "-N"));
fclose(cfg.out_msafp);
fclose(cfg.out_seqfp);
fclose(cfg.possummfp);
fclose(cfg.negsummfp);
fclose(cfg.tblfp);
if (cfg.pidfp) fclose(cfg.pidfp);
esl_randomness_Destroy(cfg.r);
esl_alphabet_Destroy(cfg.abc);
eslx_msafile_Close(afp);
esl_getopts_Destroy(go);
return 0;
}
