static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage1);
esl_usage (stdout, argv0, usage2);
esl_usage (stdout, argv0, usage3);
puts("\n where general options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\n Options for retrieving subsequences:");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\n On command line, subseq coords are separated by any nonnumeric, nonspace character(s).");
puts(" for example, -c 23..100 or -c 23/100 or -c 23-100 all work.\n");
puts(" Additionally, to retrieve a suffix to the end, omit the end coord or set it to zero; -c 23.. ");
puts(" will work, as will -c 23..0\n");
puts(" By default, the subseq will be named <source name>/<from>-<to>. To assign a name of");
puts(" your choice, use -n <newname>.\n");
puts(" In retrieving subsequences listed in a file (-C -f, or just -Cf), each line of the file");
puts(" is in GDF format: <newname> <from> <to> <source seqname>, space/tab delimited.\n");
puts(" When <start> coordinate is greater than <end>, for DNA or RNA, the reverse complement is");
puts(" retrieved; in protein sequence, this is an error. The -r option is another way to revcomp.");
puts("\n other options:");
esl_opt_DisplayHelp(stdout, go, 3, 2, 80);
exit(0);
}
static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage1);
puts("\n where general options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
exit(0);
}
static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage);
puts("\n options:");
esl_opt_DisplayHelp(stdout, go, 0, 2, 80);
exit(0);
}
static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage);
puts("\n where general options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\n options controlling segment randomization method:");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\n options declaring a particular alphabet:");
esl_opt_DisplayHelp(stdout, go, 3, 2, 80);
puts("\n other options:");
esl_opt_DisplayHelp(stdout, go, 4, 2, 80);
exit(0);
}
static int
process_commandline(int argc, char **argv, ESL_GETOPTS **ret_go, char **ret_fmfile, char **ret_qfile)
{
ESL_GETOPTS *go = esl_getopts_Create(options);
int status;
if (esl_opt_ProcessEnvironment(go) != eslOK) { if (printf("Failed to process environment: %s\n", go->errbuf) < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) { if (printf("Failed to parse command line: %s\n", go->errbuf) < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
if (esl_opt_VerifyConfig(go) != eslOK) { if (printf("Failed to parse command line: %s\n", go->errbuf) < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
/* help format: */
if (esl_opt_GetBoolean(go, "-h") == TRUE)
{
esl_banner(stdout, argv[0], banner);
esl_usage(stdout, argv[0], usage);
if (puts("\nBasic options:") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80); /* 1= group; 2 = indentation; 120=textwidth*/
if (puts("\nSpecial options:") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80); /* 2= group; 2 = indentation; 120=textwidth*/
exit(0);
}
if (esl_opt_ArgNumber(go) != 2) { if (puts("Incorrect number of command line arguments.") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
if ((*ret_qfile = esl_opt_GetArg(go, 1)) == NULL) { if (puts("Failed to get <qfile> argument on command line") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
if ((*ret_fmfile = esl_opt_GetArg(go, 2)) == NULL) { if (puts("Failed to get <fmfile> argument on command line") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
/* Validate any attempted use of stdin streams */
if (esl_strcmp(*ret_fmfile, "-") == 0 && esl_strcmp(*ret_qfile, "-") == 0)
{ if (puts("Either <fmfile> or <qfile> may be '-' (to read from stdin), but not both.") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto FAILURE; }
*ret_go = go;
return eslOK;
FAILURE: /* all errors handled here are user errors, so be polite. */
esl_usage(stdout, argv[0], usage);
puts("\nwhere basic options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80); /* 1= group; 2 = indentation; 80=textwidth*/
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
esl_getopts_Destroy(go);
exit(1);
ERROR:
if (go) esl_getopts_Destroy(go);
exit(status);
}
static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage1);
esl_usage (stdout, argv0, usage2);
esl_usage (stdout, argv0, usage3);
esl_usage (stdout, argv0, usage4);
puts("\n where general options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\n options for shuffling input sequences (default mode):");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\n options for generating sequences de novo (w/ -G option):");
esl_opt_DisplayHelp(stdout, go, 4, 2, 80);
puts("\n other infrequently used options:");
esl_opt_DisplayHelp(stdout, go, 5, 2, 80);
exit(0);
}
static void
cmdline_help(char *argv0, ESL_GETOPTS *go)
{
esl_banner(stdout, argv0, banner);
esl_usage (stdout, argv0, usage);
puts("\n where general options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("");
puts("The <seqfile> is a sequence file in any accepted format, such as FASTA.");
puts("It may be indexed (see esl-sfetch --index) for faster performance.");
puts("");
puts("The <maskfile> is a space-delimited file; each data line with 3 columns:");
puts(" field 1: <seqname> to fetch from <sqfile>");
puts(" field 2: <start> coordinate for mask operation, 1..n");
puts(" field 3: <end> coordinate for mask operation, 1..n");
puts("Lines starting with # are comments, and ignored.)");
puts("");
exit(0);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL;
ESL_ALPHABET *nt_abc = esl_alphabet_Create(eslDNA);
ESL_ALPHABET *aa_abc = esl_alphabet_Create(eslAMINO);
ESL_GENCODE *gcode = NULL;
ESL_GENCODE_WORKSTATE *wrk = NULL;
char *seqfile = NULL;
int informat = eslSQFILE_UNKNOWN;
ESL_SQFILE *sqfp = NULL;
int status;
/*****************************************************************
* command line parsing
*****************************************************************/
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, /*docgroup=*/0, /*indent=*/2, /*textwidth=*/80);
puts("\nAvailable NCBI genetic code tables (for -c <id>):");
esl_gencode_DumpAltCodeTable(stdout);
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);
}
seqfile = esl_opt_GetArg(go, 1);
if (esl_opt_IsOn(go, "--informat") &&
(informat = esl_sqio_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"));
status = esl_sqfile_OpenDigital(nt_abc, seqfile, informat, /*env=*/NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("Failed to find (or open) sequence file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Failed to recognize format of sequence file %s", seqfile);
else if (status != eslOK) esl_fatal("Failure in opening sequence file %s; code %d", seqfile, status);
/* A limitation. The esl_sqio_ReadWindow() interface needs to use SSI positioning
* to read reverse complement, and that doesn't work on nonrewindable streams.
*/
if ( esl_opt_GetBoolean(go, "-W") && ! esl_sqfile_IsRewindable(sqfp) && ! esl_opt_GetBoolean(go, "--watson"))
esl_fatal("esl-translate can't read reverse complement from a nonrewindable stream (stdin pipe, .gz file, etc).");
/* Set up the genetic code. Default = NCBI 1, the standard code; allow ORFs to start at any aa
*/
gcode = esl_gencode_Create(nt_abc, aa_abc);
esl_gencode_Set(gcode, esl_opt_GetInteger(go, "-c")); // default = 1, the standard genetic code
if (esl_opt_GetBoolean(go, "-m")) esl_gencode_SetInitiatorOnlyAUG(gcode);
else if (! esl_opt_GetBoolean(go, "-M")) esl_gencode_SetInitiatorAny(gcode); // note this is the default, if neither -m or -M are set
/* Set up the workstate structure, which contains both stateful
* info about our position in <sqfp> and the DNA <sq>, as well as
* one-time config info from options
*/
wrk = esl_gencode_WorkstateCreate(go, gcode);
/* The two styles of main processing loop:
*/
if (esl_opt_GetBoolean(go, "-W")) do_by_windows(gcode, wrk, sqfp);
else do_by_sequences(gcode, wrk, sqfp);
esl_gencode_WorkstateDestroy(wrk);
esl_sqfile_Close(sqfp);
esl_gencode_Destroy(gcode);
esl_alphabet_Destroy(aa_abc);
esl_alphabet_Destroy(nt_abc);
esl_getopts_Destroy(go);
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 */
ESL_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 = esl_msafile_Open(&abc, alifile, NULL, fmt, NULL, &afp)) != eslOK)
esl_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 = esl_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) esl_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 = esl_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);
}
esl_msafile_Close(afp);
esl_getopts_Destroy(go);
return 0;
ERROR:
if(afp) esl_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; /* application configuration */
int kstatus, tstatus;/* return code from Easel routine */
int fmt; /* expected format of kfile, tfile */
char *kfile, *tfile; /* known, test structure file */
ESL_MSAFILE *kfp, *tfp; /* open kfile, tfile */
ESL_MSA *ka, *ta; /* known, trusted alignment */
int64_t klen, tlen; /* lengths of dealigned seqs */
int i; /* counter over sequences */
int apos; /* counter over alignment columns */
int rfpos; /* counter over consensus (non-gap RF) columns */
int is_rfpos; /* TRUE if current apos is a consensus pos, FALSE if not */
int uapos; /* counter over unaligned residue positions */
int nali; /* number of alignment we're on in each file */
int **kp; /* [0..i..nseq-1][1..r..sq->n] = x known non-gap RF position of residue r in sequence i */
int **tp; /* [0..i..nseq-1][1..r..sq->n] = x predicted non-gap RF position of residue r in sequence i */
/* for both kp and pp, if x <= 0, residue r for seq i is not aligned to a non-gap RF position, but rather as an 'insert'
* after non-gap RF position (x * -1)
*/
int *km_pos; /* [0..rflen] = x, in known aln, number of residues aligned to non-gap RF column x; special case: mct[0] = 0 */
int *ki_pos; /* [0..rflen] = x, in known aln, number of residues inserted after non-gap RF column x */
int *tm_pos; /* [0..rflen] = x, in predicted aln, number of residues aligned to non-gap RF column x; special case: mct[0] = 0 */
int *ti_pos; /* [0..rflen] = x, in predicted aln, number of residues inserted after non-gap RF column x */
int *cor_tm_pos; /* [0..rflen] = x, in predicted aln, number of correctly predicted residues aligned to non-gap RF column x; special case: mct[0] = 0 */
int *cor_ti_pos; /* [0..rflen] = x, in predicted aln, number of correctly predicted residues inserted after non-gap RF column x */
int *km_seq; /* [0..i..nseq-1] = x, in known aln, number of residues aligned to non-gap RF columns in seq i; */
int *ki_seq; /* [0..i..nseq-1] = x, in known aln, number of residues inserted in seq i */
int *tm_seq; /* [0..i..nseq-1] = x, in predicted aln, number of residues aligned to non-gap RF columns in seq i; */
int *ti_seq; /* [0..i..nseq-1] = x, in predicted aln, number of residues inserted in seq i */
int *cor_tm_seq; /* [0..i..nseq-1] = x, in predicted aln, number of correctly predicted residues aligned to non-gap RF columns in seq i */
int *cor_ti_seq; /* [0..i..nseq-1] = x, in predicted aln, number of correctly predicted residues inserted in seq i */
int *seqlen; /* [0..i..nseq-1] = x, unaligned seq i has length x */
ESL_ALPHABET *abc = NULL; /* alphabet for all alignments */
int rflen, t_rflen; /* non-gap RF length (consensus lengths) */
int status;
char *namedashes;
int ni;
int namewidth = 8; /* length of 'seq name' */
int cor_tm, cor_ti, km, ki; /* correct predicted match, correct predicted insert, total match, total insert */
char *mask = NULL;
int masklen;
ESL_DSQ *ks;
ESL_DSQ *ts;
FILE *dfp = NULL; /* for --c2dfile */
/* variables needed for -p and related options */
int do_post = FALSE; /* TRUE if -p enabled */
int do_post_for_this_rfpos = FALSE; /* set for each consensus position, always TRUE unless --mask-p2xm */
int p; /* counter over integerized posteriors */
int *ptm = NULL; /* [0..p..10] number of total matches with posterior value p (10="*")*/
int *pti = NULL; /* [0..p..10] number of total inserts with posterior value p */
int *cor_ptm = NULL; /* [0..p..10] number of correct matches with posterior value p */
int *cor_pti = NULL; /* [0..p..10] number of correct inserts with posterior value p */
int npostvals = 11; /* number of posterior values 0-9, * */
int ppidx; /* index of PP */
char ppchars[11] = "0123456789*";
int cm_cor_ptm, cm_cor_pti, cm_ptm, cm_pti, cm_incor_ptm, cm_incor_pti; /* cumulative counts of posteriors */
// int tot_cor_ptm, tot_cor_pti, tot_ptm, tot_pti; /* total counts of posteriors */
// int tot_incor_ptm,tot_incor_pti; // SRE: commented out; don't seem to be used; need to silence compiler warning
char errbuf[eslERRBUFSIZE];
/***********************************************
* 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);
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
exit(EXIT_SUCCESS);
}
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);
}
kfile = esl_opt_GetArg(go, 1);
tfile = esl_opt_GetArg(go, 2);
fmt = eslMSAFILE_STOCKHOLM;
/***********************************************
* Open the two Stockholm 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 ( (kstatus = esl_msafile_Open(&abc, kfile, NULL, fmt, NULL, &kfp)) != eslOK) esl_msafile_OpenFailure(kfp, kstatus);
if ( (tstatus = esl_msafile_Open(&abc, tfile, NULL, fmt, NULL, &tfp)) != eslOK) esl_msafile_OpenFailure(tfp, tstatus);
do_post = esl_opt_GetBoolean(go, "-p");
/* read the mask file if --p-mask is enabled */
if(! esl_opt_IsDefault(go, "--p-mask")) {
if((status = read_mask_file(esl_opt_GetString(go, "--p-mask"), errbuf, &mask, &masklen)) != eslOK) esl_fatal(errbuf);
}
/* open the c2dfile for output, if nec */
if (esl_opt_IsOn(go, "--c2dfile")) {
if ((dfp = fopen(esl_opt_GetString(go, "--c2dfile"), "w")) == NULL) esl_fatal("Failed to open --c2dfile output file %s\n", esl_opt_GetString(go, "--c2dfile"));
}
/***********************************************
* Do alignment comparisons, one seq at a time;
* this means looping over all seqs in all alignments.
***********************************************/
nali = 0;
while ( (kstatus = esl_msafile_Read(kfp, &ka)) != eslEOF)
{
if ( kstatus != eslOK) esl_msafile_ReadFailure(kfp, kstatus);
if ( (tstatus = esl_msafile_Read(tfp, &ta)) != eslOK) esl_msafile_ReadFailure(tfp, tstatus);
nali++;
if((nali > 1) && (esl_opt_IsOn(go, "--c2dfile"))) esl_fatal("--c2dfile is only meant for msafiles with single alignments");
/* Sanity check on alignment
*/
if (ka->nseq != ta->nseq)
esl_fatal("trusted, test alignments don't have same seq #\n");
if (ka->rf == NULL)
esl_fatal("trusted alignment has no reference annotation\n");
if (ta->rf == NULL)
esl_fatal("test alignment has no reference annotation\n");
/* make sure the sequences are all identical */
ESL_ALLOC(seqlen, sizeof(int) * ka->nseq);
for(i = 0; i < ka->nseq; i++) {
if(strcmp(ka->sqname[i], ta->sqname[i]) != 0) esl_fatal("sequence %d of trusted alignment %s has different name than seq %d of predicted alignment %s\n", (i+1), ka->sqname[i], (i+1), ta->sqname[i]);
ESL_ALLOC(ks, sizeof(ESL_DSQ) * (ka->alen+2));
memcpy(ks, ka->ax[i], (ka->alen+2) * sizeof(ESL_DSQ));
esl_abc_XDealign(ka->abc, ks, ka->ax[i], &klen);
ESL_ALLOC(ts, sizeof(ESL_DSQ) * (ta->alen+2));
memcpy(ts, ta->ax[i], (ta->alen+2) * sizeof(ESL_DSQ));
esl_abc_XDealign(ta->abc, ts, ta->ax[i], &tlen);
if (tlen != klen)
esl_fatal("dealigned sequence mismatch, seq %d, when dealigned, is %d residues in the known alignment, but %d residues in the trusted alignment.", (i+1), klen, tlen);
if (memcmp(ks, ts, sizeof(ESL_DSQ) * klen) != 0)
esl_fatal("dealigned sequence mismatch, seq %d %s, when dealigned, are not identical.", (i+1), ka->sqname[i]);
seqlen[i] = tlen;
free(ks);
free(ts);
}
/* determine non-gap RF length */
rflen = 0;
for(apos = 1; apos <= ka->alen; apos++) {
if((! esl_abc_CIsGap (ka->abc, ka->rf[apos-1])) &&
(! esl_abc_CIsMissing(ka->abc, ka->rf[apos-1]))) rflen++;
}
t_rflen = 0;
for(apos = 1; apos <= ta->alen; apos++) {
if((! esl_abc_CIsGap (ta->abc, ta->rf[apos-1])) &&
(! esl_abc_CIsMissing (ta->abc, ta->rf[apos-1]))) t_rflen++;
}
if(t_rflen != rflen) esl_fatal("Trusted alignment non-gap RF length (%d) != predicted alignment non-gap RF length (%d).\n", rflen, t_rflen);
/* if -p, make sure the test alignment has posterior probabilities, and allocate our counters for correct/incorrect per post value */
if(do_post) {
if(! esl_opt_IsDefault(go, "--p-mask")) {
if(masklen != rflen) {
esl_fatal("Length of mask in %s (%d) not equal to non-gap RF len of alignments (%d)\n", esl_opt_GetString(go, "--p-mask"), masklen, rflen);
}
}
if(ta->pp == NULL) esl_fatal("-p requires \"#=GR PP\" annotation in the test alignment, but none exists");
ESL_ALLOC(ptm, sizeof(int) * npostvals);
ESL_ALLOC(pti, sizeof(int) * npostvals);
ESL_ALLOC(cor_ptm, sizeof(int) * npostvals);
ESL_ALLOC(cor_pti, sizeof(int) * npostvals);
esl_vec_ISet(ptm, npostvals, 0);
esl_vec_ISet(pti, npostvals, 0);
esl_vec_ISet(cor_ptm, npostvals, 0);
esl_vec_ISet(cor_pti, npostvals, 0);
}
/* allocate and initialize our counters */
ESL_ALLOC(kp, sizeof(int *) * ka->nseq);
ESL_ALLOC(tp, sizeof(int *) * ta->nseq);
for(i = 0; i < ka->nseq; i++) {
ESL_ALLOC(kp[i], sizeof(int) * (seqlen[i]+1));
ESL_ALLOC(tp[i], sizeof(int) * (seqlen[i]+1));
esl_vec_ISet(kp[i], seqlen[i]+1, -987654321);
esl_vec_ISet(tp[i], seqlen[i]+1, -987654321);
}
ESL_ALLOC(km_pos, sizeof(int) * (rflen+1));
ESL_ALLOC(ki_pos, sizeof(int) * (rflen+1));
ESL_ALLOC(tm_pos, sizeof(int) * (rflen+1));
ESL_ALLOC(ti_pos, sizeof(int) * (rflen+1));
ESL_ALLOC(cor_tm_pos, sizeof(int) * (rflen+1));
ESL_ALLOC(cor_ti_pos, sizeof(int) * (rflen+1));
esl_vec_ISet(km_pos, rflen+1, 0);
esl_vec_ISet(ki_pos, rflen+1, 0);
esl_vec_ISet(tm_pos, rflen+1, 0);
esl_vec_ISet(ti_pos, rflen+1, 0);
esl_vec_ISet(cor_tm_pos, rflen+1, 0);
esl_vec_ISet(cor_ti_pos, rflen+1, 0);
ESL_ALLOC(km_seq, sizeof(int) * ka->nseq);
ESL_ALLOC(ki_seq, sizeof(int) * ka->nseq);
ESL_ALLOC(tm_seq, sizeof(int) * ka->nseq);
ESL_ALLOC(ti_seq, sizeof(int) * ka->nseq);
ESL_ALLOC(cor_tm_seq, sizeof(int) * ka->nseq);
ESL_ALLOC(cor_ti_seq, sizeof(int) * ka->nseq);
esl_vec_ISet(km_seq, ka->nseq, 0);
esl_vec_ISet(ki_seq, ka->nseq, 0);
esl_vec_ISet(tm_seq, ka->nseq, 0);
esl_vec_ISet(ti_seq, ka->nseq, 0);
esl_vec_ISet(cor_tm_seq, ka->nseq, 0);
esl_vec_ISet(cor_ti_seq, ka->nseq, 0);
/* determine non-gap RF location of each residue in known alignment */
for(i = 0; i < ka->nseq; i++) {
uapos = rfpos = 0;
for(apos = 1; apos <= ka->alen; apos++) {
is_rfpos = FALSE;
if((! esl_abc_CIsGap (ka->abc, ka->rf[apos-1])) &&
(! esl_abc_CIsMissing (ka->abc, ka->rf[apos-1]))) {
rfpos++; is_rfpos = TRUE;
}
if(esl_abc_XIsResidue(ka->abc, ka->ax[i][apos])) {
uapos++;
kp[i][uapos] = (is_rfpos) ? rfpos : (-1 * rfpos);
if(is_rfpos) { km_pos[rfpos]++; km_seq[i]++; }
else { ki_pos[rfpos]++; ki_seq[i]++; }
}
}
}
/* determine non-gap RF location of each residue in predicted alignment */
for(i = 0; i < ta->nseq; i++) {
uapos = rfpos = 0;
for(apos = 1; apos <= ta->alen; apos++) {
is_rfpos = FALSE;
if((! esl_abc_CIsGap (abc, ta->rf[apos-1])) &&
(! esl_abc_CIsMissing (abc, ta->rf[apos-1]))) {
rfpos++; is_rfpos = TRUE;
if(do_post) {
do_post_for_this_rfpos = (mask != NULL && mask[rfpos-1] == '0') ? FALSE : TRUE;
}
}
if(esl_abc_XIsResidue(ta->abc, ta->ax[i][apos])) {
uapos++;
tp[i][uapos] = (is_rfpos) ? rfpos : (-1 * rfpos);
if(do_post) {
if(esl_abc_CIsGap(abc, ta->pp[i][(apos-1)])) esl_fatal("gap PP value for nongap residue: ali: %d seq: %d apos: %d\n", nali, i, apos);
ppidx = get_pp_idx(abc, ta->pp[i][(apos-1)]);
if(ppidx == -1) esl_fatal("unrecognized PP value (%c) for nongap residue: ali: %d seq: %d apos: %d\n", ta->pp[i][(apos-1)], nali, i, apos);
}
if(is_rfpos) {
tm_pos[rfpos]++; tm_seq[i]++;
if(do_post_for_this_rfpos) ptm[ppidx]++;
}
else {
ti_pos[rfpos]++; ti_seq[i]++;
if(do_post) pti[ppidx]++;
}
if(kp[i][uapos] == tp[i][uapos]) { /* correctly predicted this residue */
if(is_rfpos) {
cor_tm_seq[i]++; cor_tm_pos[rfpos]++;
if(do_post_for_this_rfpos) cor_ptm[ppidx]++;
}
else {
cor_ti_seq[i]++; cor_ti_pos[rfpos]++;
if(do_post) cor_pti[ppidx]++;
}
}
}
}
}
if((! (esl_opt_GetBoolean(go, "-c"))) && (! esl_opt_GetBoolean(go, "-p"))) {
/* print per sequence statistics */
/* determine the longest name in msa */
for(ni = 0; ni < ka->nseq; ni++) namewidth = ESL_MAX(namewidth, strlen(ka->sqname[ni]));
ESL_ALLOC(namedashes, sizeof(char) * namewidth+1);
namedashes[namewidth] = '\0';
for(ni = 0; ni < namewidth; ni++) namedashes[ni] = '-';
printf("# %-*s %6s %28s %28s %28s\n", namewidth, "seq name", "len", "match columns", "insert columns", "all columns");
printf("# %-*s %6s %28s %28s %28s\n", namewidth, namedashes, "------", "----------------------------", "----------------------------", "----------------------------");
for(i = 0; i < ta->nseq; i++) {
printf(" %-*s %6d %8d / %8d (%.3f) %8d / %8d (%.3f) %8d / %8d (%.3f)\n", namewidth, ka->sqname[i], seqlen[i],
cor_tm_seq[i], km_seq[i], (km_seq[i] == 0) ? 0. : ((float) cor_tm_seq[i] / (float) km_seq[i]),
cor_ti_seq[i], ki_seq[i], (ki_seq[i] == 0) ? 0. : ((float) cor_ti_seq[i] / (float) ki_seq[i]),
(cor_tm_seq[i] + cor_ti_seq[i]), (km_seq[i] + ki_seq[i]), ((float) (cor_tm_seq[i] + cor_ti_seq[i]) / ((float) km_seq[i] + ki_seq[i])));
}
cor_tm = esl_vec_ISum(cor_tm_seq, ka->nseq);
cor_ti = esl_vec_ISum(cor_ti_seq, ka->nseq);
km = esl_vec_ISum(km_seq, ka->nseq);
ki = esl_vec_ISum(ki_seq, ka->nseq);
printf("# %-*s %6s %28s %28s %28s\n", namewidth, namedashes, "-----", "----------------------------", "----------------------------", "----------------------------");
printf("# %-*s %6s %8d / %8d (%.3f) %8d / %8d (%.3f) %8d / %8d (%.3f)\n",
namewidth, "*all*", "-",
cor_tm, km, ((float) cor_tm / (float) km),
cor_ti, ki, ((float) cor_ti / (float) ki),
(cor_tm+cor_ti), (km+ki), (((float) (cor_tm + cor_ti))/ ((float) (km + ki))));
free(namedashes);
for(i = 0; i < ka->nseq; i++) {
free(kp[i]);
free(tp[i]);
}
}
else if(esl_opt_GetBoolean(go, "-c")) { /* print per column statistics */
printf("# %5s %20s %20s %20s\n", "rfpos", "match", "insert", "both");
printf("# %5s %20s %20s %20s\n", "-----", "--------------------", "--------------------", "--------------------");
for(rfpos = 0; rfpos <= rflen; rfpos++) {
printf(" %5d %4d / %4d (%.3f) %4d / %4d (%.3f) %4d / %4d (%.3f)\n", rfpos,
cor_tm_pos[rfpos], km_pos[rfpos], (km_pos[rfpos] == 0) ? 0. : ((float) cor_tm_pos[rfpos] / (float) km_pos[rfpos]),
cor_ti_pos[rfpos], ki_pos[rfpos], (ki_pos[rfpos] == 0) ? 0. : ((float) cor_ti_pos[rfpos] / (float) ki_pos[rfpos]),
(cor_tm_pos[rfpos] + cor_ti_pos[rfpos]), (km_pos[rfpos] + ki_pos[rfpos]), ((float) (cor_tm_pos[rfpos] + cor_ti_pos[rfpos]) / ((float) km_pos[rfpos] + ki_pos[rfpos])));
}
}
else if(do_post) { /* do posterior output */
if(mask == NULL) {
printf("# %2s %29s %29s\n", "", " match columns ", " insert columns ");
printf("# %2s %29s %29s\n", "", "-----------------------------", "-----------------------------") ;
printf("# %2s %8s %8s %9s %8s %8s %9s\n", "PP", "ncorrect", "ntotal", "fractcor", "ncorrect", "ntotal", "fractcor");
printf("# %2s %8s %8s %9s %8s %8s %9s\n", "--", "--------", "--------", "---------", "--------", "--------", "---------");
}
else {
printf("# %2s %29s %29s\n", "", " match columns within mask ", " insert columns ");
printf("# %2s %29s %29s\n", "", "-----------------------------", "-----------------------------") ;
printf("# %2s %8s %8s %9s %8s %8s %9s\n", "PP", "ncorrect", "ntotal", "fractcor", "ncorrect", "ntotal", "fractcor");
printf("# %2s %8s %8s %9s %8s %8s %9s\n", "--", "--------", "--------", "---------", "--------", "--------", "---------");
}
cm_ptm = cm_pti = cm_cor_ptm = cm_cor_pti = cm_incor_ptm = cm_incor_pti = 0;
//tot_ptm = esl_vec_ISum(ptm, npostvals);
//tot_pti = esl_vec_ISum(pti, npostvals);
//tot_cor_ptm = esl_vec_ISum(cor_ptm, npostvals);
//tot_cor_pti = esl_vec_ISum(cor_pti, npostvals);
//tot_incor_ptm = tot_ptm - tot_cor_ptm;
//tot_incor_pti = tot_pti - tot_cor_pti;
for(p = (npostvals-1); p >= 0; p--) {
cm_cor_ptm += cor_ptm[p];
cm_cor_pti += cor_pti[p];
cm_ptm += ptm[p];
cm_pti += pti[p];
cm_incor_ptm += ptm[p] - cor_ptm[p];
cm_incor_pti += pti[p] - cor_pti[p];
printf(" %2c %8d / %8d (%.5f) %8d / %8d (%.5f)\n",
ppchars[p], cor_ptm[p], ptm[p],
(ptm[p] == 0) ? 0. : (float) cor_ptm[p] / (float) ptm[p],
cor_pti[p], pti[p],
(pti[p] == 0) ? 0. : (float) cor_pti[p] / (float) pti[p]);
}
}
/* handle --c2dfile */
if (dfp != NULL) {
/* match stats, 4 fields, CMYK color values */
for(rfpos = 1; rfpos <= rflen; rfpos++) {
if(km_pos[rfpos] == 0) { /* special case, no known alignment residues, a blank position */
fprintf(dfp, "%.3f %.3f %.3f %.3f\n", 0., 0., 0., 0.);
}
else {
fprintf(dfp, "%.3f %.3f %.3f %.3f\n",
0., /* cyan */
1. - ((float) cor_tm_pos[rfpos] / (float) km_pos[rfpos]), /* magenta, fraction incorrect */
1. - ((float) km_pos[rfpos] / ta->nseq), /* yellow, 1 - fraction of seqs with residue in column */
0.);
}
}
fprintf(dfp, "//\n");
/* insert stats, 4 fields, CMYK color values */
rfpos = 0; /* special case, combine insert posn 0 and 1 together */
if(ki_pos[rfpos] == 0) { /* special case, no known alignment residues, a blank position */
fprintf(dfp, "%.3f %.3f %.3f %.3f\n", 0., 0., 0., 0.);
}
else {
fprintf(dfp, "%.3f %.3f %.3f %.3f\n",
0., /* cyan */
1. - ((float) (cor_ti_pos[0] + cor_ti_pos[1]) / ((float) (ki_pos[0] + ki_pos[1]))), /* magenta, fraction correct */
0.,
0.);
}
/* insert stats posn 2..rflen */
for(rfpos = 2; rfpos <= rflen; rfpos++) {
if(ki_pos[rfpos] == 0) { /* special case, no known alignment residues, a blank position */
fprintf(dfp, "%.3f %.3f %.3f %.3f\n", 0., 0., 0., 0.);
}
else {
fprintf(dfp, "%.3f %.3f %.3f %.3f\n",
0., /* cyan */
1. - ((float) cor_ti_pos[rfpos] / (float) ki_pos[rfpos]), /* magenta, fraction correct */
0.,
0.);
}
}
fprintf(dfp, "//\n");
}
if(ptm != NULL) free(ptm);
if(pti != NULL) free(pti);
if(cor_ptm != NULL) free(cor_ptm);
if(cor_ptm != NULL) free(cor_pti);
free(kp);
free(tp);
free(km_seq);
free(ki_seq);
free(tm_seq);
free(ti_seq);
free(cor_tm_seq);
free(cor_ti_seq);
free(km_pos);
free(ki_pos);
free(tm_pos);
free(ti_pos);
free(cor_tm_pos);
free(cor_ti_pos);
free(seqlen);
esl_msa_Destroy(ka);
esl_msa_Destroy(ta);
}
if(mask != NULL) free(mask);
if(dfp != NULL) {
fclose(dfp);
printf("# Draw file of per-column stats saved to file: %s\n", esl_opt_GetString(go, "--c2dfile"));
}
if(abc) esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
esl_msafile_Close(tfp);
esl_msafile_Close(kfp);
return 0;
ERROR:
return status;
}
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;
}
