/**
* @brief copies an mseq structure
*
* @param[out] prMSeqDest_p
* Copy of mseq structure
* @param[in] prMSeqSrc
* Source mseq structure to copy
*
* @note caller has to free copy by calling FreeMSeq()
*
*/
void
CopyMSeq(mseq_t **prMSeqDest_p, mseq_t *prMSeqSrc)
{
int i;
assert(prMSeqSrc != NULL && prMSeqDest_p != NULL);
NewMSeq(prMSeqDest_p);
(*prMSeqDest_p)->nseqs = prMSeqSrc->nseqs;
(*prMSeqDest_p)->seqtype = prMSeqSrc->seqtype;
if (prMSeqSrc->filename!=NULL) {
(*prMSeqDest_p)->filename = CkStrdup(prMSeqSrc->filename);
}
(*prMSeqDest_p)->seq = (char **)
CKMALLOC((*prMSeqDest_p)->nseqs * sizeof(char *));
(*prMSeqDest_p)->orig_seq = (char **)
CKMALLOC((*prMSeqDest_p)->nseqs * sizeof(char *));
(*prMSeqDest_p)->sqinfo = (SQINFO *)
CKMALLOC((*prMSeqDest_p)->nseqs * sizeof(SQINFO));
for (i=0; i<(*prMSeqDest_p)->nseqs; i++) {
(*prMSeqDest_p)->seq[i] = CkStrdup(prMSeqSrc->seq[i]);
(*prMSeqDest_p)->orig_seq[i] = CkStrdup(prMSeqSrc->orig_seq[i]);
SeqinfoCopy(&(*prMSeqDest_p)->sqinfo[i], &prMSeqSrc->sqinfo[i]);
}
}
/**
* @brief Appends an mseq structure to an already existing one.
* filename will be left untouched.
*
* @param[in] prMSeqDest_p
* MSeq structure to which to append to
* @param[out] prMSeqToAdd
* MSeq structure which is to append
*
*
*/
void
JoinMSeqs(mseq_t **prMSeqDest_p, mseq_t *prMSeqToAdd)
{
int iSrcSeqIndex;
int iNewNSeq;
assert(NULL != prMSeqDest_p && NULL != (*prMSeqDest_p));
assert(NULL != prMSeqToAdd);
if (0 == prMSeqToAdd->nseqs) {
Log(&rLog, LOG_WARN, "Was asked to add 0 sequences");
return;
}
/* warn on seqtype mismatch and keep original seqtype */
if ((*prMSeqDest_p)->seqtype != prMSeqToAdd->seqtype) {
Log(&rLog, LOG_WARN, "Joining sequences of different type");
}
/* leave filename as it is */
/*
* copy new seq/s, orig_seq/s, sqinfo/s
*/
iNewNSeq = (*prMSeqDest_p)->nseqs + prMSeqToAdd->nseqs;
(*prMSeqDest_p)->seq = (char **)
CKREALLOC((*prMSeqDest_p)->seq, iNewNSeq * sizeof(char *));
(*prMSeqDest_p)->orig_seq = (char **)
CKREALLOC((*prMSeqDest_p)->orig_seq, iNewNSeq * sizeof(char *));
(*prMSeqDest_p)->sqinfo = (SQINFO *)
CKREALLOC((*prMSeqDest_p)->sqinfo, iNewNSeq * sizeof(SQINFO));
for (iSrcSeqIndex=0; iSrcSeqIndex < prMSeqToAdd->nseqs; iSrcSeqIndex++) {
int iDstSeqIndex = (*prMSeqDest_p)->nseqs++;
(*prMSeqDest_p)->seq[iDstSeqIndex] =
CkStrdup(prMSeqToAdd->seq[iSrcSeqIndex]);
(*prMSeqDest_p)->orig_seq[iDstSeqIndex] =
CkStrdup(prMSeqToAdd->orig_seq[iSrcSeqIndex]);
SeqinfoCopy(&(*prMSeqDest_p)->sqinfo[iDstSeqIndex],
& prMSeqToAdd->sqinfo[iSrcSeqIndex]);
}
(*prMSeqDest_p)->nseqs = iNewNSeq;
(*prMSeqDest_p)->aligned = SeqsAreAligned(*prMSeqDest_p);
return;
}
/**
* @brief Creates a new sequence entry and appends it to an existing mseq
* structure.
*
* @param[out] prMSeqDest_p
* Already existing and initialised mseq structure
* @param[in] pcSeqName
* sequence name of the sequence to add
* @param[in] pcSeqRes
* the actual sequence (residues) to add
*
* @note Don't forget to update the align and type flag if necessary!
*
* FIXME allow adding of more features
*
*/
void
AddSeq(mseq_t **prMSeqDest_p, char *pcSeqName, char *pcSeqRes)
{
int iSeqIdx = 0;
SQINFO sqinfo;
assert(NULL != prMSeqDest_p);
assert(NULL != pcSeqName);
assert(NULL != pcSeqRes);
iSeqIdx = (*prMSeqDest_p)->nseqs;
(*prMSeqDest_p)->seq = (char **)
CKREALLOC((*prMSeqDest_p)->seq, (iSeqIdx+1) * sizeof(char *));
(*prMSeqDest_p)->orig_seq = (char **)
CKREALLOC((*prMSeqDest_p)->orig_seq, (iSeqIdx+1) * sizeof(char *));
(*prMSeqDest_p)->sqinfo = (SQINFO *)
CKREALLOC((*prMSeqDest_p)->sqinfo, (iSeqIdx+1) * sizeof(SQINFO));
(*prMSeqDest_p)->seq[iSeqIdx] = CkStrdup(pcSeqRes);
(*prMSeqDest_p)->orig_seq[iSeqIdx] = CkStrdup(pcSeqRes);
/* should probably get ri of SqInfo altogether in the long run and just
transfer the intersting members into our own struct
*/
sqinfo.flags = 0; /* init */
sqinfo.len = strlen(pcSeqRes);
sqinfo.flags |= SQINFO_LEN;
/* name is an array of SQINFO_NAMELEN length */
strncpy(sqinfo.name, pcSeqName, SQINFO_NAMELEN-1);
sqinfo.name[SQINFO_NAMELEN-1] = '\0';
sqinfo.flags |= SQINFO_NAME;
SeqinfoCopy(&(*prMSeqDest_p)->sqinfo[iSeqIdx],
& sqinfo);
(*prMSeqDest_p)->nseqs++;
return;
}
/**
* @brief Sort sequences by length
*
* @param[out] prMSeq
* mseq to sort by length
* @param[out] cOrder
* Sorting order. 'd' for descending, 'a' for ascending.
*
*
*/
void
SortMSeqByLength(mseq_t *prMSeq, const char cOrder)
{
int *piSeqLen;
int *piOrder;
int iSeqIndex;
mseq_t *prMSeqCopy = NULL;
assert('a'==cOrder || 'd'==cOrder);
Log(&rLog, LOG_WARN,
"FIXME: This modifies sequence ordering. Might not be what user wants. Will change output order as well");
piSeqLen = (int *) CKMALLOC(prMSeq->nseqs * sizeof(int));
piOrder = (int *) CKMALLOC(prMSeq->nseqs * sizeof(int));
for (iSeqIndex=0; iSeqIndex<prMSeq->nseqs; iSeqIndex++) {
piSeqLen[iSeqIndex] = prMSeq->sqinfo[iSeqIndex].len;
}
QSortAndTrackIndex(piOrder, piSeqLen, prMSeq->nseqs, cOrder, FALSE);
CopyMSeq(&prMSeqCopy, prMSeq);
for (iSeqIndex=0; iSeqIndex<prMSeq->nseqs; iSeqIndex++) {
/* copy mseq entry
*/
CKFREE(prMSeq->seq[iSeqIndex]);
prMSeq->seq[iSeqIndex] = CkStrdup(prMSeqCopy->seq[piOrder[iSeqIndex]]);
CKFREE(prMSeq->orig_seq[iSeqIndex]);
prMSeq->orig_seq[iSeqIndex] = CkStrdup(prMSeqCopy->orig_seq[piOrder[iSeqIndex]]);
SeqinfoCopy(&prMSeq->sqinfo[iSeqIndex], &prMSeqCopy->sqinfo[piOrder[iSeqIndex]]);
}
CKFREE(piSeqLen);
CKFREE(piOrder);
FreeMSeq(&prMSeqCopy);
return;
}
/**
* @brief reads sequences from file
*
* @param[out] prMSeq
* Multiple sequence struct. Must be preallocated.
* FIXME: would make more sense to allocate it here.
* @param[in] seqfile
* Sequence file name. If '-' sequence will be read from stdin.
* @param[in] iSeqType
* int-encoded sequence type. Set to
* SEQTYPE_UNKNOWN for autodetect (guessed from first sequence)
* @param[in] iMaxNumSeq
* Return an error, if more than iMaxNumSeq have been read
* @param[in] iMaxSeqLen
* Return an error, if a seq longer than iMaxSeqLen has been read
*
* @return 0 on success, -1 on error
*
* @note
* - Depends heavily on squid
* - Sequence file format will be guessed
* - If supported by squid, gzipped files can be read as well.
*/
int
ReadSequences(mseq_t *prMSeq, char *seqfile,
int iSeqType, int iSeqFmt, bool bIsProfile, bool bDealignInputSeqs,
int iMaxNumSeq, int iMaxSeqLen)
{
SQFILE *dbfp; /* sequence file descriptor */
char *cur_seq;
SQINFO cur_sqinfo;
int iSeqIdx; /* sequence counter */
int iSeqPos; /* sequence string position counter */
assert(NULL!=seqfile);
/* Try to work around inability to autodetect from a pipe or .gz:
* assume FASTA format
*/
if (SQFILE_UNKNOWN == iSeqFmt &&
(Strparse("^.*\\.gz$", seqfile, 0) || strcmp(seqfile, "-") == 0)) {
iSeqFmt = SQFILE_FASTA;
}
/* Using squid routines to read input. taken from seqstat_main.c. we don't
* know if input is aligned, so we use SeqfileOpen instead of MSAFileOpen
* etc. NOTE this also means we discard some information, e.g. when
* reading from and writing to a stockholm file, all extra MSA
* info/annotation will be lost.
*
*/
if (NULL == (dbfp = SeqfileOpen(seqfile, iSeqFmt, NULL))) {
Log(&rLog, LOG_ERROR, "Failed to open sequence file %s for reading", seqfile);
return -1;
}
/* FIXME squid's ReadSeq() will exit with fatal error if format is
* unknown. This will be a problem for a GUI. Same is true for many squid
* other functions.
*
* The original squid:ReadSeq() dealigns sequences on input. We
* use a patched version.
*
*/
while (ReadSeq(dbfp, dbfp->format,
&cur_seq,
&cur_sqinfo)) {
if (prMSeq->nseqs+1>iMaxNumSeq) {
Log(&rLog, LOG_ERROR, "Maximum number of sequences (=%d) exceeded after reading sequence '%s' from '%s'",
iMaxNumSeq, cur_sqinfo.name, seqfile);
return -1;
}
if ((int)strlen(cur_seq)>iMaxSeqLen) {
Log(&rLog, LOG_ERROR, "Sequence '%s' has %d residues and is therefore longer than allowed (max. sequence length is %d)",
cur_sqinfo.name, strlen(cur_seq), iMaxSeqLen);
return -1;
}
if ((int)strlen(cur_seq)==0) {
Log(&rLog, LOG_ERROR, "Sequence '%s' has 0 residues",
cur_sqinfo.name);
return -1;
}
/* FIXME: use modified version of AddSeq() that allows handing down SqInfo
*/
prMSeq->seq = (char **)
CKREALLOC(prMSeq->seq, (prMSeq->nseqs+1) * sizeof(char *));
prMSeq->seq[prMSeq->nseqs] = CkStrdup(cur_seq);
prMSeq->sqinfo = (SQINFO *)
CKREALLOC(prMSeq->sqinfo, (prMSeq->nseqs+1) * sizeof(SQINFO));
SeqinfoCopy(&prMSeq->sqinfo[prMSeq->nseqs], &cur_sqinfo);
#ifdef TRACE
Log(&rLog, LOG_FORCED_DEBUG, "seq no %d: seq = %s", prMSeq->nseqs, prMSeq->seq[prMSeq->nseqs]);
LogSqInfo(&prMSeq->sqinfo[prMSeq->nseqs]);
#endif
/* always guess type from first seq. use squid function and
* convert value
*/
if (0 == prMSeq->nseqs) {
int type = Seqtype(prMSeq->seq[prMSeq->nseqs]);
switch (type) {
case kDNA:
prMSeq->seqtype = SEQTYPE_DNA;
break;
case kRNA:
prMSeq->seqtype = SEQTYPE_RNA;
break;
case kAmino:
prMSeq->seqtype = SEQTYPE_PROTEIN;
break;
case kOtherSeq:
prMSeq->seqtype = SEQTYPE_UNKNOWN;
break;
default:
Log(&rLog, LOG_FATAL, "Internal error in %s", __FUNCTION__);
}
/* override with given sequence type but check with
* automatically detected type and warn if necessary
*/
if (SEQTYPE_UNKNOWN != iSeqType) {
if (prMSeq->seqtype != iSeqType) {
Log(&rLog, LOG_WARN, "Overriding automatically determined seq-type %s to %s as requested",
SeqTypeToStr(prMSeq->seqtype), SeqTypeToStr(iSeqType));
prMSeq->seqtype = iSeqType;
}
}
/* if type could not be determined and was not set return error */
if (SEQTYPE_UNKNOWN == iSeqType && SEQTYPE_UNKNOWN == prMSeq->seqtype) {
Log(&rLog, LOG_ERROR, "Couldn't guess sequence type from first sequence");
FreeSequence(cur_seq, &cur_sqinfo);
SeqfileClose(dbfp);
return -1;
}
}
Log(&rLog, LOG_DEBUG, "seq-no %d: type=%s name=%s len=%d seq=%s",
prMSeq->nseqs, SeqTypeToStr(prMSeq->seqtype),
prMSeq->sqinfo[prMSeq->nseqs].name, prMSeq->sqinfo[prMSeq->nseqs].len,
prMSeq->seq[prMSeq->nseqs]);
/* FIXME IPUAC and/or case conversion? If yes see
* corresponding squid functions. Special treatment of
* Stockholm tilde-gaps for ktuple code?
*/
prMSeq->nseqs++;
FreeSequence(cur_seq, &cur_sqinfo);
}
SeqfileClose(dbfp);
/*#if ALLOW_ONLY_PROTEIN
if (SEQTYPE_PROTEIN != prMSeq->seqtype) {
Log(&rLog, LOG_FATAL, "Sequence type is %s. %s only works on protein.",
SeqTypeToStr(prMSeq->seqtype), PACKAGE_NAME);
}
#endif*/
/* Check if sequences are aligned */
prMSeq->aligned = SeqsAreAligned(prMSeq, bIsProfile, bDealignInputSeqs);
/* keep original sequence as copy and convert "working" sequence
*
*/
prMSeq->orig_seq = (char**) CKMALLOC(prMSeq->nseqs * sizeof(char *));
for (iSeqIdx=0; iSeqIdx<prMSeq->nseqs; iSeqIdx++) {
prMSeq->orig_seq[iSeqIdx] = CkStrdup(prMSeq->seq[iSeqIdx]);
/* convert unknown characters according to set seqtype
* be conservative, i.e. don't allow any fancy ambiguity
* characters to make sure that ktuple code etc. works.
*/
/* first on the fly conversion between DNA and RNA
*/
if (prMSeq->seqtype==SEQTYPE_DNA)
ToDNA(prMSeq->seq[iSeqIdx]);
if (prMSeq->seqtype==SEQTYPE_RNA)
ToRNA(prMSeq->seq[iSeqIdx]);
/* then check of each character
*/
for (iSeqPos=0; iSeqPos<(int)strlen(prMSeq->seq[iSeqIdx]); iSeqPos++) {
char *res = &(prMSeq->seq[iSeqIdx][iSeqPos]);
if (isgap(*res))
continue;
if (prMSeq->seqtype==SEQTYPE_PROTEIN) {
if (NULL == strchr(AMINO_ALPHABET, toupper(*res))) {
*res = AMINOACID_ANY;
}
} else if (prMSeq->seqtype==SEQTYPE_DNA) {
if (NULL == strchr(DNA_ALPHABET, toupper(*res))) {
*res = NUCLEOTIDE_ANY;
}
} else if (prMSeq->seqtype==SEQTYPE_RNA) {
if (NULL == strchr(RNA_ALPHABET, toupper(*res))) {
*res = NUCLEOTIDE_ANY;
}
}
}
}
/* order in which sequences appear in guide-tree
* only allocate if different output-order desired */
prMSeq->tree_order = NULL;
prMSeq->filename = CkStrdup(seqfile);
Log(&rLog, LOG_INFO, "Read %d sequences (type: %s) from %s",
prMSeq->nseqs, SeqTypeToStr(prMSeq->seqtype), prMSeq->filename);
return 0;
}
/**
* @brief Appends an mseq structure to an already existing one.
* filename will be left untouched.
*
* @param[in] prMSeqDest_p
* MSeq structure to which to append to
* @param[out] prMSeqToAdd
* MSeq structure which is to append
*
*/
void
JoinMSeqs(mseq_t **prMSeqDest_p, mseq_t *prMSeqToAdd)
{
int iSrcSeqIndex;
int iNewNSeq;
assert(NULL != prMSeqDest_p && NULL != (*prMSeqDest_p));
assert(NULL != prMSeqToAdd);
if (0 == prMSeqToAdd->nseqs) {
Log(&rLog, LOG_WARN, "Was asked to add 0 sequences");
return;
}
/* warn on seqtype mismatch and keep original seqtype */
if ((*prMSeqDest_p)->seqtype != prMSeqToAdd->seqtype) {
Log(&rLog, LOG_WARN, "Joining sequences of different type");
}
/* leave filename as it is */
/*
* copy new seq/s, orig_seq/s, sqinfo/s
*/
iNewNSeq = (*prMSeqDest_p)->nseqs + prMSeqToAdd->nseqs;
(*prMSeqDest_p)->seq = (char **)
CKREALLOC((*prMSeqDest_p)->seq, iNewNSeq * sizeof(char *));
(*prMSeqDest_p)->orig_seq = (char **)
CKREALLOC((*prMSeqDest_p)->orig_seq, iNewNSeq * sizeof(char *));
(*prMSeqDest_p)->sqinfo = (SQINFO *)
CKREALLOC((*prMSeqDest_p)->sqinfo, iNewNSeq * sizeof(SQINFO));
for (iSrcSeqIndex=0; iSrcSeqIndex < prMSeqToAdd->nseqs; iSrcSeqIndex++) {
int iDstSeqIndex = (*prMSeqDest_p)->nseqs++;
(*prMSeqDest_p)->seq[iDstSeqIndex] =
CkStrdup(prMSeqToAdd->seq[iSrcSeqIndex]);
(*prMSeqDest_p)->orig_seq[iDstSeqIndex] =
CkStrdup(prMSeqToAdd->orig_seq[iSrcSeqIndex]);
SeqinfoCopy(&(*prMSeqDest_p)->sqinfo[iDstSeqIndex],
& prMSeqToAdd->sqinfo[iSrcSeqIndex]);
}
(*prMSeqDest_p)->nseqs = iNewNSeq;
#if 0
/* 2nd arg is bIsProfile, which when set TRUE skips
* the check for gaps. here always check for gaps,
* so set FALSE (main reason is that it is easier), FS, r282 -> */
/* had a problem at this stage, therefore dispense with gap check, FS, r290 -> */
/* 3rd argument is dealignment flag, do not dealign profiles */
(*prMSeqDest_p)->aligned = SeqsAreAligned(*prMSeqDest_p, TRUE/*FALSE*/, FALSE);
#else
(*prMSeqDest_p)->aligned = TRUE;
#endif
return;
}
/**
* @brief Parse command line parameters. Will exit if help/usage etc
* are called or or call Log(&rLog, LOG_FATAL, ) if an error was detected.
*
* @param[out] user_opts
* User parameter struct, with defaults already set.
* @param[in] argc
* mains argc
* @param[in] argv
* mains argv
*
*/
void
ParseCommandLine(cmdline_opts_t *user_opts, int argc, char **argv)
{
/* argtable command line parsing:
* see
* http://argtable.sourceforge.net/doc/argtable2-intro.html
*
* basic structure is: arg_xxxN:
* xxx can be int, lit, db1, str, rex, file or date
* If N = 0, arguments may appear zero-or-once; N = 1 means
* exactly once, N = n means up to maxcount times
*
*
* @note: changes here, might also affect main.cpp:ConvertOldCmdLine()
*
*/
struct arg_rem *rem_seq_input = arg_rem(NULL, "\nSequence Input:");
struct arg_file *opt_seqin = arg_file0("i", "in,infile",
"{<file>,-}",
"Multiple sequence input file (- for stdin)");
struct arg_file *opt_hmm_in = arg_filen(NULL, "hmm-in", "<file>",
/*min*/ 0, /*max*/ 128,
"HMM input files");
struct arg_lit *opt_dealign = arg_lit0(NULL, "dealign",
"Dealign input sequences");
struct arg_file *opt_profile1 = arg_file0(NULL, "profile1,p1",
"<file>",
"Pre-aligned multiple sequence file (aligned columns will be kept fix)");
struct arg_file *opt_profile2 = arg_file0(NULL, "profile2,p2",
"<file>",
"Pre-aligned multiple sequence file (aligned columns will be kept fix)");
struct arg_str *opt_seqtype = arg_str0("t", "seqtype",
"{Protein, RNA, DNA}",
"Force a sequence type (default: auto)");
/* struct arg_lit *opt_force_protein = arg_lit0(NULL, "protein",
"Set sequence type to protein even if Clustal guessed nucleic acid"); */
struct arg_str *opt_infmt = arg_str0(NULL, "infmt",
"{a2m=fa[sta],clu[stal],msf,phy[lip],selex,st[ockholm],vie[nna]}",
"Forced sequence input file format (default: auto)");
struct arg_rem *rem_guidetree = arg_rem(NULL, "\nClustering:");
struct arg_str *opt_pairdist = arg_str0("p", "pairdist",
"{ktuple}",
"Pairwise alignment distance measure");
struct arg_file *opt_distmat_in = arg_file0(NULL, "distmat-in",
"<file>",
"Pairwise distance matrix input file (skips distance computation)");
struct arg_file *opt_distmat_out = arg_file0(NULL, "distmat-out",
"<file>",
"Pairwise distance matrix output file");
struct arg_file *opt_guidetree_in = arg_file0(NULL, "guidetree-in",
"<file>",
"Guide tree input file (skips distance computation and guide-tree clustering step)");
struct arg_file *opt_guidetree_out = arg_file0(NULL, "guidetree-out",
"<file>",
"Guide tree output file");
/* AW: mbed is default since at least R253
struct arg_lit *opt_mbed = arg_lit0(NULL, "mbed",
"Fast, Mbed-like clustering for guide-tree calculation");
struct arg_lit *opt_mbed_iter = arg_lit0(NULL, "mbed-iter",
"Use Mbed-like clustering also during iteration");
*/
/* Note: might be better to use arg_str (mbed=YES/NO) but I don't want to introduce an '=' into pipeline, FS, r250 -> */
struct arg_lit *opt_full = arg_lit0(NULL, "full",
"Use full distance matrix for guide-tree calculation (might be slow; mBed is default)");
struct arg_lit *opt_full_iter = arg_lit0(NULL, "full-iter",
"Use full distance matrix for guide-tree calculation during iteration (might be slowish; mBed is default)");
struct arg_str *opt_clustering = arg_str0("c", "clustering",
"{UPGMA}",
"Clustering method for guide tree");
struct arg_rem *rem_aln_output = arg_rem(NULL, "\nAlignment Output:");
struct arg_file *opt_outfile = arg_file0("o", "out,outfile",
"{file,-}",
"Multiple sequence alignment output file (default: stdout)");
struct arg_str *opt_outfmt = arg_str0(NULL, "outfmt",
"{a2m=fa[sta],clu[stal],msf,phy[lip],selex,st[ockholm],vie[nna]}",
"MSA output file format (default: fasta)");
struct arg_rem *rem_iteration = arg_rem(NULL, "\nIteration:");
struct arg_str *opt_num_iterations = arg_str0(NULL, "iterations,iter",
/* FIXME "{<n>,auto}", "Number of combined guide-tree/HMM iterations"); */
"<n>", "Number of (combined guide-tree/HMM) iterations");
struct arg_int *opt_max_guidetree_iterations = arg_int0(NULL, "max-guidetree-iterations",
"<n>", "Maximum number guidetree iterations");
struct arg_int *opt_max_hmm_iterations = arg_int0(NULL, "max-hmm-iterations",
"<n>", "Maximum number of HMM iterations");
struct arg_rem *rem_limits = arg_rem(NULL, "\nLimits (will exit early, if exceeded):");
struct arg_int *opt_max_seq = arg_int0(NULL, "maxnumseq", "<n>",
"Maximum allowed number of sequences");
struct arg_int *opt_max_seqlen = arg_int0(NULL, "maxseqlen", "<l>",
"Maximum allowed sequence length");
struct arg_rem *rem_misc = arg_rem(NULL, "\nMiscellaneous:");
struct arg_lit *opt_autooptions = arg_lit0(NULL, "auto",
"Set options automatically (might overwrite some of your options)");
struct arg_int *opt_threads = arg_int0(NULL, "threads", "<n>",
"Number of processors to use");
struct arg_file *opt_logfile = arg_file0("l", "log",
"<file>",
"Log all non-essential output to this file");
struct arg_lit *opt_help = arg_lit0("h", "help",
"Print this help and exit");
struct arg_lit *opt_version = arg_lit0(NULL, "version",
"Print version information and exit");
struct arg_lit *opt_long_version = arg_lit0(NULL, "long-version",
"Print long version information and exit");
struct arg_lit *opt_verbose = arg_litn("v", "verbose",
0, 3,
"Verbose output (increases if given multiple times)");
struct arg_lit *opt_force = arg_lit0(NULL, "force",
"Force file overwriting");
struct arg_int *opt_macram = arg_int0(NULL, "MAC-RAM", "<n>", /* keep this quiet for the moment, FS r240 -> */
NULL/*"maximum amount of RAM to use for MAC algorithm (in MB)"*/);
struct arg_end *opt_end = arg_end(10); /* maximum number of errors
* to store */
void *argtable[] = {rem_seq_input,
opt_seqin,
opt_hmm_in,
opt_dealign,
opt_profile1,
opt_profile2,
opt_seqtype,
/* opt_force_protein, */
opt_infmt,
rem_guidetree,
#if 0
/* no other options then default available or not implemented */
opt_pairdist,
#endif
opt_distmat_in,
opt_distmat_out,
opt_guidetree_in,
opt_guidetree_out,
opt_full, /* FS, r250 -> */
opt_full_iter, /* FS, r250 -> */
#if 0
/* no other options then default available */
opt_clustering,
#endif
rem_aln_output,
opt_outfile,
opt_outfmt,
rem_iteration,
opt_num_iterations,
opt_max_guidetree_iterations,
opt_max_hmm_iterations,
rem_limits,
opt_max_seq,
opt_max_seqlen,
rem_misc,
opt_autooptions,
opt_threads,
opt_logfile,
opt_help,
opt_verbose,
opt_version,
opt_long_version,
opt_force,
opt_macram, /* FS, r240 -> r241 */
opt_end};
int nerrors;
/* Verify the argtable[] entries were allocated sucessfully
*/
if (arg_nullcheck(argtable)) {
Log(&rLog, LOG_FATAL, "insufficient memory (for argtable allocation)");
}
/* Parse the command line as defined by argtable[]
*/
nerrors = arg_parse(argc, argv, argtable);
/* Special case: '--help' takes precedence over error reporting
*/
if (opt_help->count > 0) {
printf("%s - %s (%s)\n", PACKAGE_NAME, PACKAGE_VERSION, PACKAGE_CODENAME);
printf("\n");
printf("If you like Clustal-Omega please cite:\n%s\n", CITATION);
printf("If you don't like Clustal-Omega, please let us know why (and cite us anyway).\n");
/* printf("You can contact reach us under %s\n", PACKAGE_BUGREPORT); */
printf("\n");
printf("Check http://www.clustal.org for more information and updates.\n");
printf("\n");
printf("Usage: %s", basename(argv[0]));
arg_print_syntax(stdout,argtable, "\n");
printf("\n");
printf("A typical invocation would be: %s -i my-in-seqs.fa -o my-out-seqs.fa -v\n",
basename(argv[0]));
printf("See below for a list of all options.\n");
arg_print_glossary(stdout, argtable, " %-25s %s\n");
arg_freetable(argtable, sizeof(argtable)/sizeof(argtable[0]));
exit(EXIT_SUCCESS);
}
/* Special case: '--version' takes precedence over error reporting
*/
if (opt_version->count > 0) {
printf("%s\n", PACKAGE_VERSION);
arg_freetable(argtable,sizeof(argtable)/sizeof(argtable[0]));
exit(EXIT_SUCCESS);
}
/* Special case: '--long-version' takes precedence over error reporting
*/
if (opt_long_version->count > 0) {
char zcLongVersion[1024];
PrintLongVersion(zcLongVersion, sizeof(zcLongVersion));
printf("%s\n", zcLongVersion);
arg_freetable(argtable,sizeof(argtable)/sizeof(argtable[0]));
exit(EXIT_SUCCESS);
}
/* If the parser returned any errors then display them and exit
*/
if (nerrors > 0) {
/* Display the error details contained in the arg_end struct.*/
arg_print_errors(stdout, opt_end, PACKAGE);
fprintf(stderr, "For more information try: %s --help\n", argv[0]);
arg_freetable(argtable,sizeof(argtable)/sizeof(argtable[0]));
exit(EXIT_FAILURE);
}
/* ------------------------------------------------------------
*
* Command line successfully parsed. Now transfer values to
* user_opts. While doing so, make sure that given input files
* exist and given output files are writable do not exist, or if
* they do, should be overwritten.
*
* No logic checks here! They are done in a different function
*
* ------------------------------------------------------------*/
/* not part of user_opts because it declared in src/util.h */
if (0 == opt_verbose->count) {
rLog.iLogLevelEnabled = LOG_WARN;
} else if (1 == opt_verbose->count) {
rLog.iLogLevelEnabled = LOG_INFO;
} else if (2 == opt_verbose->count) {
rLog.iLogLevelEnabled = LOG_VERBOSE;
} else if (3 == opt_verbose->count) {
rLog.iLogLevelEnabled = LOG_DEBUG;
}
user_opts->aln_opts.bAutoOptions = opt_autooptions->count;
user_opts->bDealignInputSeqs = opt_dealign->count;
/* NOTE: full distance matrix used to be default - there was
--mbed flag but no --full flag after r250 decided that mBed
should be default - now need --full flag to turn off mBed.
wanted to retain mBed Boolean, so simply added --full flag. if
both flags set (erroneously) want --mbed to overwrite --full,
therefore do --full 1st, the --mbed, FS, r250 */
if (opt_full->count){
user_opts->aln_opts.bUseMbed = FALSE;
}
if (opt_full_iter->count){
user_opts->aln_opts.bUseMbedForIteration = FALSE;
}
user_opts->bForceFileOverwrite = opt_force->count;
/* log-file
*/
if (opt_logfile->count > 0) {
user_opts->pcLogFile = CkStrdup(opt_logfile->filename[0]);
/* warn if already exists or not writable */
if (FileExists(user_opts->pcLogFile) && ! user_opts->bForceFileOverwrite) {
Log(&rLog, LOG_FATAL, "%s '%s'. %s",
"Cowardly refusing to overwrite already existing file",
user_opts->pcLogFile,
"Use --force to force overwriting.");
}
if (! FileIsWritable(user_opts->pcLogFile)) {
Log(&rLog, LOG_FATAL, "Sorry, I do not have permission to write to file '%s'.",
user_opts->pcLogFile);
}
}
/* normal sequence input (no profile)
*/
if (opt_seqin->count > 0) {
user_opts->pcSeqInfile = CkStrdup(opt_seqin->filename[0]);
}
/* Input limitations
*/
/* maximum number of sequences */
if (opt_max_seq->count > 0) {
user_opts->iMaxNumSeq = opt_max_seq->ival[0];
}
/* maximum sequence length */
if (opt_max_seqlen->count > 0) {
user_opts->iMaxSeqLen = opt_max_seqlen->ival[0];
}
/* Output format
*/
if (opt_infmt->count > 0) {
/* avoid gcc warning about discarded qualifier */
char *tmp = (char *)opt_infmt->sval[0];
user_opts->iSeqInFormat = String2SeqfileFormat(tmp);
} else {
user_opts->iSeqInFormat = SQFILE_UNKNOWN;
}
/* Sequence type
*/
if (opt_seqtype->count > 0) {
if (STR_NC_EQ(opt_seqtype->sval[0], "protein")) {
user_opts->iSeqType = SEQTYPE_PROTEIN;
} else if (STR_NC_EQ(opt_seqtype->sval[0], "rna")) {
user_opts->iSeqType = SEQTYPE_RNA;
} else if (STR_NC_EQ(opt_seqtype->sval[0], "dna")) {
user_opts->iSeqType = SEQTYPE_DNA;
} else {
Log(&rLog, LOG_FATAL, "Unknown sequence type '%s'", opt_seqtype->sval[0]);
}
}
/* if (opt_force_protein->count > 0) {
user_opts->iSeqType = SEQTYPE_PROTEIN;
} */
/* Profile input
*/
if (opt_profile1->count > 0) {
user_opts->pcProfile1Infile = CkStrdup(opt_profile1->filename[0]);
if (! FileExists(user_opts->pcProfile1Infile)) {
Log(&rLog, LOG_FATAL, "File '%s' does not exist.", user_opts->pcProfile1Infile);
}
}
if (opt_profile2->count > 0) {
user_opts->pcProfile2Infile = CkStrdup(opt_profile2->filename[0]);
if (! FileExists(user_opts->pcProfile2Infile)) {
Log(&rLog, LOG_FATAL, "File '%s' does not exist.", user_opts->pcProfile2Infile);
}
}
/* HMM input
*/
user_opts->aln_opts.iHMMInputFiles = 0;
user_opts->aln_opts.ppcHMMInput = NULL;
if (opt_hmm_in->count>0) {
int iAux;
user_opts->aln_opts.iHMMInputFiles = opt_hmm_in->count;
user_opts->aln_opts.ppcHMMInput = (char **) CKMALLOC(
user_opts->aln_opts.iHMMInputFiles * sizeof(char*));
for (iAux=0; iAux<opt_hmm_in->count; iAux++) {
user_opts->aln_opts.ppcHMMInput[iAux] = CkStrdup(opt_hmm_in->filename[iAux]);
if (! FileExists(user_opts->aln_opts.ppcHMMInput[iAux])) {
Log(&rLog, LOG_FATAL, "File '%s' does not exist.", user_opts->aln_opts.ppcHMMInput[iAux]);
}
}
}
/* Pair distance method
*/
if (opt_pairdist->count > 0) {
if (STR_NC_EQ(opt_pairdist->sval[0], "ktuple")) {
user_opts->aln_opts.iPairDistType = PAIRDIST_KTUPLE;
} else {
Log(&rLog, LOG_FATAL, "Unknown pairdist method '%s'", opt_pairdist->sval[0]);
}
}
/* Distance matrix input
*/
if (opt_distmat_in->count > 0) {
user_opts->aln_opts.pcDistmatInfile = CkStrdup(opt_distmat_in->filename[0]);
if (! FileExists(user_opts->aln_opts.pcDistmatInfile)) {
Log(&rLog, LOG_FATAL, "File '%s' does not exist.", user_opts->aln_opts.pcDistmatInfile);
}
}
/* Distance matrix output
*/
if (opt_distmat_out->count > 0) {
user_opts->aln_opts.pcDistmatOutfile = CkStrdup(opt_distmat_out->filename[0]);
/* warn if already exists or not writable */
if (FileExists(user_opts->aln_opts.pcDistmatOutfile) && ! user_opts->bForceFileOverwrite) {
Log(&rLog, LOG_FATAL, "%s '%s'. %s",
"Cowardly refusing to overwrite already existing file",
user_opts->aln_opts.pcDistmatOutfile,
"Use --force to force overwriting.");
}
if (! FileIsWritable(user_opts->aln_opts.pcDistmatOutfile)) {
Log(&rLog, LOG_FATAL, "Sorry, I do not have permission to write to file '%s'.",
user_opts->aln_opts.pcDistmatOutfile);
}
}
/* Clustering
*
*/
if (opt_clustering->count > 0) {
if (STR_NC_EQ(opt_clustering->sval[0], "upgma")) {
user_opts->aln_opts.iClusteringType = CLUSTERING_UPGMA;
} else {
Log(&rLog, LOG_FATAL, "Unknown guide-tree clustering method '%s'", opt_clustering->sval[0]);
}
}
/* Guidetree input
*/
if (opt_guidetree_in->count > 0) {
user_opts->aln_opts.pcGuidetreeInfile = CkStrdup(opt_guidetree_in->filename[0]);
if (! FileExists(user_opts->aln_opts.pcGuidetreeInfile)) {
Log(&rLog, LOG_FATAL, "File '%s' does not exist.", user_opts->aln_opts.pcGuidetreeInfile);
}
}
/* Guidetree output
*/
if (opt_guidetree_out->count > 0) {
user_opts->aln_opts.pcGuidetreeOutfile = CkStrdup(opt_guidetree_out->filename[0]);
/* warn if already exists or not writable */
if (FileExists(user_opts->aln_opts.pcGuidetreeOutfile) && ! user_opts->bForceFileOverwrite) {
Log(&rLog, LOG_FATAL, "%s '%s'. %s",
"Cowardly refusing to overwrite already existing file",
user_opts->aln_opts.pcGuidetreeOutfile,
"Use --force to force overwriting.");
}
if (! FileIsWritable(user_opts->aln_opts.pcGuidetreeOutfile)) {
Log(&rLog, LOG_FATAL, "Sorry, I do not have permission to write to file '%s'.",
user_opts->aln_opts.pcGuidetreeOutfile);
}
}
/* max guidetree iterations
*/
if (opt_max_guidetree_iterations->count > 0) {
user_opts->aln_opts.iMaxGuidetreeIterations = opt_max_guidetree_iterations->ival[0];
}
/* max guidetree iterations
*/
if (opt_max_hmm_iterations->count > 0) {
user_opts->aln_opts.iMaxHMMIterations = opt_max_hmm_iterations->ival[0];
}
/* number of iterations
*/
if (opt_num_iterations->count > 0) {
if (STR_NC_EQ(opt_num_iterations->sval[0], "auto")) {
Log(&rLog, LOG_FATAL, "Automatic iteration not supported at the moment.");
user_opts->aln_opts.bIterationsAuto = TRUE;
} else {
int iAux;
user_opts->aln_opts.bIterationsAuto = FALSE;
for (iAux=0; iAux<(int)strlen(opt_num_iterations->sval[0]); iAux++) {
if (! isdigit(opt_num_iterations->sval[0][iAux])) {
Log(&rLog, LOG_FATAL, "Couldn't iteration parameter: %s",
opt_num_iterations->sval[0]);
}
}
user_opts->aln_opts.iNumIterations = atoi(opt_num_iterations->sval[0]);
}
}
/* Alignment output
*/
if (opt_outfile->count > 0) {
user_opts->pcAlnOutfile = CkStrdup(opt_outfile->filename[0]);
/* warn if already exists or not writable */
if (FileExists(user_opts->pcAlnOutfile) && ! user_opts->bForceFileOverwrite) {
Log(&rLog, LOG_FATAL, "%s '%s'. %s",
"Cowardly refusing to overwrite already existing file",
user_opts->pcAlnOutfile,
"Use --force to force overwriting.");
}
if (! FileIsWritable(user_opts->pcAlnOutfile)) {
Log(&rLog, LOG_FATAL, "Sorry, I do not have permission to write to file '%s'.",
user_opts->pcAlnOutfile);
}
}
/* Output format
*/
if (opt_outfmt->count > 0) {
/* avoid gcc warning about discarded qualifier */
char *tmp = (char *)opt_outfmt->sval[0];
user_opts->iAlnOutFormat = String2SeqfileFormat(tmp);
if (SQFILE_UNKNOWN == user_opts->iAlnOutFormat) {
Log(&rLog, LOG_FATAL, "Unknown output format '%s'", opt_outfmt->sval[0]);
}
}
/* Number of threads
*/
#ifdef HAVE_OPENMP
if (opt_threads->count > 0) {
if (opt_threads->ival[0] <= 0) {
Log(&rLog, LOG_FATAL, "Changing number of threads to %d doesn't make sense.",
opt_threads->ival[0]);
}
user_opts->iThreads = opt_threads->ival[0];
}
#else
if (opt_threads->count > 0) {
if (opt_threads->ival[0] > 1) {
Log(&rLog, LOG_FATAL, "Cannot change number of threads to %d. %s was build without OpenMP support.",
opt_threads->ival[0], PACKAGE_NAME);
}
}
#endif
/* max MAC RAM (maximum amount of RAM set aside for MAC algorithm)
*/
if (opt_macram->count > 0) { /* FS, r240 -> r241 */
user_opts->aln_opts.rHhalignPara.iMacRamMB = opt_macram->ival[0];
}
arg_freetable(argtable,sizeof(argtable)/sizeof(argtable[0]));
UserOptsLogicCheck(user_opts);
return;
}
