int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* command line processing */
ESL_ALPHABET *abc = NULL; /* sequence alphabet */
ESL_RANDOMNESS *r = NULL; /* source of randomness */
char *hmmfile = NULL; /* file to read HMM(s) from */
P7_HMMFILE *hfp = NULL; /* open hmmfile */
P7_HMM *hmm = NULL; /* HMM to emit from */
FILE *ofp = NULL; /* output stream */
int outfmt = 0;
int nhmms = 0;
int status;
char errbuf[eslERRBUFSIZE];
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h")) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
if ((hmmfile = esl_opt_GetArg(go, 1)) == NULL) cmdline_failure(argv[0], "Failed to get <hmmfile> on cmdline: %s\n", go->errbuf);
if ( esl_opt_IsOn(go, "-o") ) {
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL) esl_fatal("Failed to open output file %s", esl_opt_GetString(go, "-o"));
} else ofp = stdout;
if (esl_opt_GetBoolean(go, "-a")) outfmt = eslMSAFILE_STOCKHOLM;
else outfmt = eslSQFILE_FASTA;
r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "--seed"));
status = p7_hmmfile_OpenE(hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, hmmfile, errbuf);
while ((status = p7_hmmfile_Read(hfp, &abc, &hmm)) != eslEOF)
{
if (status == eslEFORMAT) esl_fatal("Bad file format in HMM file %s:\n%s\n", hfp->fname, hfp->errbuf);
else if (status == eslEINCOMPAT) esl_fatal("HMM in %s is not in the expected %s alphabet\n", hfp->fname, esl_abc_DecodeType(abc->type));
else if (status != eslOK) esl_fatal("Unexpected error in reading HMMs from %s\n", hfp->fname);
nhmms++;
if (esl_opt_GetBoolean(go, "-c")) emit_consensus(go, ofp, outfmt, hmm);
else if (esl_opt_GetBoolean(go, "-C")) emit_fancycons(go, ofp, outfmt, hmm);
else if (esl_opt_GetBoolean(go, "-a")) emit_alignment(go, ofp, outfmt, r, hmm);
else emit_sequences(go, ofp, outfmt, r, hmm);
p7_hmm_Destroy(hmm);
}
if (nhmms == 0) esl_fatal("Empty HMM file %s? No HMM data found.\n");
if (esl_opt_IsOn(go, "-o")) { fclose(ofp); }
esl_randomness_Destroy(r);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
p7_hmmfile_Close(hfp);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
char *msafile = esl_opt_GetArg(go, 1);
int infmt = eslMSAFILE_UNKNOWN;
int outfmt = eslMSAFILE_UNKNOWN;
ESL_ALPHABET *abc = NULL;
ESL_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
int nali = 0;
int status;
/* Alphabet specification from cmdline? */
if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
/* MSA input file format from cmdline? */
if (esl_opt_IsOn(go, "--informat") &&
(infmt = esl_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"))) == eslMSAFILE_UNKNOWN)
esl_fatal("Your --informat, %s, is not a recognized multiple alignment file format",
esl_opt_GetString(go, "--informat"));
/* Open in digital mode. Autoguess alphabet, format if we haven't set them already. */
if (( status = esl_msafile_Open(&abc, msafile, NULL, infmt, NULL, &afp)) != eslOK)
esl_msafile_OpenFailure(afp, status);
/* Reverse complementation only makes sense for alphabets that have abc->complement set */
if (! abc->complement)
esl_fatal("File %s appears to use the %s alphabet.\nThat alphabet cannot be reverse complemented.\n",
msafile, esl_abc_DecodeType(abc->type));
/* Set the output format, if requested.
* By default, write in the same format we read in.
* Remember, it's afp->format that gets set; infmt was only a hint.
*/
if ( esl_opt_IsOn(go, "--outformat") )
{
outfmt = esl_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN)
esl_fatal("Your --outformat, %s, is not a recognized multiple alignment file format.\n",
esl_opt_GetString(go, "--outformat"));
}
else outfmt = afp->format;
/* Here we go. */
while ((status = esl_msafile_Read(afp, &msa)) == eslOK)
{
nali++;
status = esl_msa_ReverseComplement(msa);
esl_msafile_Write(stdout, msa, outfmt);
esl_msa_Destroy(msa);
}
if (nali == 0) esl_fatal("No alignments found in input file %s\n", msafile);
if (status != eslEOF) esl_msafile_ReadFailure(afp, status);
esl_msafile_Close(afp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return 0;
}
/* Function: main()
* Synopsis: break input sequence set into chunks, for each one building the
* Burrows-Wheeler transform and corresponding FM-index. Maintain requisite
* meta data.
* Notes: Currently depends on the divsufsort-lite code of Yuta Mori, though this
* could easily be replaced.
*/
int
main(int argc, char **argv)
{
char tmp_filename[16] = "fmtmpXXXXXX";
FILE *fptmp = NULL;
FILE *fp = NULL;
uint8_t *T = NULL;
uint8_t *BWT = NULL;
int *SA = NULL; //what I write will be 32-bit ints, but I need to keep this as int so it'll work with libdivsufsort
uint32_t *SAsamp = NULL;
uint32_t *occCnts_sb = NULL; // same indexing as above
uint32_t *cnts_sb = NULL;
uint16_t *occCnts_b = NULL; // this is logically a 2D array, but will be indexed as occ_cnts[alph_size*index + char] (instead of occ_cnts[index][char])
uint16_t *cnts_b = NULL;
FM_METADATA *meta = NULL;
clock_t t1, t2;
struct tms ts1, ts2;
long i,j,c;
int status = eslOK;
int chars_per_byte;
int num_freq_cnts_sb ;
int num_freq_cnts_b ;
int num_SA_samples ;
int infmt = eslSQFILE_UNKNOWN;
int alphatype = eslUNKNOWN;
int alphaguess =eslUNKNOWN;
ESL_ALPHABET *abc = NULL;
ESL_SQ *sq = NULL;
ESL_SQFILE *sqfp = NULL;
ESL_SQ *tmpsq = NULL;
ESL_SQ_BLOCK *block = NULL;
char *fname_in = NULL;
char *fname_out= NULL;
int block_size = 50000000;
int sq_cnt = 0;
int use_tmpsq = 0;
uint64_t block_length;
uint64_t total_char_count = 0;
int max_block_size;
int numblocks = 0;
uint32_t numseqs = 0;
int allocedseqs = 1000;
uint32_t seq_offset = 0;
uint32_t ambig_offset = 0;
uint32_t overlap = 0;
uint16_t seq_cnt;
uint16_t ambig_cnt;
uint32_t prev_numseqs = 0;
int compressed_bytes;
uint32_t term_loc;
ESL_GETOPTS *go = NULL; /* command line processing */
uint8_t ambig_repl = 0;
int in_ambig_run = 0;
FM_AMBIGLIST ambig_list;
ESL_ALLOC (meta, sizeof(FM_METADATA));
if (meta == NULL)
esl_fatal("unable to allocate memory to store FM meta data\n");
ESL_ALLOC (meta->ambig_list, sizeof(FM_AMBIGLIST));
if (meta->ambig_list == NULL)
esl_fatal("unable to allocate memory to store FM ambiguity data\n");
fm_initAmbiguityList(meta->ambig_list);
meta->alph_type = fm_DNA;
meta->freq_SA = 8;
meta->freq_cnt_b = 256;
meta->freq_cnt_sb = pow(2,16); //65536 - that's the # values in a short
meta->seq_count = 0;
ESL_ALLOC (meta->seq_data, allocedseqs * sizeof(FM_SEQDATA));
if (meta->seq_data == NULL )
esl_fatal("unable to allocate memory to store FM sequence data\n");
process_commandline(argc, argv, &go, &fname_in, &fname_out);
if (esl_opt_IsOn(go, "--bin_length")) meta->freq_cnt_b = esl_opt_GetInteger(go, "--bin_length");
if ( meta->freq_cnt_b < 32 || meta->freq_cnt_b >4096 || (meta->freq_cnt_b & (meta->freq_cnt_b - 1)) ) // test power of 2
esl_fatal("bin_length must be a power of 2, at least 128, and at most 4096\n");
if (esl_opt_IsOn(go, "--sa_freq")) meta->freq_SA = esl_opt_GetInteger(go, "--sa_freq");
if ( (meta->freq_SA & (meta->freq_SA - 1)) ) // test power of 2
esl_fatal ("SA_freq must be a power of 2\n");
if (esl_opt_IsOn(go, "--block_size")) block_size = 1000000 * esl_opt_GetInteger(go, "--block_size");
if ( block_size <=0 )
esl_fatal ("block_size must be a positive number\n");
//start timer
t1 = times(&ts1);
output_header(stdout, go, fname_in, fname_out);
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(fname_in, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", fname_in);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", fname_in);
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
meta->fwd_only = 0;
if (esl_opt_IsUsed(go, "--alph")) {
meta->alph = esl_opt_GetString(go, "--alph") ;
if ( esl_strcmp(meta->alph, "dna")==0 || esl_strcmp(meta->alph, "rna")==0) {
meta->alph_type = fm_DNA;
alphatype = eslDNA;
} else if (esl_strcmp(meta->alph, "dna_full")==0 || esl_strcmp(meta->alph, "rna_full")==0) {
meta->alph_type = fm_DNA_full;
alphatype = eslDNA;
} else if (esl_strcmp(meta->alph, "amino")==0) {
meta->alph_type = fm_AMINO;
alphatype = eslAMINO;
meta->fwd_only = 1;
} else {
esl_fatal("Unknown alphabet type. Try 'dna', 'dna_full', or 'amino'\n%s", "");
}
} else {
esl_sqfile_GuessAlphabet(sqfp, &alphaguess);
if (alphaguess == eslDNA || alphaguess == eslRNA) {
meta->alph_type = fm_DNA;
alphatype = eslDNA;
} else if (alphaguess == eslAMINO) {
meta->alph_type = fm_AMINO;
alphatype = eslAMINO;
meta->fwd_only = 1;
} else {
esl_fatal("Unknown alphabet type. Try 'dna', 'dna_full', or 'amino'\n%s", "");
}
}
if (esl_opt_IsOn(go, "--fwd_only") )
meta->fwd_only = 1;
meta->alph = NULL;
//getInverseAlphabet
fm_alphabetCreate(meta, &(meta->charBits));
chars_per_byte = 8/meta->charBits;
//shift inv_alph up one, to make space for '$' at 0
for (i=0; i<256; i++)
if ( meta->inv_alph[i] >= 0)
meta->inv_alph[i]++;
abc = esl_alphabet_Create(alphatype);
sq = esl_sq_CreateDigital(abc);
tmpsq = esl_sq_CreateDigital(abc);
esl_sqfile_SetDigital(sqfp, abc);
block = esl_sq_CreateDigitalBlock(FM_BLOCK_COUNT, abc);
block->complete = FALSE;
// max_block_size = FM_BLOCK_OVERLAP+block_size+1 + block_size*.2; // +1 for the '$'
max_block_size = FM_BLOCK_OVERLAP+block_size+1 + block_size; // temporary hack to avoid memory over-runs (see end of 1101_fmindex_benchmarking/00NOTES)
if (alphatype == fm_DNA)
fm_initAmbiguityList(&ambig_list);
/* Allocate BWT, Text, SA, and FM-index data structures, allowing storage of maximally large sequence*/
ESL_ALLOC (T, max_block_size * sizeof(uint8_t));
ESL_ALLOC (BWT, max_block_size * sizeof(uint8_t));
ESL_ALLOC (SA, max_block_size * sizeof(int));
ESL_ALLOC (SAsamp, (1+floor((double)max_block_size/meta->freq_SA) ) * sizeof(uint32_t));
ESL_ALLOC (occCnts_sb, (1+ceil((double)max_block_size/meta->freq_cnt_sb)) * meta->alph_size * sizeof(uint32_t)); // every freq_cnt_sb positions, store an array of ints
ESL_ALLOC (cnts_sb, meta->alph_size * sizeof(uint32_t));
ESL_ALLOC (occCnts_b, ( 1+ceil((double)max_block_size/meta->freq_cnt_b)) * meta->alph_size * sizeof(uint16_t)); // every freq_cnt_b positions, store an array of 8-byte ints
ESL_ALLOC (cnts_b, meta->alph_size * sizeof(uint16_t));
if((T == NULL) || (BWT == NULL) || (SA==NULL) || (SAsamp==NULL) || (BWT==NULL) || (cnts_b==NULL) || (occCnts_b==NULL) || (cnts_sb==NULL) || (occCnts_sb==NULL) ) {
esl_fatal( "%s: Cannot allocate memory.\n", argv[0]);
}
// Open a temporary file, to which FM-index data will be written
if (esl_tmpfile(tmp_filename, &fptmp) != eslOK) esl_fatal("unable to open fm-index tmpfile");
/* Main loop: */
while (status == eslOK ) {
//reset block as an empty vessel
for (i=0; i<block->count; i++)
esl_sq_Reuse(block->list + i);
if (use_tmpsq) {
esl_sq_Copy(tmpsq , block->list);
block->complete = FALSE; //this lets ReadBlock know that it needs to append to a small bit of previously-read seqeunce
block->list->C = FM_BLOCK_OVERLAP; // overload the ->C value, which ReadBlock uses to determine how much
// overlap should be retained in the ReadWindow step
} else {
block->complete = TRUE;
}
status = esl_sqio_ReadBlock(sqfp, block, block_size, -1, alphatype != eslAMINO);
if (status == eslEOF) continue;
if (status != eslOK) ESL_XEXCEPTION(status, "failure reading sequence block");
seq_offset = numseqs;
ambig_offset = meta->ambig_list->count;
if (block->complete || block->count == 0) {
use_tmpsq = FALSE;
} else {
/* The final sequence on the block was a probably-incomplete window of the active sequence.
* Grab a copy of the end for use in the next pass, to ensure we don't miss hits crossing
* the boundary between two blocks.
*/
esl_sq_Copy(block->list + (block->count - 1) , tmpsq);
use_tmpsq = TRUE;
}
block->first_seqidx = sq_cnt;
sq_cnt += block->count - (use_tmpsq ? 1 : 0);// if there's an incomplete sequence read into the block wait to count it until it's complete.
/* Read dseqs from block into text element T.
* Convert the dsq from esl-alphabet to fm-alphabet (1..k for alphabet of size k).
* (a) collapsing upper/lower case for appropriate sorting.
* (b) reserving 0 for '$', which must be lexicographically smallest
* (these will later be shifted to 0-based alphabet, once SA has been built)
*
*/
block_length = 0;
for (i=0; i<block->count; i++) {
//start a new block, with space for the name
allocateSeqdata(meta, block->list+i, numseqs, &allocedseqs);
//meta data
meta->seq_data[numseqs].target_id = block->first_seqidx + i ;
meta->seq_data[numseqs].target_start = block->list[i].start;
meta->seq_data[numseqs].fm_start = block_length;
if (block->list[i].name == NULL) meta->seq_data[numseqs].name[0] = '\0';
else strcpy(meta->seq_data[numseqs].name, block->list[i].name );
if (block->list[i].acc == NULL) meta->seq_data[numseqs].acc[0] = '\0';
else strcpy(meta->seq_data[numseqs].acc, block->list[i].acc );
if (block->list[i].source == NULL) meta->seq_data[numseqs].source[0] = '\0';
else strcpy(meta->seq_data[numseqs].source, block->list[i].source );
if (block->list[i].desc == NULL) meta->seq_data[numseqs].desc[0] = '\0';
else strcpy(meta->seq_data[numseqs].desc, block->list[i].desc );
for (j=1; j<=block->list[i].n; j++) {
c = abc->sym[block->list[i].dsq[j]];
if ( meta->alph_type == fm_DNA) {
if (meta->inv_alph[c] == -1) {
// replace ambiguity characters by rotating through A,C,G, and T.
c = meta->alph[ambig_repl];
ambig_repl = (ambig_repl+1)%4;
if (!in_ambig_run) {
fm_addAmbiguityRange(meta->ambig_list, block_length, block_length);
in_ambig_run=1;
} else {
meta->ambig_list->ranges[meta->ambig_list->count - 1].upper = block_length;
}
} else {
in_ambig_run=0;
}
} else if (meta->inv_alph[c] == -1) {
esl_fatal("requested alphabet doesn't match input text\n");
}
T[block_length] = meta->inv_alph[c];
block_length++;
if (j>block->list[i].C) total_char_count++; // add to total count, only if it's not redundant with earlier read
meta->seq_data[numseqs].length++;
}
numseqs++;
}
T[block_length] = 0; // last character 0 is effectively '$' for suffix array
block_length++;
seq_cnt = numseqs-seq_offset;
ambig_cnt = meta->ambig_list->count - ambig_offset;
//build and write FM-index for T. This will be a BWT on the reverse of the sequence, required for reverse-traversal of the BWT
buildAndWriteFMIndex(meta, seq_offset, ambig_offset, seq_cnt, ambig_cnt, (uint32_t)block->list[0].C, T, BWT, SA, SAsamp,
occCnts_sb, cnts_sb, occCnts_b, cnts_b, block_length, fptmp);
if ( ! meta->fwd_only ) {
//build and write FM-index for un-reversed T (used to find reverse hits using forward traversal of the BWT
buildAndWriteFMIndex(meta, seq_offset, ambig_offset, seq_cnt, ambig_cnt, 0, T, BWT, SA, NULL,
occCnts_sb, cnts_sb, occCnts_b, cnts_b, block_length, fptmp);
}
prev_numseqs = numseqs;
numblocks++;
}
esl_sqfile_Close(sqfp);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
esl_sq_Destroy(tmpsq);
esl_sq_DestroyBlock(block);
meta->seq_count = numseqs;
meta->block_count = numblocks;
/* Finished writing the FM-index data to a temporary file. Now write
* metadata to fname_out, than append FM-index data from temp file
*/
if((fp = fopen(fname_out, "wb")) == NULL)
esl_fatal( "%s: Cannot open file `%s': ", argv[0], fname_out);
//write out meta data
if( fwrite(&(meta->fwd_only), sizeof(meta->fwd_only), 1, fp) != 1 ||
fwrite(&(meta->alph_type), sizeof(meta->alph_type), 1, fp) != 1 ||
fwrite(&(meta->alph_size), sizeof(meta->alph_size), 1, fp) != 1 ||
fwrite(&(meta->charBits), sizeof(meta->charBits), 1, fp) != 1 ||
fwrite(&(meta->freq_SA), sizeof(meta->freq_SA), 1, fp) != 1 ||
fwrite(&(meta->freq_cnt_sb), sizeof(meta->freq_cnt_sb), 1, fp) != 1 ||
fwrite(&(meta->freq_cnt_b), sizeof(meta->freq_cnt_b), 1, fp) != 1 ||
fwrite(&(meta->block_count), sizeof(meta->block_count), 1, fp) != 1 ||
fwrite(&(meta->seq_count), sizeof(meta->seq_count), 1, fp) != 1 ||
fwrite(&(meta->ambig_list->count), sizeof(meta->ambig_list->count), 1, fp) != 1 ||
fwrite(&total_char_count, sizeof(total_char_count), 1, fp) != 1
)
esl_fatal( "%s: Error writing meta data for FM index.\n", argv[0]);
for (i=0; i<meta->seq_count; i++) {
if( fwrite(&(meta->seq_data[i].target_id), sizeof(meta->seq_data[i].target_id), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].target_start), sizeof(meta->seq_data[i].target_start), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].fm_start), sizeof(meta->seq_data[i].fm_start), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].length), sizeof(meta->seq_data[i].length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].name_length), sizeof(meta->seq_data[i].name_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].acc_length), sizeof(meta->seq_data[i].acc_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].source_length),sizeof(meta->seq_data[i].source_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].desc_length), sizeof(meta->seq_data[i].desc_length), 1, fp) != 1 ||
fwrite(meta->seq_data[i].name, sizeof(char), meta->seq_data[i].name_length+1 , fp) != meta->seq_data[i].name_length+1 ||
fwrite(meta->seq_data[i].acc, sizeof(char), meta->seq_data[i].acc_length+1 , fp) != meta->seq_data[i].acc_length+1 ||
fwrite(meta->seq_data[i].source, sizeof(char), meta->seq_data[i].source_length+1, fp) != meta->seq_data[i].source_length+1 ||
fwrite(meta->seq_data[i].desc, sizeof(char), meta->seq_data[i].desc_length+1 , fp) != meta->seq_data[i].desc_length+1
)
esl_fatal( "%s: Error writing meta data for FM index.\n", argv[0]);
}
for (i=0; i<meta->ambig_list->count; i++) {
if( fwrite(&(meta->ambig_list->ranges[i].lower), sizeof(meta->ambig_list->ranges[i].lower), 1, fp) != 1 ||
fwrite(&(meta->ambig_list->ranges[i].upper), sizeof(meta->ambig_list->ranges[i].upper), 1, fp) != 1
)
esl_fatal( "%s: Error writing ambiguity data for FM index.\n", argv[0]);
}
/* now append the FM-index data in fptmp to the desired output file, fp */
rewind(fptmp);
for (i=0; i<numblocks; i++) {
for(j=0; j< (meta->fwd_only?1:2); j++ ) { //do this once or twice, once for forward-T index, and possibly once for reversed
//first, read
if(fread(&block_length, sizeof(block_length), 1, fptmp) != 1)
esl_fatal( "%s: Error reading block_length in FM index.\n", argv[0]);
if(fread(&term_loc, sizeof(term_loc), 1, fptmp) != 1)
esl_fatal( "%s: Error reading terminal location in FM index.\n", argv[0]);
if(fread(&seq_offset, sizeof(seq_offset), 1, fptmp) != 1)
esl_fatal( "%s: Error reading seq_offset in FM index.\n", argv[0]);
if(fread(&ambig_offset, sizeof(ambig_offset ), 1, fptmp) != 1)
esl_fatal( "%s: Error reading ambig_offset in FM index.\n", argv[0]);
if(fread(&overlap, sizeof(overlap), 1, fptmp) != 1)
esl_fatal( "%s: Error reading overlap in FM index.\n", argv[0]);
if(fread(&seq_cnt, sizeof(seq_cnt), 1, fptmp) != 1)
esl_fatal( "%s: Error reading seq_cnt in FM index.\n", argv[0]);
if(fread(&ambig_cnt, sizeof(ambig_cnt), 1, fptmp) != 1)
esl_fatal( "%s: Error reading ambig_cnt in FM index.\n", argv[0]);
compressed_bytes = ((chars_per_byte-1+block_length)/chars_per_byte);
num_freq_cnts_b = 1+ceil((double)block_length/meta->freq_cnt_b);
num_freq_cnts_sb = 1+ceil((double)block_length/meta->freq_cnt_sb);
num_SA_samples = 1+floor((double)block_length/meta->freq_SA);
//j==0 test cause T and SA to be written only for forward sequence
if(j==0 && fread(T, sizeof(uint8_t), compressed_bytes, fptmp) != compressed_bytes)
esl_fatal( "%s: Error reading T in FM index.\n", argv[0]);
if(fread(BWT, sizeof(uint8_t), compressed_bytes, fptmp) != compressed_bytes)
esl_fatal( "%s: Error reading BWT in FM index.\n", argv[0]);
if(j==0 && fread(SAsamp, sizeof(uint32_t), (size_t)num_SA_samples, fptmp) != (size_t)num_SA_samples)
esl_fatal( "%s: Error reading SA in FM index.\n", argv[0]);
if(fread(occCnts_b, sizeof(uint16_t)*(meta->alph_size), (size_t)num_freq_cnts_b, fptmp) != (size_t)num_freq_cnts_b)
esl_fatal( "%s: Error reading occCnts_b in FM index.\n", argv[0]);
if(fread(occCnts_sb, sizeof(uint32_t)*(meta->alph_size), (size_t)num_freq_cnts_sb, fptmp) != (size_t)num_freq_cnts_sb)
esl_fatal( "%s: Error reading occCnts_sb in FM index.\n", argv[0]);
//then, write
if(fwrite(&block_length, sizeof(block_length), 1, fp) != 1)
esl_fatal( "%s: Error writing block_length in FM index.\n", argv[0]);
if(fwrite(&term_loc, sizeof(term_loc), 1, fp) != 1)
esl_fatal( "%s: Error writing terminal location in FM index.\n", argv[0]);
if(fwrite(&seq_offset, sizeof(seq_offset), 1, fp) != 1)
esl_fatal( "%s: Error writing seq_offset in FM index.\n", argv[0]);
if(fwrite(&ambig_offset, sizeof(ambig_offset), 1, fp) != 1)
esl_fatal( "%s: Error writing ambig_offset in FM index.\n", argv[0]);
if(fwrite(&overlap, sizeof(overlap), 1, fp) != 1)
esl_fatal( "%s: Error writing overlap in FM index.\n", argv[0]);
if(fwrite(&seq_cnt, sizeof(seq_cnt), 1, fp) != 1)
esl_fatal( "%s: Error writing seq_cnt in FM index.\n", argv[0]);
if(fwrite(&ambig_cnt, sizeof(ambig_cnt), 1, fp) != 1)
esl_fatal( "%s: Error writing ambig_cnt in FM index.\n", argv[0]);
if(j==0 && fwrite(T, sizeof(uint8_t), compressed_bytes, fp) != compressed_bytes)
esl_fatal( "%s: Error writing T in FM index.\n", argv[0]);
if(fwrite(BWT, sizeof(uint8_t), compressed_bytes, fp) != compressed_bytes)
esl_fatal( "%s: Error writing BWT in FM index.\n", argv[0]);
if(j==0 && fwrite(SAsamp, sizeof(uint32_t), (size_t)num_SA_samples, fp) != (size_t)num_SA_samples)
esl_fatal( "%s: Error writing SA in FM index.\n", argv[0]);
if(fwrite(occCnts_b, sizeof(uint16_t)*(meta->alph_size), (size_t)num_freq_cnts_b, fp) != (size_t)num_freq_cnts_b)
esl_fatal( "%s: Error writing occCnts_b in FM index.\n", argv[0]);
if(fwrite(occCnts_sb, sizeof(uint32_t)*(meta->alph_size), (size_t)num_freq_cnts_sb, fp) != (size_t)num_freq_cnts_sb)
esl_fatal( "%s: Error writing occCnts_sb in FM index.\n", argv[0]);
}
}
fprintf (stderr, "Number of characters in index: %ld\n", (long)total_char_count);
fprintf (stderr, "Number of FM-index blocks: %ld\n", (long)meta->block_count);
fclose(fp);
fclose(fptmp);
free(T);
free(BWT);
free(SA);
free(SAsamp);
free(occCnts_b);
free(cnts_b);
free(occCnts_sb);
free(cnts_sb);
fm_metaDestroy(meta);
esl_getopts_Destroy(go);
// compute and print the elapsed time in millisec
t2 = times(&ts2);
{
double clk_ticks = sysconf(_SC_CLK_TCK);
double elapsedTime = (t2-t1)/clk_ticks;
fprintf (stderr, "run time: %.2f seconds\n", elapsedTime);
}
return (eslOK);
ERROR:
/* Deallocate memory. */
if (fp) fclose(fp);
if (T) free(T);
if (BWT) free(BWT);
if (SA) free(SA);
if (SAsamp) free(SAsamp);
if (occCnts_b) free(occCnts_b);
if (cnts_b) free(cnts_b);
if (occCnts_sb) free(occCnts_sb);
if (cnts_sb) free(cnts_sb);
if (ambig_list.ranges) free(ambig_list.ranges);
fm_metaDestroy(meta);
esl_getopts_Destroy(go);
esl_sqfile_Close(sqfp);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
if (tmpsq) esl_sq_Destroy(tmpsq);
if (block) esl_sq_DestroyBlock(block);
fprintf (stderr, "failure during memory allocation\n");
exit(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 = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
double mu = esl_opt_GetReal(go, "-m");
double lambda = esl_opt_GetReal(go, "-l");
double tau = esl_opt_GetReal(go, "-t");
ESL_RANDOMNESS *r = esl_randomness_Create(esl_opt_GetInteger(go, "-s"));
ESL_HISTOGRAM *h = esl_histogram_CreateFull(mu, 100., esl_opt_GetReal(go, "-w"));;
int n = esl_opt_GetInteger(go, "-n");
int be_verbose = esl_opt_GetBoolean(go, "-v");
char *plotfile = esl_opt_GetString(go, "-o");
FILE *pfp = stdout;
int plot_pdf = esl_opt_GetBoolean(go, "--P");
int plot_logpdf = esl_opt_GetBoolean(go, "--LP");
int plot_cdf = esl_opt_GetBoolean(go, "--C");
int plot_logcdf = esl_opt_GetBoolean(go, "--LC");
int plot_surv = esl_opt_GetBoolean(go, "--S");
int plot_logsurv = esl_opt_GetBoolean(go, "--LS");
double xmin = esl_opt_IsOn(go, "--XL") ? esl_opt_GetReal(go, "--XL") : mu;
double xmax = esl_opt_IsOn(go, "--XH") ? esl_opt_GetReal(go, "--XH") : mu+40*(1./lambda);
double xstep = esl_opt_IsOn(go, "--XS") ? esl_opt_GetReal(go, "--XS") : 0.1;
double emu, elambda, etau;
int i;
double x;
double *data;
int ndata;
if (be_verbose)
printf("Parametric: mu = %f lambda = %f tau = %f\n", mu, lambda, tau);
if (plotfile != NULL) {
if ((pfp = fopen(plotfile, "w")) == NULL)
esl_fatal("Failed to open plotfile");
}
for (i = 0; i < n; i++)
{
x = esl_sxp_Sample(r, mu, lambda, tau);
esl_histogram_Add(h, x);
}
esl_histogram_GetData(h, &data, &ndata);
esl_sxp_FitComplete(data, ndata, &emu, &elambda, &etau);
if (be_verbose)
printf("Complete data fit: mu = %f lambda = %f tau = %f\n",
emu, elambda, etau);
if (fabs( (emu-mu)/mu ) > 0.01) esl_fatal("Error in (complete) fitted mu > 1%\n");
if (fabs( (elambda-lambda)/lambda ) > 0.10) esl_fatal("Error in (complete) fitted lambda > 10%\n");
if (fabs( (etau-tau)/tau ) > 0.10) esl_fatal("Error in (complete) fitted tau > 10%\n");
esl_sxp_FitCompleteBinned(h, &emu, &elambda, &etau);
if (be_verbose)
printf("Binned data fit: mu = %f lambda = %f tau = %f\n",
emu, elambda, etau);
if (fabs( (emu-mu)/mu ) > 0.01) esl_fatal("Error in (binned) fitted mu > 1%\n");
if (fabs( (elambda-lambda)/lambda ) > 0.10) esl_fatal("Error in (binned) fitted lambda > 10%\n");
if (fabs( (etau-tau)/tau ) > 0.10) esl_fatal("Error in (binned) fitted tau > 10%\n");
if (plot_pdf) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_pdf, xmin, xmax, xstep);
if (plot_logpdf) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_logpdf, xmin, xmax, xstep);
if (plot_cdf) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_cdf, xmin, xmax, xstep);
if (plot_logcdf) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_logcdf, xmin, xmax, xstep);
if (plot_surv) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_surv, xmin, xmax, xstep);
if (plot_logsurv) esl_sxp_Plot(pfp, mu, lambda, tau, &esl_sxp_logsurv, xmin, xmax, xstep);
if (plotfile != NULL) fclose(pfp);
esl_histogram_Destroy(h);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
static int
output_header(FILE *ofp, ESL_GETOPTS *go, char *hmmfile, char *seqfile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# query sequence file: %s\n", seqfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# target HMM database: %s\n", hmmfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--tblout") && fprintf(ofp, "# per-seq hits tabular output: %s\n", esl_opt_GetString(go, "--tblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--dfamtblout") && fprintf(ofp, "# hits output in Dfam format: %s\n", esl_opt_GetString(go, "--dfamtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--aliscoresout") && fprintf(ofp, "# alignment scores output: %s\n", esl_opt_GetString(go, "--aliscoresout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--acc") && fprintf(ofp, "# prefer accessions over names: yes\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--noali") && fprintf(ofp, "# show alignments in output: no\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--notextw") && fprintf(ofp, "# max ASCII text line length: unlimited\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--textw") && fprintf(ofp, "# max ASCII text line length: %d\n", esl_opt_GetInteger(go, "--textw")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-E") && fprintf(ofp, "# profile reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "-E")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-T") && fprintf(ofp, "# profile reporting threshold: score >= %g\n", esl_opt_GetReal(go, "-T")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incE") && fprintf(ofp, "# profile inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incT") && fprintf(ofp, "# profile inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_ga") && fprintf(ofp, "# model-specific thresholding: GA cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_nc") && fprintf(ofp, "# model-specific thresholding: NC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_tc") && fprintf(ofp, "# model-specific thresholding: TC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--max") && fprintf(ofp, "# Max sensitivity mode: on [all heuristic filters off]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F1") && fprintf(ofp, "# MSV filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F2") && fprintf(ofp, "# Vit filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F3") && fprintf(ofp, "# Fwd filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nobias") && fprintf(ofp, "# biased composition HMM filter: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B1") && fprintf(ofp, "# biased comp MSV window len: %d\n", esl_opt_GetInteger(go, "--B1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B2") && fprintf(ofp, "# biased comp Viterbi window len: %d\n", esl_opt_GetInteger(go, "--B2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B3") && fprintf(ofp, "# biased comp Forward window len: %d\n", esl_opt_GetInteger(go, "--B3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nonull2") && fprintf(ofp, "# null2 bias corrections: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--toponly") && fprintf(ofp, "# search only top strand: on\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--bottomonly") && fprintf(ofp, "# search only bottom strand: on\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-Z") && fprintf(ofp, "# sequence search space set to: %.0f\n", esl_opt_GetReal(go, "-Z")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed")==0 && fprintf(ofp, "# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp, "# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--qformat") && fprintf(ofp, "# input seqfile format asserted: %s\n", esl_opt_GetString(go, "--qformat")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--w_beta") && fprintf(ofp, "# window length beta value: %g\n", esl_opt_GetReal(go, "--w_beta")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--w_length") && fprintf(ofp, "# window length : %d\n", esl_opt_GetInteger(go, "--w_length")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// if (esl_opt_IsUsed(go, "--daemon") && fprintf(ofp, "run as a daemon process\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
#ifdef HMMER_THREADS
if (esl_opt_IsUsed(go, "--cpu") && fprintf(ofp, "# number of worker threads: %d\n", esl_opt_GetInteger(go, "--cpu")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
#endif
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
static int
output_header(FILE *ofp, const ESL_GETOPTS *go, char *hmmfile, char *seqfile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# query HMM file: %s\n", hmmfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# target sequence database: %s\n", seqfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-A") && fprintf(ofp, "# MSA of all hits saved to file: %s\n", esl_opt_GetString(go, "-A")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--tblout") && fprintf(ofp, "# per-seq hits tabular output: %s\n", esl_opt_GetString(go, "--tblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domtblout") && fprintf(ofp, "# per-dom hits tabular output: %s\n", esl_opt_GetString(go, "--domtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--pfamtblout") && fprintf(ofp, "# pfam-style tabular hit output: %s\n", esl_opt_GetString(go, "--pfamtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--acc") && fprintf(ofp, "# prefer accessions over names: yes\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--noali") && fprintf(ofp, "# show alignments in output: no\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--notextw") && fprintf(ofp, "# max ASCII text line length: unlimited\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--textw") && fprintf(ofp, "# max ASCII text line length: %d\n", esl_opt_GetInteger(go, "--textw")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-E") && fprintf(ofp, "# sequence reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "-E")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-T") && fprintf(ofp, "# sequence reporting threshold: score >= %g\n", esl_opt_GetReal(go, "-T")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domE") && fprintf(ofp, "# domain reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "--domE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domT") && fprintf(ofp, "# domain reporting threshold: score >= %g\n", esl_opt_GetReal(go, "--domT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incE") && fprintf(ofp, "# sequence inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incT") && fprintf(ofp, "# sequence inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incdomE") && fprintf(ofp, "# domain inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incdomE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incdomT") && fprintf(ofp, "# domain inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incdomT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_ga") && fprintf(ofp, "# model-specific thresholding: GA cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_nc") && fprintf(ofp, "# model-specific thresholding: NC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_tc") && fprintf(ofp, "# model-specific thresholding: TC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--max") && fprintf(ofp, "# Max sensitivity mode: on [all heuristic filters off]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F1") && fprintf(ofp, "# MSV filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F2") && fprintf(ofp, "# Vit filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F3") && fprintf(ofp, "# Fwd filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nobias") && fprintf(ofp, "# biased composition HMM filter: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--restrictdb_stkey") && fprintf(ofp, "# Restrict db to start at seq key: %s\n", esl_opt_GetString(go, "--restrictdb_stkey")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--restrictdb_n") && fprintf(ofp, "# Restrict db to # target seqs: %d\n", esl_opt_GetInteger(go, "--restrictdb_n")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--ssifile") && fprintf(ofp, "# Override ssi file to: %s\n", esl_opt_GetString(go, "--ssifile")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nonull2") && fprintf(ofp, "# null2 bias corrections: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-Z") && fprintf(ofp, "# sequence search space set to: %.0f\n", esl_opt_GetReal(go, "-Z")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domZ") && fprintf(ofp, "# domain search space set to: %.0f\n", esl_opt_GetReal(go, "--domZ")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed") == 0 && fprintf(ofp, "# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp, "# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--tformat") && fprintf(ofp, "# targ <seqfile> format asserted: %s\n", esl_opt_GetString(go, "--tformat")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
int
main(int argc, char **argv)
{
int i,j;
ESL_GETOPTS *go = NULL; /* command line processing */
ESL_STOPWATCH *w = esl_stopwatch_Create();
int status;
ESL_MSA *msa = NULL;
FILE *ofp = NULL; /* output file (default is stdout) */
ESL_ALPHABET *abc = NULL; /* digital alphabet */
char *alifile; /* name of the alignment file we're building HMMs from */
ESLX_MSAFILE *afp = NULL; /* open alifile */
int infmt = eslMSAFILE_UNKNOWN; /* autodetect alignment format by default. */
int outfmt = eslMSAFILE_STOCKHOLM;
char *postmsafile; /* optional file to resave annotated, modified MSAs to */
FILE *postmsafp = NULL; /* open <postmsafile>, or NULL */
int mask_range_cnt = 0;
uint32_t mask_starts[100]; // over-the-top allocation.
uint32_t mask_ends[100];
char *rangestr;
char *range;
int *map = NULL; /* map[i]=j, means model position i comes from column j of the alignment; 1..alen */
int keep_mm;
/* Set processor specific flags */
impl_Init();
alifile = NULL;
postmsafile = NULL;
/* Parse the command line
*/
process_commandline(argc, argv, &go, &alifile, &postmsafile);
keep_mm = esl_opt_IsUsed(go, "--apendmask");
/* Initialize what we can in the config structure (without knowing the alphabet yet).
* Fields controlled by masters are set up in usual_master() or mpi_master()
* Fields used by workers are set up in mpi_worker()
*/
ofp = NULL;
infmt = eslMSAFILE_UNKNOWN;
afp = NULL;
abc = NULL;
if (esl_opt_IsOn(go, "--informat")) {
infmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslMSAFILE_UNKNOWN) p7_Fail("%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--informat"));
}
/* Determine output alignment file format */
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) p7_Fail(argv[0], "%s is not a recognized output MSA file format\n", esl_opt_GetString(go, "--outformat"));
/* Parse the ranges */
if (esl_opt_IsUsed(go, "--alirange")) {
esl_strdup(esl_opt_GetString(go, "--alirange"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--modelrange")) {
esl_strdup(esl_opt_GetString(go, "--modelrange"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--model2ali")) {
esl_strdup(esl_opt_GetString(go, "--model2ali"), -1, &rangestr) ;
} else if (esl_opt_IsUsed(go, "--ali2model")) {
esl_strdup(esl_opt_GetString(go, "--ali2model"), -1, &rangestr) ;
} else {
if (puts("Must specify mask range with --modelrange, --alirange, --model2ali, or --ali2model\n") < 0) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed"); goto ERROR;
}
while ( (status = esl_strtok(&rangestr, ",", &range) ) == eslOK) {
status = esl_regexp_ParseCoordString(range, mask_starts + mask_range_cnt, mask_ends + mask_range_cnt );
if (status == eslESYNTAX) esl_fatal("range flags take coords <from>..<to>; %s not recognized", range);
if (status == eslFAIL) esl_fatal("Failed to find <from> or <to> coord in %s", range);
mask_range_cnt++;
}
/* Start timing. */
esl_stopwatch_Start(w);
/* Open files, set alphabet.
* afp - open alignment file for input
* abc - alphabet expected or guessed in ali file
* postmsafp - open MSA output file
* ofp - optional open output file, or stdout
*/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
else abc = NULL;
status = eslx_msafile_Open(&abc, alifile, NULL, infmt, NULL, &afp);
if (status != eslOK) eslx_msafile_OpenFailure(afp, status);
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
postmsafp = fopen(postmsafile, "w");
if (postmsafp == NULL) p7_Fail("Failed to MSA output file %s for writing", postmsafile);
}
if (esl_opt_IsUsed(go, "-o"))
{
ofp = fopen(esl_opt_GetString(go, "-o"), "w");
if (ofp == NULL) p7_Fail("Failed to open -o output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Looks like the i/o is set up successfully...
* Initial output to the user
*/
output_header(go, ofp, alifile, postmsafile); /* cheery output header */
/* read the alignment */
if ((status = eslx_msafile_Read(afp, &msa)) != eslOK) eslx_msafile_ReadFailure(afp, status);
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
/* add/modify mmline for the mask */
if (msa->mm == NULL) {
ESL_ALLOC(msa->mm, msa->alen);
keep_mm = FALSE;
}
if (!keep_mm)
for (i=0; i<msa->alen; i++) msa->mm[i] = '.';
}
// convert model coordinates to alignment coordinates, if necessary
if (esl_opt_IsUsed(go, "--modelrange") || esl_opt_IsUsed(go, "--model2ali") || esl_opt_IsUsed(go, "--ali2model") ) {
float symfrac = esl_opt_GetReal(go, "--symfrac");
int do_hand = esl_opt_IsOn(go, "--hand");
int L;
//same as p7_builder relative_weights
if (esl_opt_IsOn(go, "--wnone") ) { esl_vec_DSet(msa->wgt, msa->nseq, 1.); }
else if (esl_opt_IsOn(go, "--wgiven") ) ;
else if (esl_opt_IsOn(go, "--wpb") ) status = esl_msaweight_PB(msa);
else if (esl_opt_IsOn(go, "--wgsc") ) status = esl_msaweight_GSC(msa);
else if (esl_opt_IsOn(go, "--wblosum")) status = esl_msaweight_BLOSUM(msa, esl_opt_GetReal(go, "--wid"));
if ((status = esl_msa_MarkFragments(msa, esl_opt_GetReal(go, "--fragthresh"))) != eslOK) goto ERROR;
//build a map of model mask coordinates to alignment coords
ESL_ALLOC(map, sizeof(int) * (msa->alen+1));
L = p7_Alimask_MakeModel2AliMap(msa, do_hand, symfrac, map );
if ( esl_opt_IsUsed(go, "--model2ali") ) {
//print mapping
printf ("model coordinates alignment coordinates\n");
for (i=0; i<mask_range_cnt; i++)
printf ("%8d..%-8d -> %8d..%-8d\n", mask_starts[i], mask_ends[i], map[mask_starts[i]-1], map[mask_ends[i]-1]);
/* If I wanted to, I could print all the map values independently:
printf("\n\n-----------\n");
printf("Map\n");
printf("---\n");
for (i=0; i<L; i++)
printf("%d -> %d\n", i+1, map[i]);
*/
} else if ( esl_opt_IsUsed(go, "--ali2model") ) {
//print mapping (requires scanning the inverted map
int alistart = 0;
int aliend = 0;
printf ("alignment coordinates model coordinates\n");
for (i=0; i<mask_range_cnt; i++) {
//find j for ali positions
while (map[alistart] < mask_starts[i] )
alistart++;
aliend = alistart;
while (map[aliend] < mask_ends[i] )
aliend++;
printf (" %8d..%-8d -> %8d..%-8d\n", map[alistart], map[aliend], alistart+1, aliend+1);
}
} else {
//convert the mask coords based on map
for (i=0; i<mask_range_cnt; i++) {
mask_starts[i] = map[mask_starts[i]-1]; //-1 because mmline is offset by one relative to the 1-base alignment
mask_ends[i] = map[mask_ends[i]-1];
}
}
}
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
//overwrite '.' with 'm' everywhere the range says to do it
for (i=0; i<mask_range_cnt; i++)
for (j=mask_starts[i]; j<=mask_ends[i]; j++)
msa->mm[j-1] = 'm';
if ((status = eslx_msafile_Write(postmsafp, msa, outfmt)) != eslOK) ESL_XEXCEPTION_SYS(eslEWRITE, "write failed");
}
esl_stopwatch_Stop(w);
if (esl_opt_IsOn(go, "-o")) fclose(ofp);
if (postmsafp) fclose(postmsafp);
if (afp) eslx_msafile_Close(afp);
if (abc) esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
esl_stopwatch_Destroy(w);
return 0;
ERROR:
return eslFAIL;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 2, argc, argv, banner, usage);
ESL_ALPHABET *abc = esl_alphabet_Create(eslAMINO);
char *hmmfile = esl_opt_GetArg(go, 1);
char *qfile = esl_opt_GetArg(go, 2);
ESL_SQ *qsq = esl_sq_CreateDigital(abc);
ESL_SQFILE *qfp = NULL;
FILE *hmmfp = NULL;
ESL_SCOREMATRIX *S = esl_scorematrix_Create(abc);
ESL_DMATRIX *Q = NULL;
P7_BG *bg = p7_bg_Create(abc);
P7_HMM *hmm = NULL;
double *fa = NULL;
double popen = esl_opt_GetReal (go, "-q");
double pextend = esl_opt_GetReal (go, "-r");
char *mxfile = esl_opt_GetString(go, "-m");
char errbuf[eslERRBUFSIZE];
double slambda;
int a,b;
int status;
/* Reverse engineer a scoring matrix to obtain conditional prob's
* that we'll use for the single-seq query HMM. Because score mx is
* symmetric, we can set up P[a][b] = P(b | a), so we can use the
* matrix rows as HMM match emission vectors. This means dividing
* the joint probs through by f_a.
*/
if (mxfile == NULL) {
if (esl_scorematrix_Set("BLOSUM62", S) != eslOK) esl_fatal("failed to set BLOSUM62 scores");
} else {
ESL_FILEPARSER *efp = NULL;
if ( esl_fileparser_Open(mxfile, NULL, &efp) != eslOK) esl_fatal("failed to open score file %s", mxfile);
if ( esl_scorematrix_Read(efp, abc, &S) != eslOK) esl_fatal("failed to read matrix from %s", mxfile);
esl_fileparser_Close(efp);
}
/* A wasteful conversion of the HMMER single-precision background probs to Easel double-prec */
ESL_ALLOC(fa, sizeof(double) * bg->abc->K);
esl_vec_F2D(bg->f, bg->abc->K, fa);
/* Backcalculate joint probabilities Q, given score matrix S and background frequencies fa */
status = esl_scorematrix_ProbifyGivenBG(S, fa, fa, &slambda, &Q);
if (status == eslEINVAL) esl_fatal("built-in score matrix %s has no valid solution for lambda", matrix);
else if (status == eslENOHALT) esl_fatal("failed to solve score matrix %s for lambda", matrix);
else if (status != eslOK) esl_fatal("unexpected error in solving score matrix %s for probability parameters", matrix);
esl_scorematrix_JointToConditionalOnQuery(abc, Q);
/* Open the query sequence file in FASTA format */
status = esl_sqfile_Open(qfile, eslSQFILE_FASTA, NULL, &qfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s.", qfile);
else if (status == eslEFORMAT) esl_fatal("Format of %s unrecognized.", qfile);
else if (status == eslEINVAL) esl_fatal("Can't autodetect stdin or .gz.");
else if (status != eslOK) esl_fatal("Open of %s failed, code %d.", qfile, status);
/* Open the output HMM file */
if ((hmmfp = fopen(hmmfile, "w")) == NULL) esl_fatal("Failed to open output HMM file %s", hmmfile);
/* For each sequence, build a model and save it.
*/
while ((status = esl_sqio_Read(qfp, qsq)) == eslOK)
{
p7_Seqmodel(abc, qsq->dsq, qsq->n, qsq->name, Q, bg->f, popen, pextend, &hmm);
if ( p7_hmm_Validate(hmm, errbuf, 1e-5) != eslOK) esl_fatal("HMM validation failed: %s\n", errbuf);
if ( p7_hmmfile_WriteASCII(hmmfp, -1, hmm) != eslOK) esl_fatal("HMM save failed");
p7_hmm_Destroy(hmm);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s line %" PRId64 "):\n%s\n",
qfp->filename, qfp->linenumber, qfp->errbuf);
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, qfp->filename);
esl_dmatrix_Destroy(Q);
esl_scorematrix_Destroy(S);
free(fa);
free(fb);
esl_sq_Destroy(qsq);
esl_sqfile_Close(qfp);
fclose(hmmfp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *hmmfile = NULL; /* HMM file name */
char *seqfile = NULL; /* sequence file name */
char *mapfile = NULL; /* optional mapped MSA file name */
int infmt = eslSQFILE_UNKNOWN;
int outfmt = eslMSAFILE_STOCKHOLM;
P7_HMMFILE *hfp = NULL; /* open HMM file */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
char *outfile = NULL; /* output filename */
FILE *ofp = stdout; /* output stream */
ESL_SQ **sq = NULL; /* array of sequences */
void *p = NULL; /* tmp ptr for reallocation */
int nseq = 0; /* # of sequences in <seqfile> */
int mapseq = 0; /* # of sequences in mapped MSA */
int totseq = 0; /* # of seqs in all sources */
ESL_ALPHABET *abc = NULL; /* alphabet (set from the HMM file)*/
P7_HMM *hmm = NULL;
P7_TRACE **tr = NULL; /* array of tracebacks */
ESL_MSA *msa = NULL; /* resulting multiple alignment */
int msaopts = 0; /* flags to p7_tracealign_Seqs() */
int idx; /* counter over seqs, traces */
int status; /* easel/hmmer return code */
char errbuf[eslERRBUFSIZE];
/* Parse the command line
*/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
hmmfile = esl_opt_GetArg(go, 1);
seqfile = esl_opt_GetArg(go, 2);
if (strcmp(hmmfile, "-") == 0 && strcmp(seqfile, "-") == 0)
cmdline_failure(argv[0], "Either <hmmfile> or <seqfile> may be '-' (to read from stdin), but not both.\n");
msaopts |= p7_ALL_CONSENSUS_COLS; /* default as of 3.1 */
if (esl_opt_GetBoolean(go, "--trim")) msaopts |= p7_TRIM;
/* If caller declared an input format, decode it
*/
if (esl_opt_IsOn(go, "--informat")) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--informat"));
}
/* Determine output alignment file format */
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a recognized output MSA file format\n", esl_opt_GetString(go, "--outformat"));
/* Open output stream */
if ( (outfile = esl_opt_GetString(go, "-o")) != NULL)
{
if ((ofp = fopen(outfile, "w")) == NULL)
cmdline_failure(argv[0], "failed to open -o output file %s for writing\n", outfile);
}
/* If caller forced an alphabet on us, create the one the caller wants
*/
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);
/* Read one HMM, and make sure there's only one.
*/
status = p7_hmmfile_OpenE(hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, hmmfile, errbuf);
status = p7_hmmfile_Read(hfp, &abc, &hmm);
if (status == eslEFORMAT) p7_Fail("Bad file format in HMM file %s:\n%s\n", hfp->fname, hfp->errbuf);
else if (status == eslEINCOMPAT) p7_Fail("HMM in %s is not in the expected %s alphabet\n", hfp->fname, esl_abc_DecodeType(abc->type));
else if (status == eslEOF) p7_Fail("Empty HMM file %s? No HMM data found.\n", hfp->fname);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s\n", hfp->fname);
status = p7_hmmfile_Read(hfp, &abc, NULL);
if (status != eslEOF) p7_Fail("HMM file %s does not contain just one HMM\n", hfp->fname);
p7_hmmfile_Close(hfp);
/* We're going to build up two arrays: sequences and traces.
* If --mapali option is chosen, the first set of sequences/traces is from the provided alignment
*/
if ( (mapfile = esl_opt_GetString(go, "--mapali")) != NULL)
{
map_alignment(mapfile, hmm, &sq, &tr, &mapseq);
}
totseq = mapseq;
/* Read digital sequences into an array (possibly concat'ed onto mapped seqs)
*/
status = esl_sqfile_OpenDigital(abc, seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", seqfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", seqfile);
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, seqfile);
ESL_RALLOC(sq, p, sizeof(ESL_SQ *) * (totseq + 1));
sq[totseq] = esl_sq_CreateDigital(abc);
nseq = 0;
while ((status = esl_sqio_Read(sqfp, sq[totseq+nseq])) == eslOK)
{
nseq++;
ESL_RALLOC(sq, p, sizeof(ESL_SQ *) * (totseq+nseq+1));
sq[totseq+nseq] = esl_sq_CreateDigital(abc);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s", status, sqfp->filename);
esl_sqfile_Close(sqfp);
totseq += nseq;
/* Remaining initializations, including trace array allocation
*/
ESL_RALLOC(tr, p, sizeof(P7_TRACE *) * totseq);
for (idx = mapseq; idx < totseq; idx++)
tr[idx] = p7_trace_CreateWithPP();
p7_tracealign_computeTraces(hmm, sq, mapseq, totseq - mapseq, tr);
p7_tracealign_Seqs(sq, tr, totseq, hmm->M, msaopts, hmm, &msa);
eslx_msafile_Write(ofp, msa, outfmt);
for (idx = 0; idx <= totseq; idx++) esl_sq_Destroy(sq[idx]); /* including sq[nseq] because we overallocated */
for (idx = 0; idx < totseq; idx++) p7_trace_Destroy(tr[idx]);
free(sq);
free(tr);
esl_msa_Destroy(msa);
p7_hmm_Destroy(hmm);
if (ofp != stdout) fclose(ofp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return eslOK;
ERROR:
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
ESL_ALPHABET *abc = NULL; /* biological alphabet */
char *alifile = NULL; /* alignment file name */
int fmt = eslMSAFILE_UNKNOWN; /* format code for alifile */
ESLX_MSAFILE *afp = NULL; /* open msa file */
ESL_MSAFILE2 *old_afp = NULL; /* open msa file, legacy (--small) */
ESL_MSA *msa = NULL; /* one multiple sequence alignment */
int nali; /* number of alignments read */
int i; /* counter over seqs */
int64_t alen; /* alignment length */
int nseq; /* number of sequences in the msa */
int64_t rlen; /* a raw (unaligned) seq length */
int64_t small, large; /* smallest, largest sequence */
int64_t nres; /* total # of residues in msa */
double avgid; /* average fractional pair id */
int max_comparisons; /* maximum # comparisons for avg id */
int do_stall; /* used to stall when debugging */
double **abc_ct = NULL; /* [0..msa->alen-1][0..abc->K] number of each residue at each position (abc->K is gap) */
double ***bp_ct = NULL; /* [0..msa->alen-1][0..abc->Kp-1][0..abc->Kp-1] per (non-pknotted) consensus basepair *
* count of each possible basepair over all seqs basepairs are indexed by 'i' the minimum *
* of 'i:j' for a pair between i and j, where i < j. */
double **pp_ct = NULL; /* [0..msa->alen-1][0..11], count of each posterior probability (PP) code, over all sequences, gap is 11 */
int *i_am_rf = NULL; /* [0..i..msa->alen-1]: TRUE if pos i is non-gap RF posn, if msa->rf == NULL remains NULL */
int *rf2a_map = NULL; /* [0..rfpos..rflen-1] = apos,
* apos is the alignment position (0..msa->alen-1) that
* is non-gap RF position rfpos+1 (for rfpos in 0..rflen-1) */
int rflen = -1; /* nongap RF length */
char errbuf[eslERRBUFSIZE];
int status; /* easel return code */
/* optional output files */
FILE *iinfofp = NULL; /* output file for --iinfo */
FILE *pcinfofp = NULL; /* output file for --pcinfo */
FILE *psinfofp = NULL; /* output file for --psinfo */
FILE *rinfofp = NULL; /* output file for --rinfo */
FILE *icinfofp = NULL; /* output file for --icinfo */
FILE *listfp = NULL; /* output file for --list */
FILE *cinfofp = NULL; /* output file for --cinfo */
FILE *bpinfofp = NULL; /* output file for --bpinfo */
int use_weights; /* TRUE if --weight, reported weighted counts (using msa->wgt) to all output files */
int weights_exist; /* TRUE if at least one msa->wgt value differs from 1.0, FALSE if not (or if msa->wgt==NULL) */
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK ||
esl_opt_VerifyConfig(go) != eslOK)
{
printf("Failed to parse command line: %s\n", go->errbuf);
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
if (esl_opt_GetBoolean(go, "-h") )
{
esl_banner(stdout, argv[0], banner);
esl_usage (stdout, argv[0], usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, 1, 2, 80);
puts("\n small memory mode, requires --amino,--dna, or --rna and --informat pfam:");
esl_opt_DisplayHelp(stdout, go, 2, 2, 80);
puts("\n optional output files:");
esl_opt_DisplayHelp(stdout, go, 3, 2, 80);
exit(0);
}
if (esl_opt_ArgNumber(go) != 1)
{
printf("Incorrect number of command line arguments.\n");
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
alifile = esl_opt_GetArg(go, 1);
if (esl_opt_IsOn(go, "--informat") &&
(fmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"))) == eslMSAFILE_UNKNOWN)
esl_fatal("%s is not a valid input sequence file format for --informat", esl_opt_GetString(go, "--informat"));
if (esl_opt_GetBoolean(go, "--small") && fmt != eslMSAFILE_PFAM) esl_fatal("--small requires --informat pfam\n");
max_comparisons = 1000;
do_stall = esl_opt_GetBoolean(go, "--stall"); /* a stall point for attaching gdb */
while (do_stall);
/***********************************************
* Open the MSA file; determine alphabet; set for digital input
***********************************************/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
/* We'd like to get rid of the legacy msafile interface, but it
* includes small memory functionality for Pfam format which we have
* to replace first. For now, use both interfaces, new and legacy
*/
if ( esl_opt_GetBoolean(go, "--small") )
{
if (! abc) esl_fatal("--small requires one of --amino, --dna, --rna be specified.");
status = esl_msafile2_OpenDigital(abc, alifile, NULL, &old_afp);
if (status == eslENOTFOUND) esl_fatal("Alignment file %s doesn't exist or is not readable\n", alifile);
else if (status == eslEFORMAT) esl_fatal("Couldn't determine format of alignment %s\n", alifile);
else if (status != eslOK) esl_fatal("Alignment file open failed with error %d\n", status);
}
else
{
if ( (status = eslx_msafile_Open(&abc, alifile, NULL, fmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
}
/**************************************
* Open optional output files, as nec *
**************************************/
/* determine name for first list file, if nec */
if( esl_opt_IsOn(go, "--list")) {
if ((listfp = fopen(esl_opt_GetString(go, "--list"), "w")) == NULL)
esl_fatal("Failed to open --list output file %s\n", esl_opt_GetString(go, "--list"));
}
if( esl_opt_IsOn(go, "--icinfo")) {
if ((icinfofp = fopen(esl_opt_GetString(go, "--icinfo"), "w")) == NULL)
esl_fatal("Failed to open --icinfo output file %s\n", esl_opt_GetString(go, "--icinfo"));
}
if( esl_opt_IsOn(go, "--rinfo")) {
if ((rinfofp = fopen(esl_opt_GetString(go, "--rinfo"), "w")) == NULL)
esl_fatal("Failed to open --rinfo output file %s\n", esl_opt_GetString(go, "--rinfo"));
}
if( esl_opt_IsOn(go, "--pcinfo")) {
if ((pcinfofp = fopen(esl_opt_GetString(go, "--pcinfo"), "w")) == NULL)
esl_fatal("Failed to open --pcinfo output file %s\n", esl_opt_GetString(go, "--pcinfo"));
}
if( esl_opt_IsOn(go, "--psinfo")) {
if ((psinfofp = fopen(esl_opt_GetString(go, "--psinfo"), "w")) == NULL)
esl_fatal("Failed to open --psinfo output file %s\n", esl_opt_GetString(go, "--psinfo"));
}
if( esl_opt_IsOn(go, "--iinfo")) {
if ((iinfofp = fopen(esl_opt_GetString(go, "--iinfo"), "w")) == NULL)
esl_fatal("Failed to open --iinfo output file %s\n", esl_opt_GetString(go, "--iinfo"));
}
if( esl_opt_IsOn(go, "--cinfo")) {
if ((cinfofp = fopen(esl_opt_GetString(go, "--cinfo"), "w")) == NULL)
esl_fatal("Failed to open --cinfo output file %s\n", esl_opt_GetString(go, "--cinfo"));
}
if( esl_opt_IsOn(go, "--bpinfo")) {
if ((bpinfofp = fopen(esl_opt_GetString(go, "--bpinfo"), "w")) == NULL)
esl_fatal("Failed to open --bpinfo output file %s\n", esl_opt_GetString(go, "--bpinfo"));
}
/***********************************************
* Read MSAs one at a time.
***********************************************/
if (esl_opt_GetBoolean(go, "-1")) {
puts("#");
if(! esl_opt_GetBoolean(go, "--small")) {
printf("# %-4s %-20s %10s %7s %7s %12s %6s %6s %10s %3s\n", "idx", "name", "format", "nseq", "alen", "nres", "small", "large", "avlen", "%id");
printf("# %-4s %-20s %10s %7s %7s %12s %6s %6s %10s %3s\n", "----", "--------------------", "----------", "-------", "-------", "------------", "------", "------", "----------", "---");
}
else {
printf("# %-4s %-20s %10s %7s %7s %12s %10s\n", "idx", "name", "format", "nseq", "alen", "nres", "avlen");
printf("# %-4s %-20s %10s %7s %7s %12s %10s\n", "----", "--------------------", "----------", "-------", "-------", "------------", "----------");
}
}
nali = 0;
fmt = (esl_opt_GetBoolean(go, "--small") ? old_afp->format : afp->format);
while ( (status = ( esl_opt_GetBoolean(go, "--small") ?
esl_msafile2_ReadInfoPfam(old_afp, listfp, abc, -1, NULL, NULL, &msa, &nseq, &alen, NULL, NULL, NULL, NULL, NULL, &abc_ct, &pp_ct, NULL, NULL, NULL) :
eslx_msafile_Read (afp, &msa))) == eslOK)
{
nali++;
nres = 0;
if (! esl_opt_GetBoolean(go, "--small")) {
nseq = msa->nseq;
alen = msa->alen;
small = large = -1;
for (i = 0; i < msa->nseq; i++)
{
rlen = esl_abc_dsqrlen(msa->abc, msa->ax[i]);
nres += rlen;
if (small == -1 || rlen < small) small = rlen;
if (large == -1 || rlen > large) large = rlen;
}
esl_dst_XAverageId(abc, msa->ax, msa->nseq, max_comparisons, &avgid);
}
else { /* --small invoked */
for(i = 0; i < alen; i++) nres += (int) esl_vec_DSum(abc_ct[i], abc->K);
}
if (esl_opt_GetBoolean(go, "-1"))
{
printf("%-6d %-20s %10s %7d %7" PRId64 " %12" PRId64,
nali,
msa->name,
eslx_msafile_DecodeFormat(fmt),
nseq,
alen,
nres);
if (! esl_opt_GetBoolean(go, "--small")) {
printf(" %6" PRId64 " %6" PRId64 " %10.1f %3.0f\n",
small,
large,
(double) nres / (double) msa->nseq,
100.*avgid);
}
else {
printf(" %10.1f\n", (double) nres / (double) nseq);
}
}
else
{
printf("Alignment number: %d\n", nali);
if (msa->name != NULL)
printf("Alignment name: %s\n", msa->name);
printf("Format: %s\n", eslx_msafile_DecodeFormat(fmt));
printf("Number of sequences: %d\n", nseq);
printf("Alignment length: %" PRId64 "\n", alen);
printf("Total # residues: %" PRId64 "\n", nres);
if(! esl_opt_GetBoolean(go, "--small")) {
printf("Smallest: %" PRId64 "\n", small);
printf("Largest: %" PRId64 "\n", large);
}
printf("Average length: %.1f\n", (double) nres / (double) nseq);
if(! esl_opt_GetBoolean(go, "--small")) {
printf("Average identity: %.0f%%\n", 100.*avgid);
}
printf("//\n");
}
/* Dump data to optional output files, if nec */
if(esl_opt_IsOn(go, "--list")) {
if(! esl_opt_GetBoolean(go, "--small")) {
/* only print sequence name to list file if ! --small, else we already have in esl_msafile2_ReadInfoPfam() */
for(i = 0; i < msa->nseq; i++) fprintf(listfp, "%s\n", msa->sqname[i]);
}
}
/* if RF exists, get i_am_rf array[0..alen] which tells us which positions are non-gap RF positions
* and rf2a_map, a map of non-gap RF positions to overall alignment positions */
if(msa->rf != NULL) {
if((status = map_rfpos_to_apos(msa, abc, errbuf, alen, &i_am_rf, &rf2a_map, &rflen)) != eslOK) esl_fatal(errbuf);
}
else i_am_rf = NULL;
weights_exist = check_msa_weights(msa);
use_weights = (weights_exist && esl_opt_GetBoolean(go, "--weight")) ? TRUE : FALSE;
if( (! esl_opt_GetBoolean(go, "--small")) &&
(esl_opt_IsOn(go, "--icinfo") || esl_opt_IsOn(go, "--rinfo") || esl_opt_IsOn(go, "--pcinfo") ||
esl_opt_IsOn(go, "--cinfo") || esl_opt_IsOn(go, "--bpinfo"))) {
/* collect counts of each residue and PPs (if they exist) from the msa */
if(esl_opt_GetBoolean(go, "--weight") && msa->wgt == NULL) esl_fatal("--weight requires all alignments have #=GS WT annotation, but aln %d does not", nali);
if((status = count_msa(msa, errbuf, nali,
esl_opt_GetBoolean(go, "--noambig"), /* ignore ambiguous residues? */
esl_opt_GetBoolean(go, "--weight"), /* use msa->wgt sequence weights? */
&abc_ct,
((bpinfofp != NULL && msa->ss_cons != NULL) ? &bp_ct : NULL), /* get basepair counts? */
(msa->pp != NULL ? &pp_ct : NULL))) /* get PP counts? */
!= eslOK) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--icinfo")) {
if((status = dump_infocontent_info(icinfofp, abc, abc_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--rinfo")) {
if((status = dump_residue_info(rinfofp, abc, abc_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if(esl_opt_IsOn(go, "--pcinfo")) {
if(pp_ct == NULL) esl_fatal("Error: --pcinfo requires all alignments have #=GR PP annotation, but alignment %d does not", nali);
if((status = dump_posterior_column_info(pcinfofp, pp_ct, use_weights, nali, alen, nseq, i_am_rf, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if(esl_opt_IsOn(go, "--psinfo")) {
if(msa->pp == NULL) esl_fatal("Error: --psinfo requires all alignments have #=GR PP annotation, but alignment %d does not", nali);
if((status = dump_posterior_sequence_info(psinfofp, msa, nali, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--iinfo")) {
if(msa->rf == NULL) esl_fatal("--iinfo requires all alignments have #=GC RF annotation, but alignment %d does not", nali);
if(esl_opt_GetBoolean(go, "--weight") && msa->wgt == NULL) esl_fatal("--weight requires all alignments have #=GS WT annotation, but aln %d does not", nali);
if((status = dump_insert_info(iinfofp, msa, use_weights, nali, i_am_rf, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--cinfo")) {
if((status = dump_column_residue_counts(cinfofp, abc, abc_ct, esl_opt_GetBoolean(go, "--noambig"), use_weights, nali, alen, nseq, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
if( esl_opt_IsOn(go, "--bpinfo")) {
if(msa->ss_cons == NULL) esl_fatal("--bpinfo requires all alignments have #=GC SS_cons annotation, but alignment %d does not", nali);
if((status = dump_basepair_counts(bpinfofp, msa, abc, bp_ct, use_weights, nali, nseq, msa->name, alifile, errbuf) != eslOK)) esl_fatal(errbuf);
}
esl_msa_Destroy(msa);
if(abc_ct != NULL) { esl_Free2D((void **) abc_ct, alen); abc_ct = NULL; }
if(bp_ct != NULL) { esl_Free3D((void ***) bp_ct, alen, abc->Kp); bp_ct = NULL; }
if(pp_ct != NULL) { esl_Free2D((void **) pp_ct, alen); pp_ct = NULL; }
if(i_am_rf != NULL) { free(i_am_rf); i_am_rf = NULL; }
if(rf2a_map != NULL) { free(rf2a_map); rf2a_map = NULL; }
}
/* If an msa read failed, we've dropped out to here with an informative status code.
* we have to handle failures from new vs. legacy msa parsing differently
*/
if (esl_opt_GetBoolean(go, "--small"))
{
if (status == eslEFORMAT) esl_fatal("Alignment file parse error, line %d of file %s:\n%s\nOffending line is:\n%s\n", old_afp->linenumber, old_afp->fname, old_afp->errbuf, old_afp->buf);
else if (status != eslEOF) esl_fatal("Alignment file read failed with error code %d\n", status);
else if (nali == 0) esl_fatal("No alignments found in file %s\n", alifile);
}
else
{
if (nali == 0 || status != eslEOF) eslx_msafile_ReadFailure(afp, status);
}
/* Cleanup, normal return
*/
if(listfp != NULL) {
fclose(listfp);
printf("# List of sequences in %d alignment(s) saved to file %s\n", nali, esl_opt_GetString(go, "--list"));
}
if(icinfofp != NULL) {
fclose(icinfofp);
printf("# Information content data saved to file %s.\n", esl_opt_GetString(go, "--icinfo"));
}
if(rinfofp != NULL) {
fclose(rinfofp);
printf("# Residue data saved to file %s.\n", esl_opt_GetString(go, "--rinfo"));
}
if(pcinfofp != NULL) {
fclose(pcinfofp);
printf("# Per-column posterior probability data saved to file %s.\n", esl_opt_GetString(go, "--pcinfo"));
}
if(psinfofp != NULL) {
fclose(psinfofp);
printf("# Per-sequence posterior probability data saved to file %s.\n", esl_opt_GetString(go, "--psinfo"));
}
if(iinfofp != NULL) {
printf("# Insert data saved to file %s.\n", esl_opt_GetString(go, "--iinfo"));
fclose(iinfofp);
}
if(cinfofp != NULL) {
printf("# Per-column counts data saved to file %s.\n", esl_opt_GetString(go, "--cinfo"));
fclose(cinfofp);
}
if(bpinfofp != NULL) {
printf("# Per-column basepair counts data saved to file %s.\n", esl_opt_GetString(go, "--bpinfo"));
fclose(bpinfofp);
}
if (afp) eslx_msafile_Close(afp);
if (old_afp) esl_msafile2_Close(old_afp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *alifile = NULL; /* alignment file name */
int infmt = eslMSAFILE_UNKNOWN; /* format code for alifile */
int outfmt = eslMSAFILE_UNKNOWN; /* output format for fetched msa's */
ESLX_MSAFILE *afp = NULL; /* open alignment file */
FILE *ofp = NULL; /* output stream for alignments */
int status; /* easel return code */
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) < 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
if (esl_opt_IsOn(go, "--informat")) {
infmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslMSAFILE_UNKNOWN) esl_fatal("%s is not a valid input alignment file format for --informat", esl_opt_GetString(go, "--informat"));
}
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) esl_fatal("%s is not a valid output alignment file format for --outformat", esl_opt_GetString(go, "--outformat"));
alifile = esl_opt_GetArg(go, 1);
/* Open the alignment file. */
if ( (status = eslx_msafile_Open(NULL, alifile, NULL, infmt, NULL, &afp)) != eslOK)
eslx_msafile_OpenFailure(afp, status);
/* Open the SSI index, if any */
if (! esl_opt_GetBoolean(go, "--index"))
{
if (afp->bf->mode_is == eslBUFFER_FILE ||
afp->bf->mode_is == eslBUFFER_ALLFILE ||
afp->bf->mode_is == eslBUFFER_MMAP)
{
char *ssifile = NULL;
esl_sprintf(&ssifile, "%s.ssi", afp->bf->filename);
status = esl_ssi_Open(ssifile, &(afp->ssi));
if (status == eslERANGE ) esl_fatal("SSI index %s has 64-bit offsets; this system doesn't support them", ssifile);
else if (status == eslEFORMAT) esl_fatal("SSI index %s has an unrecognized format. Try recreating, w/ esl-afetch --index", ssifile);
else if (status == eslENOTFOUND) afp->ssi = NULL;
else if (status != eslOK) esl_fatal("SSI index %s: open failed, error code %d\n", ssifile, status);
free(ssifile);
}
}
/* Open the output file, if any
*/
if (esl_opt_GetBoolean(go, "-O"))
{
if ((ofp = fopen(esl_opt_GetArg(go, 2), "w")) == NULL)
esl_fatal("Failed to open output file %s\n", esl_opt_GetArg(go, 2));
}
else if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
esl_fatal("Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Hand off control flow as appropriate */
if (esl_opt_GetBoolean(go, "--index"))
{
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
create_ssi_index(go, afp);
}
else if (esl_opt_GetBoolean(go, "-f"))
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
multifetch(go, ofp, outfmt, esl_opt_GetArg(go, 2), afp);
}
else
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
onefetch(go, ofp, outfmt, esl_opt_GetArg(go, 2), afp);
if (ofp != stdout) printf("\n\nRetrieved alignment %s.\n", esl_opt_GetArg(go, 2));
}
eslx_msafile_Close(afp);
esl_getopts_Destroy(go);
exit(0);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
FILE *ofp = NULL; /* output stream for sequences */
int status; /* easel return code */
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) < 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
/* Open the sequence file */
seqfile = esl_opt_GetArg(go, 1);
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat", esl_opt_GetString(go, "--informat"));
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
/* Open the output file, if any */
if (esl_opt_GetBoolean(go, "-O"))
{
if ((ofp = fopen(esl_opt_GetArg(go, 2), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetArg(go, 2));
}
else if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Indexing mode */
if (esl_opt_GetBoolean(go, "--index"))
{
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
if (sqfp->data.ascii.do_gzip) cmdline_failure(argv[0], "Can't index a .gz compressed file");
if (sqfp->data.ascii.do_stdin) cmdline_failure(argv[0], "Can't index a standard input pipe");
create_ssi_index(go, sqfp);
}
/* List retrieval mode */
else if (esl_opt_GetBoolean(go, "-f"))
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
/* Open the SSI index for retrieval */
if (! sqfp->data.ascii.do_gzip && ! sqfp->data.ascii.do_stdin && ! esl_sqio_IsAlignment(sqfp->format))
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
if (esl_opt_GetBoolean(go, "-C")) multifetch_subseq(go, ofp, esl_opt_GetArg(go, 2), sqfp);
else multifetch (go, ofp, esl_opt_GetArg(go, 2), sqfp);
}
/* Single sequence retrieval mode */
else
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
char *key = esl_opt_GetArg(go, 2);
char *cstring = esl_opt_GetString(go, "-c");
char *newname = esl_opt_GetString(go, "-n");
/* Open the SSI index for retrieval */
if (! sqfp->data.ascii.do_gzip && ! sqfp->data.ascii.do_stdin && ! esl_sqio_IsAlignment(sqfp->format))
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
/* -c: subsequence retrieval; else full sequence retrieval */
if (cstring != NULL)
{
uint32_t start, end;
status = esl_regexp_ParseCoordString(cstring, &start, &end);
if (status == eslESYNTAX) esl_fatal("-c takes arg of subseq coords <from>..<to>; %s not recognized", cstring);
if (status == eslFAIL) esl_fatal("Failed to find <from> or <to> coord in %s", cstring);
onefetch_subseq(go, ofp, sqfp, newname, key, start, end);
if (ofp != stdout) printf("\n\nRetrieved subsequence %s/%d-%d.\n", key, start, end);
}
else
{
onefetch(go, ofp, esl_opt_GetArg(go, 2), sqfp);
if (ofp != stdout) printf("\n\nRetrieved sequence %s.\n", esl_opt_GetArg(go, 2));
}
}
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_Create(esl_opt_GetInteger(go, "-s"));
double mu = esl_opt_GetReal (go, "-m");
double lambda = esl_opt_GetReal (go, "-l");
double tau = esl_opt_GetReal (go, "-t");
int n = esl_opt_GetInteger(go, "-n");
double binwidth = esl_opt_GetReal (go, "-w");
int plot_cdf = esl_opt_GetBoolean(go, "--cdf");
int plot_logcdf = esl_opt_GetBoolean(go, "--logcdf");
int plot_pdf = esl_opt_GetBoolean(go, "--pdf");
int plot_logpdf = esl_opt_GetBoolean(go, "--logpdf");
int plot_surv = esl_opt_GetBoolean(go, "--surv");
int plot_logsurv = esl_opt_GetBoolean(go, "--logsurv");
int be_verbose = esl_opt_GetBoolean(go, "-v");
char *plotfile = esl_opt_GetString (go, "-o");
ESL_HISTOGRAM *h = NULL;
int xmin_set = esl_opt_IsOn(go, "--xL");
double xmin = xmin_set ? esl_opt_GetReal(go, "--xL") : mu;
int xmax_set = esl_opt_IsOn(go, "--xH");
double xmax = xmax_set ? esl_opt_GetReal(go, "--xH") : mu+40*(1./lambda);
int xstep_set = esl_opt_IsOn(go, "--xH");
double xstep = xstep_set ? esl_opt_GetReal(go, "--xS") : 0.1;
FILE *pfp = stdout;
double emu, elambda, etau;
int i;
double x;
double *data;
int ndata;
fprintf(stderr, "## %s\n", argv[0]);
fprintf(stderr, "# rng seed = %" PRIu32 "\n", esl_randomness_GetSeed(rng));
if (be_verbose) printf("Parametric: mu = %f lambda = %f tau = %f\n", mu, lambda, tau);
h = esl_histogram_CreateFull(mu, 100., binwidth);
if (plotfile && (pfp = fopen(plotfile, "w")) == NULL) ESL_EXCEPTION(eslFAIL, "Failed to open plotfile");
for (i = 0; i < n; i++)
{
x = esl_wei_Sample(rng, mu, lambda, tau);
esl_histogram_Add(h, x);
}
esl_histogram_GetData(h, &data, &ndata);
esl_wei_FitComplete(data, ndata, &emu, &elambda, &etau);
if (be_verbose) printf("Complete data fit: mu = %f lambda = %f tau = %f\n", emu, elambda, etau);
if (fabs( (emu-mu)/mu ) > 0.01) ESL_EXCEPTION(eslFAIL, "Error in (complete) fitted mu > 1%\n");
if (fabs( (elambda-lambda)/lambda ) > 0.10) ESL_EXCEPTION(eslFAIL, "Error in (complete) fitted lambda > 10%\n");
if (fabs( (etau-tau)/tau ) > 0.10) ESL_EXCEPTION(eslFAIL, "Error in (complete) fitted tau > 10%\n");
esl_wei_FitCompleteBinned(h, &emu, &elambda, &etau);
if (be_verbose) printf("Binned data fit: mu = %f lambda = %f tau = %f\n", emu, elambda, etau);
if (fabs( (emu-mu)/mu ) > 0.01) ESL_EXCEPTION(eslFAIL, "Error in (binned) fitted mu > 1%\n");
if (fabs( (elambda-lambda)/lambda ) > 0.10) ESL_EXCEPTION(eslFAIL, "Error in (binned) fitted lambda > 10%\n");
if (fabs( (etau-tau)/tau ) > 0.10) ESL_EXCEPTION(eslFAIL, "Error in (binned) fitted lambda > 10%\n");
if (plot_pdf) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_pdf, xmin, xmax, xstep);
if (plot_logpdf) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_logpdf, xmin, xmax, xstep);
if (plot_cdf) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_cdf, xmin, xmax, xstep);
if (plot_logcdf) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_logcdf, xmin, xmax, xstep);
if (plot_surv) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_surv, xmin, xmax, xstep);
if (plot_logsurv) esl_wei_Plot(pfp, mu, lambda, tau, &esl_wei_logsurv, xmin, xmax, xstep);
if (plotfile) fclose(pfp);
esl_histogram_Destroy(h);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
fprintf(stderr, "# status = ok\n");
return 0;
}
/* serial_master()
* The serial version of hmmsearch.
* For each query HMM in <hmmfile> search the database for hits.
*
* A master can only return if it's successful. All errors are handled
* immediately and fatally with p7_Fail(). We also use the
* ESL_EXCEPTION and ERROR: mechanisms, but only because we know we're
* using a fatal exception handler.
*/
static int
serial_master(ESL_GETOPTS *go, struct cfg_s *cfg)
{
FILE *ofp = stdout; /* results output file (-o) */
P7_HMMFILE *hfp = NULL; /* open input HMM file */
ESL_SQFILE *dbfp = NULL; /* open input sequence file */
P7_HMM *hmm = NULL; /* one HMM query */
ESL_ALPHABET *abc = NULL; /* digital alphabet */
int dbfmt = eslSQFILE_UNKNOWN; /* format code for sequence database file */
ESL_STOPWATCH *w;
int textw = 0;
int nquery = 0;
int status = eslOK;
int hstatus = eslOK;
int sstatus = eslOK;
int i;
int ncpus = 0;
int infocnt = 0;
WORKER_INFO *info = NULL;
char errbuf[eslERRBUFSIZE];
w = esl_stopwatch_Create();
if (esl_opt_GetBoolean(go, "--notextw")) textw = 0;
else textw = esl_opt_GetInteger(go, "--textw");
if (esl_opt_IsOn(go, "--tformat")) {
dbfmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--tformat"));
if (dbfmt == eslSQFILE_UNKNOWN) p7_Fail("%s is not a recognized sequence database file format\n", esl_opt_GetString(go, "--tformat"));
}
/* Open the target sequence database */
status = esl_sqfile_Open(cfg->dbfile, dbfmt, p7_SEQDBENV, &dbfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", cfg->dbfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", cfg->dbfile);
else if (status == eslEINVAL) p7_Fail("Can't autodetect format of a stdin or .gz seqfile");
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, cfg->dbfile);
if (esl_opt_IsUsed(go, "--restrictdb_stkey") || esl_opt_IsUsed(go, "--restrictdb_n")) {
if (esl_opt_IsUsed(go, "--ssifile"))
esl_sqfile_OpenSSI(dbfp, esl_opt_GetString(go, "--ssifile"));
else
esl_sqfile_OpenSSI(dbfp, NULL);
}
/* Open the query profile HMM file */
status = p7_hmmfile_OpenE(cfg->hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, cfg->hmmfile, errbuf);
/* Open the results output files */
if (esl_opt_IsOn(go, "-o")) { if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w+")) == NULL) p7_Fail("Failed to open output file %s for writing\n", esl_opt_GetString(go, "-o")); }
infocnt = 1;
ESL_ALLOC(info, sizeof(*info) * infocnt);
/* <abc> is not known 'til first HMM is read. */
hstatus = p7_hmmfile_Read(hfp, &abc, &hmm);
if (hstatus == eslOK)
{
/* One-time initializations after alphabet <abc> becomes known */
// output_header(ofp, go, cfg->hmmfile, cfg->dbfile);
// dbfp->abc = abc; //ReadBlock requires knowledge of the alphabet to decide how best to read blocks
// for (i = 0; i < infocnt; ++i)
// {
// info[i].bg = p7_bg_Create(abc);
// }
}
/* Outer loop: over each query HMM in <hmmfile>. */
// while (hstatus == eslOK)
// {
P7_PROFILE *gm = NULL;
P7_OPROFILE *om = NULL; /* optimized query profile */
nquery++;
esl_stopwatch_Start(w);
/* seqfile may need to be rewound (multiquery mode) */
if (nquery > 1)
{
if (! esl_sqfile_IsRewindable(dbfp))
esl_fatal("Target sequence file %s isn't rewindable; can't search it with multiple queries", cfg->dbfile);
if (! esl_opt_IsUsed(go, "--restrictdb_stkey") )
esl_sqfile_Position(dbfp, 0); //only re-set current position to 0 if we're not planning to set it in a moment
}
if ( cfg->firstseq_key != NULL ) { //it's tempting to want to do this once and capture the offset position for future passes, but ncbi files make this non-trivial, so this keeps it general
sstatus = esl_sqfile_PositionByKey(dbfp, cfg->firstseq_key);
if (sstatus != eslOK)
p7_Fail("Failure setting restrictdb_stkey to %d\n", cfg->firstseq_key);
}
// if (fprintf(ofp, "Query: %s [M=%d]\n", hmm->name, hmm->M) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// if (hmm->acc) { if (fprintf(ofp, "Accession: %s\n", hmm->acc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
// if (hmm->desc) { if (fprintf(ofp, "Description: %s\n", hmm->desc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
/* Convert to an optimized model */
gm = p7_profile_Create (hmm->M, abc);
om = p7_oprofile_Create(hmm->M, abc);
// p7_ProfileConfig(hmm, info->bg, gm, 100, p7_LOCAL); /* 100 is a dummy length for now; and MSVFilter requires local mode */
p7_oprofile_Convert(gm, om); /* <om> is now p7_LOCAL, multihit */
for (i = 0; i < infocnt; ++i)
{
/* Create processing pipeline and hit list */
info[i].th = p7_tophits_Create();
info[i].om = p7_oprofile_Clone(om);
info[i].pli = p7_pipeline_Create(go, om->M, 100, FALSE, p7_SEARCH_SEQS); /* L_hint = 100 is just a dummy for now */
P7_PIPELINE *pli = info[i].pli;
pli->nmodels++;
pli->nnodes += info[i].om->M;
// if (pli->Z_setby == p7_ZSETBY_NTARGETS && pli->mode == p7_SCAN_MODELS) pli->Z = pli->nmodels;
// if (pli->do_biasfilter) p7_bg_SetFilter(info[i].bg, info[i].om->M, info[i].om->compo);
// if (pli->mode == p7_SEARCH_SEQS)
// status = p7_pli_NewModelThresholds(pli, info[i].om);
pli->W = info[i].om->max_length;
}
sstatus = serial_loop(info, dbfp, cfg->n_targetseq, ofp);
switch(sstatus)
{
case eslEFORMAT:
esl_fatal("Parse failed (sequence file %s):\n%s\n",
dbfp->filename, esl_sqfile_GetErrorBuf(dbfp));
break;
case eslEOF:
/* do nothing */
break;
default:
esl_fatal("Unexpected error %d reading sequence file %s", sstatus, dbfp->filename);
}
/* merge the results of the search results */
for (i = 1; i < infocnt; ++i)
{
p7_tophits_Merge(info[0].th, info[i].th);
p7_pipeline_Merge(info[0].pli, info[i].pli);
p7_pipeline_Destroy(info[i].pli);
p7_tophits_Destroy(info[i].th);
p7_oprofile_Destroy(info[i].om);
}
/* Print the results. */
p7_tophits_SortBySortkey(info->th);
p7_tophits_Threshold(info->th, info->pli);
// p7_tophits_Targets(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// p7_tophits_Domains(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
esl_stopwatch_Stop(w);
// p7_pli_Statistics(ofp, info->pli, w);
// if (fprintf(ofp, "//\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
p7_pipeline_Destroy(info->pli);
p7_tophits_Destroy(info->th);
p7_oprofile_Destroy(info->om);
p7_oprofile_Destroy(om);
p7_profile_Destroy(gm);
p7_hmm_Destroy(hmm);
// hstatus = p7_hmmfile_Read(hfp, &abc, &hmm);
// } /* end outer loop over query HMMs */
switch(hstatus) {
case eslEOD: p7_Fail("read failed, HMM file %s may be truncated?", cfg->hmmfile); break;
case eslEFORMAT: p7_Fail("bad file format in HMM file %s", cfg->hmmfile); break;
case eslEINCOMPAT: p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile); break;
case eslEOF:
case eslOK: /* do nothing. EOF is what we want. */ break;
default: p7_Fail("Unexpected error (%d) in reading HMMs from %s", hstatus, cfg->hmmfile);
}
/* Terminate outputs... any last words?
*/
/* Cleanup - prepare for exit
*/
// for (i = 0; i < infocnt; ++i)
// p7_bg_Destroy(info[i].bg);
free(info);
p7_hmmfile_Close(hfp);
esl_sqfile_Close(dbfp);
esl_alphabet_Destroy(abc);
esl_stopwatch_Destroy(w);
if (ofp != stdout) fclose(ofp);
printf("44HHHH \n");
return eslOK;
ERROR:
return eslFAIL;
}
static int
output_header(const ESL_GETOPTS *go, FILE *ofp, char *alifile, char *postmsafile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# input alignment file: %s\n", alifile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
if (fprintf(ofp, "# output alignment file: %s\n", postmsafile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--alirange") && fprintf(ofp, "# alignment range: %s\n", esl_opt_GetString(go, "--alirange")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--modelrange") && fprintf(ofp, "# model range: %s\n", esl_opt_GetString(go, "--modelrange"))< 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--apendmask") && fprintf(ofp, "# add to existing mask: [on]\n" )< 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--model2ali") && fprintf(ofp, "# ali ranges for model range: %s\n", esl_opt_GetString(go, "--model2ali")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--ali2model") && fprintf(ofp, "# model ranges for ali range: %s\n", esl_opt_GetString(go, "--ali2model")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--amino") && fprintf(ofp, "# input alignment is asserted as: protein\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--dna") && fprintf(ofp, "# input alignment is asserted as: DNA\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--rna") && fprintf(ofp, "# input alignment is asserted as: RNA\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--fast") && fprintf(ofp, "# model architecture construction: fast/heuristic\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--hand") && fprintf(ofp, "# model architecture construction: hand-specified by RF annotation\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--symfrac") && fprintf(ofp, "# sym fraction for model structure: %.3f\n", esl_opt_GetReal(go, "--symfrac")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--fragthresh") && fprintf(ofp, "# seq called frag if L <= x*alen: %.3f\n", esl_opt_GetReal(go, "--fragthresh")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wpb") && fprintf(ofp, "# relative weighting scheme: Henikoff PB\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wgsc") && fprintf(ofp, "# relative weighting scheme: G/S/C\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wblosum") && fprintf(ofp, "# relative weighting scheme: BLOSUM filter\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wnone") && fprintf(ofp, "# relative weighting scheme: none\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wid") && fprintf(ofp, "# frac id cutoff for BLOSUM wgts: %f\n", esl_opt_GetReal(go, "--wid")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed") == 0 && fprintf(ofp,"# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp,"# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
int
main(int argc, char **argv)
{
int status;
ESL_ALPHABET *abc = NULL; /* sequence alphabet */
ESL_GETOPTS *go = NULL; /* command line processing */
ESL_RANDOMNESS *r = NULL; /* source of randomness */
P7_HMM *hmm = NULL; /* sampled HMM to emit from */
P7_HMM *core = NULL; /* safe copy of the HMM, before config */
P7_BG *bg = NULL; /* null model */
ESL_SQ *sq = NULL; /* sampled sequence */
P7_TRACE *tr = NULL; /* sampled trace */
P7_PROFILE *gm = NULL; /* profile */
int i,j;
int i1,i2;
int k1,k2;
int iseq;
FILE *fp = NULL;
double expected;
int do_ilocal;
char *hmmfile = NULL;
int nseq;
int do_swlike;
int do_ungapped;
int L;
int M;
int do_h2;
char *ipsfile = NULL;
char *kpsfile = NULL;
ESL_DMATRIX *imx = NULL;
ESL_DMATRIX *kmx = NULL;
ESL_DMATRIX *iref = NULL; /* reference matrix: expected i distribution under ideality */
int Lbins;
/*****************************************************************
* Parse the command line
*****************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) esl_fatal("Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) esl_fatal("Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") == TRUE) {
puts(usage);
puts("\n where options are:\n");
esl_opt_DisplayHelp(stdout, go, 0, 2, 80); /* 0=all docgroups; 2 = indentation; 80=textwidth*/
return eslOK;
}
do_ilocal = esl_opt_GetBoolean(go, "-i");
hmmfile = esl_opt_GetString (go, "-m");
nseq = esl_opt_GetInteger(go, "-n");
do_swlike = esl_opt_GetBoolean(go, "-s");
do_ungapped = esl_opt_GetBoolean(go, "-u");
L = esl_opt_GetInteger(go, "-L");
M = esl_opt_GetInteger(go, "-M");
do_h2 = esl_opt_GetBoolean(go, "-2");
ipsfile = esl_opt_GetString (go, "--ips");
kpsfile = esl_opt_GetString (go, "--kps");
if (esl_opt_ArgNumber(go) != 0) {
puts("Incorrect number of command line arguments.");
printf("Usage: %s [options]\n", argv[0]);
return eslFAIL;
}
r = esl_randomness_CreateFast(0);
if (hmmfile != NULL)
{ /* Read the HMM (and get alphabet from it) */
P7_HMMFILE *hfp = NULL;
status = p7_hmmfile_Open(hmmfile, NULL, &hfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open HMM file %s for reading.\n", hmmfile);
else if (status == eslEFORMAT) p7_Fail("File %s does not appear to be in a recognized HMM format.\n", hmmfile);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n", status, hmmfile);
if ((status = p7_hmmfile_Read(hfp, &abc, &hmm)) != eslOK) {
if (status == eslEOD) esl_fatal("read failed, HMM file %s may be truncated?", hmmfile);
else if (status == eslEFORMAT) esl_fatal("bad file format in HMM file %s", hmmfile);
else if (status == eslEINCOMPAT) esl_fatal("HMM file %s contains different alphabets", hmmfile);
else esl_fatal("Unexpected error in reading HMMs");
}
M = hmm->M;
p7_hmmfile_Close(hfp);
}
else
{ /* Or sample the HMM (create alphabet first) */
abc = esl_alphabet_Create(eslAMINO);
if (do_ungapped) p7_hmm_SampleUngapped(r, M, abc, &hmm);
else if (do_swlike) p7_hmm_SampleUniform (r, M, abc, 0.05, 0.5, 0.05, 0.2, &hmm); /* tmi, tii, tmd, tdd */
else p7_hmm_Sample (r, M, abc, &hmm);
}
Lbins = M;
imx = esl_dmatrix_Create(Lbins, Lbins);
iref = esl_dmatrix_Create(Lbins, Lbins);
kmx = esl_dmatrix_Create(M, M);
esl_dmatrix_SetZero(imx);
esl_dmatrix_SetZero(iref);
esl_dmatrix_SetZero(kmx);
tr = p7_trace_Create();
sq = esl_sq_CreateDigital(abc);
bg = p7_bg_Create(abc);
core = p7_hmm_Clone(hmm);
if (do_h2) {
gm = p7_profile_Create(hmm->M, abc);
p7_H2_ProfileConfig(hmm, bg, gm, p7_UNILOCAL);
} else {
gm = p7_profile_Create(hmm->M, abc);
p7_ProfileConfig(hmm, bg, gm, L, p7_UNILOCAL);
if (p7_hmm_Validate (hmm, NULL, 0.0001) != eslOK) esl_fatal("whoops, HMM is bad!");
if (p7_profile_Validate(gm, NULL, 0.0001) != eslOK) esl_fatal("whoops, profile is bad!");
}
/* Sample endpoints.
* Also sample an ideal reference distribution for i endpoints. i
* endpoints are prone to discretization artifacts, when emitted
* sequences have varying lengths. Taking log odds w.r.t. an ideal
* reference that is subject to the same discretization artifacts
* cancels out the effect.
*/
for (iseq = 0; iseq < nseq; iseq++)
{
if (do_ilocal) ideal_local_endpoints (r, core, sq, tr, Lbins, &i1, &i2, &k1, &k2);
else profile_local_endpoints(r, core, gm, sq, tr, Lbins, &i1, &i2, &k1, &k2);
imx->mx[i1-1][i2-1] += 1.;
kmx->mx[k1-1][k2-1] += 1.;
/* reference distribution for i */
ideal_local_endpoints (r, core, sq, tr, Lbins, &i1, &i2, &k1, &k2);
iref->mx[i1-1][i2-1] += 1.;
}
/* Adjust both mx's to log_2(obs/exp) ratio */
printf("Before normalization/log-odds:\n");
printf(" i matrix values range from %f to %f\n", dmx_upper_min(imx), dmx_upper_max(imx));
printf(" k matrix values range from %f to %f\n", dmx_upper_min(kmx), dmx_upper_max(kmx));
printf("iref matrix values range from %f to %f\n", dmx_upper_min(iref), dmx_upper_max(iref));
expected = (double) nseq * 2. / (double) (M*(M+1));
for (i = 0; i < kmx->m; i++)
for (j = i; j < kmx->n; j++)
kmx->mx[i][j] = log(kmx->mx[i][j] / expected) / log(2.0);
for (i = 0; i < imx->m; i++)
for (j = i; j < imx->m; j++)
if (iref->mx[i][j] == 0. && imx->mx[i][j] == 0.)
imx->mx[i][j] = 0.;
else if (iref->mx[i][j] == 0.)
imx->mx[i][j] = eslINFINITY;
else if (imx->mx[i][j] == 0.)
imx->mx[i][j] = -eslINFINITY;
else
imx->mx[i][j] = log(imx->mx[i][j] / iref->mx[i][j]) / log(2.0);
/* Print ps files */
if (kpsfile != NULL) {
if ((fp = fopen(kpsfile, "w")) == NULL) esl_fatal("Failed to open output postscript file %s", kpsfile);
dmx_Visualize(fp, kmx, -4., 5.);
fclose(fp);
}
if (ipsfile != NULL) {
if ((fp = fopen(ipsfile, "w")) == NULL) esl_fatal("Failed to open output postscript file %s", ipsfile);
dmx_Visualize(fp, imx, -4., 5.);
/* dmx_Visualize(fp, imx, dmx_upper_min(imx), dmx_upper_max(imx)); */
fclose(fp);
}
printf("After normalization/log-odds:\n");
printf("i matrix values range from %f to %f\n", dmx_upper_min(imx), dmx_upper_max(imx));
printf("k matrix values range from %f to %f\n", dmx_upper_min(kmx), dmx_upper_max(kmx));
p7_profile_Destroy(gm);
p7_bg_Destroy(bg);
p7_hmm_Destroy(core);
p7_hmm_Destroy(hmm);
p7_trace_Destroy(tr);
esl_sq_Destroy(sq);
esl_dmatrix_Destroy(imx);
esl_dmatrix_Destroy(kmx);
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return eslOK;
}
/* Function: main()
* Synopsis: Run set of queries against an FM
* Purpose: Read in a FM and a file of query sequences.
* For each query, find matching FM interval, then collect positions in
* the original text T for the corresponding occurrences. These positions
* are 0-based (so first character is position 0).
*/
int
main(int argc, char *argv[])
{
void* tmp; // used for RALLOC calls
clock_t t1, t2;
struct tms ts1, ts2;
char *fname_fm = NULL;
char *fname_queries = NULL;
FM_HIT *hits = NULL;
char *line = NULL;
int status = eslOK;
int hit_cnt = 0;
int hit_indiv_cnt = 0;
int miss_cnt = 0;
int hit_num = 0;
int hit_num2 = 0;
int hits_size = 0;
int i,j;
int count_only = 0;
FM_INTERVAL interval;
FM_DATA *fmsf = NULL;
FM_DATA *fmsb = NULL;
FILE* fp_fm = NULL;
FILE* fp = NULL;
FILE* out = NULL;
char *outname = NULL;
ESL_GETOPTS *go = NULL; /* command line processing */
FM_CFG *cfg;
FM_METADATA *meta;
ESL_SQ *tmpseq; // used for sequence validation
ESL_ALPHABET *abc = NULL;
//start timer
t1 = times(&ts1);
process_commandline(argc, argv, &go, &fname_fm, &fname_queries);
if (esl_opt_IsOn(go, "--out")) {
outname = esl_opt_GetString(go, "--out");
if ( esl_strcmp ("-", outname) == 0 ) {
out = stdout;
outname = "stdout";
} else {
out = fopen(outname,"w");
}
}
if (esl_opt_IsOn(go, "--count_only"))
count_only = 1;
if((fp_fm = fopen(fname_fm, "rb")) == NULL)
esl_fatal("Cannot open file `%s': ", fname_fm);
fm_configAlloc(&cfg);
cfg->occCallCnt = 0;
meta = cfg->meta;
meta->fp = fp_fm;
fm_readFMmeta( meta);
if (meta->alph_type == fm_DNA) abc = esl_alphabet_Create(eslDNA);
else if (meta->alph_type == fm_AMINO) abc = esl_alphabet_Create(eslAMINO);
tmpseq = esl_sq_CreateDigital(abc);
//read in FM-index blocks
ESL_ALLOC(fmsf, meta->block_count * sizeof(FM_DATA) );
if (!meta->fwd_only)
ESL_ALLOC(fmsb, meta->block_count * sizeof(FM_DATA) );
for (i=0; i<meta->block_count; i++) {
fm_FM_read( fmsf+i,meta, TRUE );
if (!meta->fwd_only) {
fm_FM_read(fmsb+i, meta, FALSE );
fmsb[i].SA = fmsf[i].SA;
fmsb[i].T = fmsf[i].T;
}
}
fclose(fp_fm);
output_header(meta, stdout, go, fname_fm, fname_queries);
/* initialize a few global variables, then call initGlobals
* to do architecture-specific initialization
*/
fm_configInit(cfg, NULL);
fm_alphabetCreate(meta, NULL); // don't override charBits
fp = fopen(fname_queries,"r");
if (fp == NULL)
esl_fatal("Unable to open file %s\n", fname_queries);
ESL_ALLOC(line, FM_MAX_LINE * sizeof(char));
hits_size = 200;
ESL_ALLOC(hits, hits_size * sizeof(FM_HIT));
while(fgets(line, FM_MAX_LINE, fp) ) {
int qlen=0;
while (line[qlen] != '\0' && line[qlen] != '\n') qlen++;
if (line[qlen] == '\n') line[qlen] = '\0';
hit_num = 0;
for (i=0; i<meta->block_count; i++) {
fm_getSARangeReverse(fmsf+i, cfg, line, meta->inv_alph, &interval);
if (interval.lower>=0 && interval.lower <= interval.upper) {
int new_hit_num = interval.upper - interval.lower + 1;
hit_num += new_hit_num;
if (!count_only) {
if (hit_num > hits_size) {
hits_size = 2*hit_num;
ESL_RALLOC(hits, tmp, hits_size * sizeof(FM_HIT));
}
getFMHits(fmsf+i, cfg, &interval, i, hit_num-new_hit_num, qlen, hits, fm_forward);
}
}
/* find reverse hits, using backward search on the forward FM*/
if (!meta->fwd_only) {
fm_getSARangeForward(fmsb+i, cfg, line, meta->inv_alph, &interval);// yes, use the backward fm to produce the equivalent of a forward search on the forward fm
if (interval.lower>=0 && interval.lower <= interval.upper) {
int new_hit_num = interval.upper - interval.lower + 1;
hit_num += new_hit_num;
if (!count_only) {
if (hit_num > hits_size) {
hits_size = 2*hit_num;
ESL_RALLOC(hits, tmp, hits_size * sizeof(FM_HIT));
}
//even though I used fmsb above, use fmsf here, since we'll now do a backward trace
//in the FM-index to find the next sampled SA position
getFMHits(fmsf+i, cfg, &interval, i, hit_num-new_hit_num, qlen, hits, fm_backward);
}
}
}
}
if (hit_num > 0) {
if (count_only) {
hit_cnt++;
hit_indiv_cnt += hit_num;
} else {
hit_num2 = 0;
//for each hit, identify the sequence id and position within that sequence
for (i = 0; i< hit_num; i++) {
status = fm_getOriginalPosition (fmsf, meta, hits[i].block, hits[i].length, fm_forward, hits[i].start, &(hits[i].block), &(hits[i].start) );
hits[i].sortkey = (status==eslERANGE ? -1 : meta->seq_data[ hits[i].block ].target_id);
//validate match - if any characters in orig sequence were ambiguities, reject
fm_convertRange2DSQ( fmsf, meta, hits[i].start, hits[i].length, p7_NOCOMPLEMENT, tmpseq, TRUE );
for (j=1; j<=hits[i].length; j++) {
if (tmpseq->dsq[j] >= abc->K) {
hits[i].sortkey = -1; //reject
j = hits[i].length+1; //quit looking
}
}
if (hits[i].sortkey != -1)
hit_num2++; // legitimate hit
}
if (hit_num2 > 0)
hit_cnt++;
//now sort according the the sequence_id corresponding to that seq_offset
qsort(hits, hit_num, sizeof(FM_HIT), hit_sorter);
//skim past the skipped entries
i = 0;
while ( i < hit_num ) {
if (hits[i].sortkey != -1 )
break; //
i++;
}
if (i < hit_num) {
if (out != NULL) {
fprintf (out, "%s\n",line);
//fprintf (out, "\t%10s (%8d %s)\n",meta->seq_data[ hits[i].block ].name, hits[i].start, (hits[i].direction==fm_forward?"+":"-"));
fprintf (out, " %8ld %s %10s\n", (long)(hits[i].start), (hits[i].direction==fm_forward?"f":"r"), meta->seq_data[ hits[i].block ].name);
}
hit_indiv_cnt++;
i++; // skip the first one, since I'll be comparing each to the previous
for ( ; i< hit_num; i++) {
if ( //meta->seq_data[ hits[i].block ].id != meta->seq_data[ hits[i-1].block ].id ||
hits[i].sortkey != hits[i-1].sortkey || //sortkey is seq_data[].id
hits[i].direction != hits[i-1].direction ||
hits[i].start != hits[i-1].start )
{
if (out != NULL)
//fprintf (out, "\t%10s (%8d %s)\n",meta->seq_data[ hits[i].block ].name, hits[i].start, (hits[i].direction==fm_forward?"+":"-"));
fprintf (out, " %8ld %s %10s\n", (long)(hits[i].start), (hits[i].direction==fm_forward?"f":"r"), meta->seq_data[ hits[i].block ].name);
hit_indiv_cnt++;
}
}
if (out != NULL)
fprintf (out, "\n");
}
}
} else {
miss_cnt++;
}
}
for (i=0; i<meta->block_count; i++) {
fm_FM_destroy( fmsf+i, 1 );
if (!meta->fwd_only)
fm_FM_destroy( fmsb+i, 0 );
}
free (hits);
free (line);
fclose(fp);
fm_configDestroy(cfg);
// compute and print the elapsed time in millisec
t2 = times(&ts2);
{
double clk_ticks = sysconf(_SC_CLK_TCK);
double elapsedTime = (t2-t1)/clk_ticks;
double throughput = cfg->occCallCnt/elapsedTime;
fprintf (stderr, "hit: %-10d (%d)\n", hit_cnt, hit_indiv_cnt);
fprintf (stderr, "miss:%-10d\n", miss_cnt);
fprintf (stderr, "run time: %.2f seconds\n", elapsedTime);
fprintf (stderr, "occ calls: %12s\n", commaprint(cfg->occCallCnt));
fprintf (stderr, "occ/sec: %12s\n", commaprint(throughput));
}
exit(eslOK);
ERROR:
printf ("failure allocating memory for hits\n");
exit(status);
}
/* map_sub_msas
*
* msa1 and msa2 contain the same named sequences, msa1 contains a superset
* of the columns in msa2. Determine which of the msa1 columns the msa2
* columns correspond to.
*/
static int
map_sub_msas(const ESL_GETOPTS *go, char *errbuf, ESL_MSA *msa1, ESL_MSA *msa2, char **ret_msa1_to_msa2_mask)
{
int status;
int apos1, apos2; /* counters over alignment position in msa1, msa2 respectively */
int i;
int *msa1_to_msa2_map; /* [0..apos1..msa1->alen] msa2 alignment position that apos1 corresponds to */
char *mask;
/* contract check */
if(! (msa1->flags & eslMSA_DIGITAL)) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas() msa1 (%s) not digitized.\n", esl_opt_GetArg(go, 1));
if(! (msa2->flags & eslMSA_DIGITAL)) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas() msa2 (%s) not digitized.\n", esl_opt_GetString(go, "--submap"));
if(msa1->alen <= msa2->alen) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas() alignment length for msa1 (%" PRId64 "d) <= length for msa2 (%" PRId64 ")\n", msa1->alen, msa2->alen);
ESL_ALLOC(mask, sizeof(char) * (msa1->alen+1));
for(apos1 = 0; apos1 < msa1->alen; apos1++) mask[apos1] = '0';
mask[msa1->alen] = '\0';
ESL_ALLOC(msa1_to_msa2_map, sizeof(int) * (msa1->alen+1));
esl_vec_ISet(msa1_to_msa2_map, (msa1->alen+1), -1);
/* both alignments must have same 'named' sequences in same order */
if(msa1->nseq != msa2->nseq) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas() msa1 has %d sequences, msa2 has %d sequences\n", msa1->nseq, msa2->nseq);
for(i = 0; i < msa1->nseq; i++) {
if(strcmp(msa1->sqname[i], msa2->sqname[i]) != 0) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas() msa1 seq %d is named %s, msa2 seq %d is named %s\n", i, msa1->sqname[i], i, msa2->sqname[i]);
}
apos1 = 1;
apos2 = 1;
while((apos2 <= msa2->alen) || (apos1 <= msa1->alen)) { /* determine which apos1 (alignment column in msa1), apos2 (alignment column in msa2) corresponds to */
for(i = 0; i < msa1->nseq; i++) {
if(msa1->ax[i][apos1] != msa2->ax[i][apos2]) {
apos1++;
break; /* try next apos1 */
}
}
if(i == msa1->nseq) { /* found a match */
msa1_to_msa2_map[apos1] = apos2;
mask[(apos1-1)] = '1';
apos1++;
apos2++;
}
}
if((apos1 != (msa1->alen+1)) || (apos2 != (msa2->alen+1))) ESL_FAIL(eslEINVAL, errbuf, "in map_sub_msas(), failure mapping alignments, end of loop apos1-1 = %d (msa1->alen: %" PRId64 ") and apos2-1 = %d (msa2->alen: %" PRId64 ")\n", apos1-1, msa1->alen, apos2-1, msa2->alen);
free(msa1_to_msa2_map);
*ret_msa1_to_msa2_mask = mask;
return eslOK;
ERROR:
return status;
}
/* seq_shuffling()
* SRE, Tue Jan 22 08:35:51 2008 [Market Street Cafe, Leesburg]
*
* Shuffling of input sequences.
*
* Fixed-length (L>0) vs. full-length (L=0) modes handled differently.
* In fixed-length mode:
* <shuff->seq> only needs to be allocated once, for L
* <targ> is an allocated copy of a random subseq of length L
* sequences < L residues long can't be shuffled
* In full-length mode:
* <shuff->seq> is grown to length <sq->n> for each input seq
* <targ> just points to <sq->seq>
*/
static int
seq_shuffling(ESL_GETOPTS *go, ESL_RANDOMNESS *r, FILE *ofp, int outfmt)
{
char *seqfile = esl_opt_GetArg(go, 1);
int infmt = eslSQFILE_UNKNOWN;
ESL_SQFILE *sqfp = NULL;
ESL_SQ *sq = esl_sq_Create();
ESL_SQ *shuff = esl_sq_Create();
char *targ = NULL;
int N = esl_opt_GetInteger(go, "-N");
int L = esl_opt_GetInteger(go, "-L"); /* L>0 means select random fixed-len subseqs */
int kmers = 0;
int i;
int status;
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
if (esl_opt_IsOn(go, "-k")) kmers = esl_opt_GetInteger(go, "-k");
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", seqfile);
else if (status == eslEINVAL) esl_fatal("Can't autodetect stdin or .gz.");
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
if (L>0) {
esl_sq_GrowTo(shuff, L);
shuff->n = L;
ESL_ALLOC(targ, sizeof(char) * (L+1));
}
while ((status = esl_sqio_Read(sqfp, sq)) == eslOK)
{
if (L == 0) { /* shuffling entire sequence */
esl_sq_GrowTo(shuff, sq->n); /* make sure shuff can hold sq */
shuff->n = sq->n;
targ = sq->seq;
} else {
if (sq->n < L) continue; /* reject seqs < L long */
}
for (i = 0; i < N; i++)
{
if (L > 0) { /* fixed-len mode: copy a random subseq */
int pos = esl_rnd_Roll(r, sq->n - L + 1);
strncpy(targ, sq->seq + pos, L);
targ[L] = '\0';
}
/* Do the requested kind of shuffling */
if (esl_opt_GetBoolean(go, "-m")) esl_rsq_CShuffle (r, targ, shuff->seq); /* monoresidue shuffling */
else if (esl_opt_GetBoolean(go, "-d")) esl_rsq_CShuffleDP (r, targ, shuff->seq); /* diresidue shuffling */
else if (esl_opt_IsOn (go, "-k")) esl_rsq_CShuffleKmers(r, targ, kmers, shuff->seq); /* diresidue shuffling */
else if (esl_opt_GetBoolean(go, "-0")) esl_rsq_CMarkov0 (r, targ, shuff->seq); /* 0th order Markov */
else if (esl_opt_GetBoolean(go, "-1")) esl_rsq_CMarkov1 (r, targ, shuff->seq); /* 1st order Markov */
else if (esl_opt_GetBoolean(go, "-r")) esl_rsq_CReverse ( targ, shuff->seq); /* reverse */
else if (esl_opt_IsOn (go, "-w")) { /* regionally shuffle */
int W= esl_opt_GetInteger(go, "-w"); esl_rsq_CShuffleWindows(r, targ, W, shuff->seq);
}
/* Set the name of the shuffled sequence */
if (N > 1) esl_sq_FormatName(shuff, "%s-shuffled-%d", sq->name, i);
else esl_sq_FormatName(shuff, "%s-shuffled", sq->name);
/* Output the resulting sequence */
esl_sqio_Write(ofp, shuff, outfmt, FALSE);
/* don't need to reuse the shuffled sequence: we will use exactly the same memory */
}
esl_sq_Reuse(sq);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
if (L>0) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return eslOK;
ERROR:
if (targ != NULL) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return status;
}
/* map2masks
*
* Given a map of alignment columns in msa1 to alignment columns
* to msa2, construct and output masks as per command-line options.
*
* Args: msa1_to_msa2_map: [1..apos..msa1->alen]: '0': msa1 apos maps to a gap in msa2 (doesn't map to any column in msa2)
* 'x': msa1 apos maps to posn x in msa2 (x>0)
*/
static int
map2masks(const ESL_GETOPTS *go, char *errbuf, int alen1, int alen2, int *a2rf_map1, int *a2rf_map2, int *rf2a_map1, int *rf2a_map2, int rflen1, int rflen2, int *msa1_to_msa2_map)
{
int status;
int apos1, apos2; /* counters over alignment position in msa1, msa2 respectively */
int rfpos1, rfpos2; /* counters over reference positions */
int num_ones; /* number of 1s in current mask */
int num_zeroes; /* number of 0s in current mask */
FILE *fp;
char *mask = NULL;
if(esl_opt_GetString(go, "--mask-a2a")) {
if ((fp = fopen(esl_opt_GetString(go, "--mask-a2a"), "w")) == NULL)
ESL_FAIL(eslFAIL, errbuf, "Failed to open --mask-a2a mask output file %s", esl_opt_GetString(go, "--mask-a2a"));
/* construct mask as follows:
* mask[0..apos1..alen1-1] = '1' if column apos1+1 maps to an alignment column of msa2
* = '0' if column apos1+1 maps to a gap in msa2 (doesn't map to any column in msa2)
*/
ESL_ALLOC(mask, sizeof(char) * (alen1+1));
num_ones = num_zeroes = 0;
for(apos1 = 1; apos1 <= alen1; apos1++) {
if(msa1_to_msa2_map[apos1] == 0) { mask[(apos1-1)] = '0'; num_zeroes++; }
else { mask[(apos1-1)] = '1'; num_ones++; }
}
mask[alen1] = '\0';
fprintf(fp, "%s\n", mask);
free(mask);
fclose(fp);
printf("# Mask of 1/0s with 1 indicating aln column in %s maps to aln column in %s saved to file %s.\n# (Length: %d; '1's: %d; '0's: %d)\n", esl_opt_GetArg(go, 1), esl_opt_GetArg(go, 2), esl_opt_GetString(go, "--mask-a2a"), (num_ones+num_zeroes), num_ones, num_zeroes);
}
if(esl_opt_GetString(go, "--mask-a2rf")) {
if (a2rf_map2 == NULL) ESL_FAIL(eslFAIL, errbuf, "with --mask-a2rf, <msafile2> %s must have #=GC RF annotation, but it doesn't.", esl_opt_GetArg(go, 2));
if ((fp = fopen(esl_opt_GetString(go, "--mask-a2rf"), "w")) == NULL)
ESL_FAIL(eslFAIL, errbuf, "Failed to open --mask-a2rf mask output file %s\n", esl_opt_GetString(go, "--mask-a2rf"));
/* construct mask as follows:
* mask[0..apos1..alen1-1] = '1' if column apos1+1 maps to a reference column (non-gap in RF) of msa2
* = '0' if column apos1+1 maps to a gap (doesn't map to any column in msa2) or an insert (gap in RF) in msa2
*/
ESL_ALLOC(mask, sizeof(char) * (alen1+1));
num_ones = num_zeroes = 0;
for(apos1 = 1; apos1 <= alen1; apos1++) {
apos2 = msa1_to_msa2_map[apos1];
if(apos2 == 0) { mask[(apos1-1)] = '0'; num_zeroes++; } /* apos1 doesn't map to any column in msa2 */
else {
rfpos2 = a2rf_map2[apos2];
if(rfpos2 <= 0) { mask[(apos1-1)] = '0'; num_zeroes++; } /* apos1 maps to a gap RF (insert) in msa2 */
else { mask[(apos1-1)] = '1'; num_ones++; } /* apos1 maps to a non-gap RF (reference) column in msa2 */
}
}
mask[alen1] = '\0';
fprintf(fp, "%s\n", mask);
free(mask);
fclose(fp);
printf("# Mask of 1/0s with 1 indicating aln column in %s maps to reference (non-gap RF) column in %s saved to file %s.\n# (Length: %d; '1's: %d; '0's: %d)\n", esl_opt_GetArg(go, 1), esl_opt_GetArg(go, 2), esl_opt_GetString(go, "--mask-a2rf"), (num_ones+num_zeroes), num_ones, num_zeroes);
}
if(esl_opt_GetString(go, "--mask-rf2a")) {
if (a2rf_map1 == NULL) ESL_FAIL(eslFAIL, errbuf, "with --mask-rf2a, <msafile1> %s must have #=GC RF annotation, but it doesn't.", esl_opt_GetArg(go, 1));
if ((fp = fopen(esl_opt_GetString(go, "--mask-rf2a"), "w")) == NULL)
ESL_FAIL(eslFAIL, errbuf, "Failed to open --mask-rf2a mask output file %s\n", esl_opt_GetString(go, "--mask-rf2a"));
/* construct mask as follows:
* mask[0..rfpos1..rflen-1] = '1' if non-gap RF msa1 column rfpos1+1 maps to an alignment column of msa2
* = '0' if non-gap RF msa1 column rfpos1+1 maps to a gap in msa2 (doesn't map to any column in msa2)
*/
ESL_ALLOC(mask, sizeof(char) * (rflen1+1));
num_ones = num_zeroes = 0;
for(rfpos1 = 1; rfpos1 <= rflen1; rfpos1++) {
apos1 = rf2a_map1[rfpos1];
apos2 = msa1_to_msa2_map[apos1];
if(apos2 == 0) { mask[(rfpos1-1)] = '0'; num_zeroes++; }
else { mask[(rfpos1-1)] = '1'; num_ones++; }
}
mask[rflen1] = '\0';
fprintf(fp, "%s\n", mask);
free(mask);
fclose(fp);
printf("# Mask of 1/0s with 1 indicating reference (non-gap RF) column in %s maps to aln column in %s saved to file %s.\n# (Length: %d; '1's: %d; '0's: %d)\n", esl_opt_GetArg(go, 1), esl_opt_GetArg(go, 2), esl_opt_GetString(go, "--mask-rf2a"), (num_ones+num_zeroes), num_ones, num_zeroes);
}
if(esl_opt_GetString(go, "--mask-rf2rf")) {
if (a2rf_map1 == NULL) ESL_FAIL(eslFAIL, errbuf, "with --mask-rf2rf, <msafile1> %s must have #=GC RF annotation, but it doesn't.", esl_opt_GetArg(go, 1));
if (a2rf_map2 == NULL) ESL_FAIL(eslFAIL, errbuf, "with --mask-rf2rf, <msafile2> %s must have #=GC RF annotation, but it doesn't.", esl_opt_GetArg(go, 2));
if ((fp = fopen(esl_opt_GetString(go, "--mask-rf2rf"), "w")) == NULL)
ESL_FAIL(eslFAIL, errbuf, "Failed to open --mask-rf2rf mask output file %s\n", esl_opt_GetString(go, "--mask-rf2rf"));
/* construct mask as follows:
* mask[0..apos1..alen-1] = '1' if column apos1+1 maps to a reference column (non-gap in RF) of msa2
* = '0' if column apos1+1 maps to a gap (doesn't map to any column in msa2) or an insert (gap in RF) in msa2
*/
ESL_ALLOC(mask, sizeof(char) * (alen1+1));
num_ones = num_zeroes = 0;
for(rfpos1 = 1; rfpos1 <= rflen1; rfpos1++) {
apos1 = rf2a_map1[rfpos1];
apos2 = msa1_to_msa2_map[apos1];
if(apos2 == 0) { mask[(rfpos1-1)] = '0'; num_zeroes++; } /* rfpos1 doesn't map to any column in msa2 */
else {
rfpos2 = a2rf_map2[apos2];
if(rfpos2 <= 0) { mask[(rfpos1-1)] = '0'; num_zeroes++; } /* rfpos1 maps to a gap RF (insert) in msa2 */
else { mask[(rfpos1-1)] = '1'; num_ones++; } /* rfpos1 maps to a non-gap RF (reference) column in msa2 */
}
}
mask[rflen1] = '\0';
fprintf(fp, "%s\n", mask);
free(mask);
fclose(fp);
printf("# Mask of 1/0s with 1 indicating reference (non-gap RF) column in %s maps to reference (non-gap RF) column in %s saved to file %s.\n# (Length: %d; '1's: %d; '0's: %d)\n", esl_opt_GetArg(go, 1), esl_opt_GetArg(go, 2), esl_opt_GetString(go, "--mask-rf2rf"), (num_ones+num_zeroes), num_ones, num_zeroes);
}
return eslOK;
ERROR:
ESL_FAIL(eslEMEM, errbuf, "map2masks(): memory allocation error.");
return status; /* NEVERREACHED */
}
/* serial_master()
* The serial version of hmmsearch.
* For each query HMM in <hmmdb> search the database for hits.
*
* A master can only return if it's successful. All errors are handled
* immediately and fatally with p7_Fail().
*/
static int
serial_master(ESL_GETOPTS *go, struct cfg_s *cfg)
{
FILE *ofp = stdout; /* output file for results (default stdout) */
FILE *tblfp = NULL; /* output stream for tabular per-seq (--tblout) */
FILE *dfamtblfp = NULL; /* output stream for tabular Dfam format (--dfamtblout) */
FILE *aliscoresfp = NULL; /* output stream for alignment scores (--aliscoresout) */
// P7_HMM *hmm = NULL; /* one HMM query */
// P7_SCOREDATA *scoredata = NULL;
int seqfmt = eslSQFILE_UNKNOWN; /* format of seqfile */
ESL_SQFILE *sqfp = NULL; /* open seqfile */
P7_HMMFILE *hfp = NULL; /* open HMM database file */
ESL_ALPHABET *abc = NULL; /* sequence alphabet */
P7_OPROFILE *om = NULL; /* target profile */
ESL_STOPWATCH *w = NULL; /* timing */
ESL_SQ *qsq = NULL; /* query sequence */
int nquery = 0;
int textw;
int status = eslOK;
int hstatus = eslOK;
int sstatus = eslOK;
int i;
int ncpus = 0;
int infocnt = 0;
WORKER_INFO *info = NULL;
#ifdef HMMER_THREADS
P7_OM_BLOCK *block = NULL;
ESL_THREADS *threadObj= NULL;
ESL_WORK_QUEUE *queue = NULL;
#endif
char errbuf[eslERRBUFSIZE];
double window_beta = -1.0 ;
int window_length = -1;
if (esl_opt_IsUsed(go, "--w_beta")) { if ( ( window_beta = esl_opt_GetReal(go, "--w_beta") ) < 0 || window_beta > 1 ) esl_fatal("Invalid window-length beta value\n"); }
if (esl_opt_IsUsed(go, "--w_length")) { if (( window_length = esl_opt_GetInteger(go, "--w_length")) < 4 ) esl_fatal("Invalid window length value\n"); }
w = esl_stopwatch_Create();
if (esl_opt_GetBoolean(go, "--notextw")) textw = 0;
else textw = esl_opt_GetInteger(go, "--textw");
/* If caller declared an input format, decode it */
if (esl_opt_IsOn(go, "--qformat")) {
seqfmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--qformat"));
if (seqfmt == eslSQFILE_UNKNOWN) p7_Fail("%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--qformat"));
}
/* validate options if running as a daemon */
// if (esl_opt_IsOn(go, "--daemon")) {
/* running as a daemon, the input format must be type daemon */
// if (seqfmt != eslSQFILE_UNKNOWN && seqfmt != eslSQFILE_DAEMON)
// esl_fatal("Input format %s not supported. Must be daemon\n", esl_opt_GetString(go, "--qformat"));
// seqfmt = eslSQFILE_DAEMON;
// if (strcmp(cfg->seqfile, "-") != 0) esl_fatal("Query sequence file must be '-'\n");
// }
/* Open the target profile database to get the sequence alphabet */
status = p7_hmmfile_OpenE(cfg->hmmfile, p7_HMMDBENV, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem, trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, cfg->hmmfile, errbuf);
if (! hfp->is_pressed) p7_Fail("Failed to open binary auxfiles for %s: use hmmpress first\n", hfp->fname);
hstatus = p7_oprofile_ReadMSV(hfp, &abc, &om);
if (hstatus == eslEFORMAT) p7_Fail("bad format, binary auxfiles, %s:\n%s", cfg->hmmfile, hfp->errbuf);
else if (hstatus == eslEINCOMPAT) p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile);
else if (hstatus != eslOK) p7_Fail("Unexpected error in reading HMMs from %s", cfg->hmmfile);
p7_oprofile_Destroy(om);
p7_hmmfile_Close(hfp);
/* Open the query sequence database */
status = esl_sqfile_OpenDigital(abc, cfg->seqfile, seqfmt, NULL, &sqfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", cfg->seqfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", cfg->seqfile);
else if (status == eslEINVAL) p7_Fail("Can't autodetect format of a stdin or .gz seqfile");
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, cfg->seqfile);
if (sqfp->format > 100) // breaking the law! That range is reserved for msa, for aligned formats
p7_Fail("%s contains a multiple sequence alignment; expect unaligned sequences, like FASTA\n", cfg->seqfile);
qsq = esl_sq_CreateDigital(abc);
/* Open the results output files */
if (esl_opt_IsOn(go, "-o")) { if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL) esl_fatal("Failed to open output file %s for writing\n", esl_opt_GetString(go, "-o")); }
if (esl_opt_IsOn(go, "--tblout")) { if ((tblfp = fopen(esl_opt_GetString(go, "--tblout"), "w")) == NULL) esl_fatal("Failed to open tabular per-seq output file %s for writing\n", esl_opt_GetString(go, "--tblfp")); }
if (esl_opt_IsOn(go, "--dfamtblout")) { if ((dfamtblfp = fopen(esl_opt_GetString(go, "--dfamtblout"),"w")) == NULL) esl_fatal("Failed to open tabular dfam output file %s for writing\n", esl_opt_GetString(go, "--dfamtblout")); }
if (esl_opt_IsOn(go, "--aliscoresout")) { if ((aliscoresfp = fopen(esl_opt_GetString(go, "--aliscoresout"),"w")) == NULL) esl_fatal("Failed to open alignment scores output file %s for writing\n", esl_opt_GetString(go, "--aliscoresout")); }
output_header(ofp, go, cfg->hmmfile, cfg->seqfile);
#ifdef HMMER_THREADS
/* initialize thread data */
if (esl_opt_IsOn(go, "--cpu")) ncpus = esl_opt_GetInteger(go, "--cpu");
else esl_threads_CPUCount(&ncpus);
if (ncpus > 0)
{
threadObj = esl_threads_Create(&pipeline_thread);
queue = esl_workqueue_Create(ncpus * 2);
}
#endif
infocnt = (ncpus == 0) ? 1 : ncpus;
ESL_ALLOC(info, sizeof(*info) * infocnt);
for (i = 0; i < infocnt; ++i)
{
info[i].bg = p7_bg_Create(abc);
#ifdef HMMER_THREADS
info[i].queue = queue;
#endif
}
#ifdef HMMER_THREADS
for (i = 0; i < ncpus * 2; ++i)
{
block = p7_oprofile_CreateBlock(BLOCK_SIZE);
if (block == NULL) esl_fatal("Failed to allocate sequence block");
status = esl_workqueue_Init(queue, block);
if (status != eslOK) esl_fatal("Failed to add block to work queue");
}
#endif
/* Outside loop: over each query sequence in <seqfile>. */
while ((sstatus = esl_sqio_Read(sqfp, qsq)) == eslOK)
{
if (sstatus == eslEMEM) p7_Fail("Memory allocation error reading sequence file\n", status);
if (sstatus == eslEINCONCEIVABLE) p7_Fail("Unexpected error %d reading sequence file\n", status);
// if (qsq->L > NHMMER_MAX_RESIDUE_COUNT) p7_Fail("Input sequence %s in file %s exceeds maximum length of %d bases.\n", qsq->name, cfg->seqfile, NHMMER_MAX_RESIDUE_COUNT);
nquery++;
esl_stopwatch_Start(w);
/* Open the target profile database */
status = p7_hmmfile_OpenE(cfg->hmmfile, p7_HMMDBENV, &hfp, NULL);
if (status != eslOK) p7_Fail("Unexpected error %d in opening hmm file %s.\n", status, cfg->hmmfile);
#ifdef HMMER_THREADS
/* if we are threaded, create a lock to prevent multiple readers */
if (ncpus > 0)
{
status = p7_hmmfile_CreateLock(hfp);
if (status != eslOK) p7_Fail("Unexpected error %d creating lock\n", status);
}
#endif
if (fprintf(ofp, "Query: %s [L=%ld]\n", qsq->name, (long) qsq->n) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (qsq->acc[0] != 0 && fprintf(ofp, "Accession: %s\n", qsq->acc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (qsq->desc[0] != 0 && fprintf(ofp, "Description: %s\n", qsq->desc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
for (i = 0; i < infocnt; ++i)
{
/* Create processing pipeline and hit list */
info[i].th = p7_tophits_Create();
info[i].pli = p7_pipeline_Create(go, 100, 100, TRUE, p7_SCAN_MODELS); /* M_hint = 100, L_hint = 100 are just dummies for now */
info[i].pli->hfp = hfp; /* for two-stage input, pipeline needs <hfp> */
p7_pli_NewSeq(info[i].pli, qsq);
info[i].qsq = qsq;
if ( esl_opt_IsUsed(go, "--toponly") )
info[i].pli->strand = p7_STRAND_TOPONLY;
else if ( esl_opt_IsUsed(go, "--bottomonly") )
info[i].pli->strand = p7_STRAND_BOTTOMONLY;
else
info[i].pli->strand = p7_STRAND_BOTH;
#ifdef HMMER_THREADS
if (ncpus > 0) esl_threads_AddThread(threadObj, &info[i]);
#endif
}
#ifdef HMMER_THREADS
if (ncpus > 0) hstatus = thread_loop(threadObj, queue, hfp);
else hstatus = serial_loop(info, hfp);
#else
hstatus = serial_loop(info, hfp);
#endif
switch(hstatus)
{
case eslEFORMAT: p7_Fail("bad file format in HMM file %s", cfg->hmmfile); break;
case eslEINCOMPAT: p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile); break;
case eslEOF:
case eslOK: /* do nothing */
break;
default: p7_Fail("Unexpected error in reading HMMs from %s", cfg->hmmfile);
}
/* merge the results of the search results */
for (i = 1; i < infocnt; ++i)
{
p7_tophits_Merge(info[0].th, info[i].th);
p7_pipeline_Merge(info[0].pli, info[i].pli);
p7_pipeline_Destroy(info[i].pli);
p7_tophits_Destroy(info[i].th);
}
/* modify e-value to account for number of models */
for (i = 0; i < info->th->N ; i++)
{
info->th->unsrt[i].lnP += log((float)info->pli->nmodels);
info->th->unsrt[i].dcl[0].lnP = info->th->unsrt[i].lnP;
info->th->unsrt[i].sortkey = -1.0 * info->th->unsrt[i].lnP;
}
/* it's possible to have duplicates based on how viterbi ranges can overlap */
p7_tophits_SortByModelnameAndAlipos(info->th);
p7_tophits_RemoveDuplicates(info->th, info->pli->use_bit_cutoffs);
/* Print results */
p7_tophits_SortBySortkey(info->th);
p7_tophits_Threshold(info->th, info->pli);
//tally up total number of hits and target coverage
info->pli->n_output = info->pli->pos_output = 0;
for (i = 0; i < info->th->N; i++) {
if ( (info->th->hit[i]->flags & p7_IS_REPORTED) || info->th->hit[i]->flags & p7_IS_INCLUDED) {
info->pli->n_output++;
info->pli->pos_output += abs(info->th->hit[i]->dcl[0].jali - info->th->hit[i]->dcl[0].iali) + 1;
}
}
p7_tophits_Targets(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
p7_tophits_Domains(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (tblfp) p7_tophits_TabularTargets(tblfp, qsq->name, qsq->acc, info->th, info->pli, (nquery == 1));
if (dfamtblfp) p7_tophits_TabularXfam(dfamtblfp, qsq->name, NULL, info->th, info->pli);
if (aliscoresfp) p7_tophits_AliScores(aliscoresfp, qsq->name, info->th );
esl_stopwatch_Stop(w);
info->pli->nseqs = 1;
p7_pli_Statistics(ofp, info->pli, w);
if (fprintf(ofp, "//\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
fflush(ofp);
p7_hmmfile_Close(hfp);
p7_pipeline_Destroy(info->pli);
p7_tophits_Destroy(info->th);
esl_sq_Reuse(qsq);
}
if (sstatus == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (sstatus != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
sstatus, sqfp->filename);
/* Terminate outputs - any last words?
*/
if (tblfp) p7_tophits_TabularTail(tblfp, "hmmscan", p7_SCAN_MODELS, cfg->seqfile, cfg->hmmfile, go);
if (ofp) { if (fprintf(ofp, "[ok]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
/* Cleanup - prepare for successful exit
*/
for (i = 0; i < infocnt; ++i)
p7_bg_Destroy(info[i].bg);
#ifdef HMMER_THREADS
if (ncpus > 0)
{
esl_workqueue_Reset(queue);
while (esl_workqueue_Remove(queue, (void **) &block) == eslOK)
p7_oprofile_DestroyBlock(block);
esl_workqueue_Destroy(queue);
esl_threads_Destroy(threadObj);
}
#endif
free(info);
esl_sq_Destroy(qsq);
esl_stopwatch_Destroy(w);
esl_alphabet_Destroy(abc);
esl_sqfile_Close(sqfp);
if (ofp != stdout) fclose(ofp);
if (tblfp) fclose(tblfp);
if (dfamtblfp) fclose(dfamtblfp);
if (aliscoresfp) fclose(aliscoresfp);
return eslOK;
ERROR:
if (ofp != stdout) fclose(ofp);
if (tblfp) fclose(tblfp);
if (dfamtblfp) fclose(dfamtblfp);
if (aliscoresfp) fclose(aliscoresfp);
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
char *maskfile = NULL; /* mask coordinate file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
int outfmt = eslSQFILE_FASTA; /* format code for output seqs */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
ESL_FILEPARSER *maskefp = NULL; /* open mask coord file */
FILE *ofp = NULL; /* output stream for masked seqs */
char *source = NULL; /* name of current seq to mask */
char *p1, *p2; /* pointers used in parsing */
int64_t start, end; /* start, end coord for masking */
int64_t i, j, pos; /* coords in a sequence */
int64_t overmask; /* # of extra residues to mask */
ESL_SQ *sq = esl_sq_Create(); /* current sequence */
int do_fetching;
int do_lowercase;
int maskchar;
int status; /* easel return code */
/****************************************************************************
* Parse command line
****************************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
do_fetching = esl_opt_GetBoolean(go, "-R");
do_lowercase = esl_opt_GetBoolean(go, "-l");
overmask = (esl_opt_IsOn(go, "-x") ? esl_opt_GetInteger(go, "-x") : 0);
maskchar = (esl_opt_IsOn(go, "-m") ? esl_opt_GetChar(go, "-m") : 'X');
seqfile = esl_opt_GetArg(go, 1);
maskfile = esl_opt_GetArg(go, 2);
/* Open the <seqfile>: text mode, not digital */
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
if (do_fetching && sqfp->data.ascii.ssi == NULL)
cmdline_failure(argv[0], "-R option (random access/fetching) requires %s to be SSI indexed\n", seqfile);
/* Open the <maskfile> */
if (esl_fileparser_Open(maskfile, NULL, &maskefp) != eslOK)
cmdline_failure(argv[0], "Failed to open mask coordinate file %s\n", maskfile);
esl_fileparser_SetCommentChar(maskefp, '#');
/* Open the output file, if any */
if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/****************************************************************************
* Main loop over lines in <maskfile>
****************************************************************************/
/* Read one data line at a time from the <maskfile>;
* parse into data fields <seqname> <start> <end>
*/
while (esl_fileparser_NextLine(maskefp) == eslOK)
{
/* First field is sequence name */
if (esl_fileparser_GetTokenOnLine(maskefp, &source, NULL) != eslOK)
esl_fatal("Failed to read source seq name on line %d of file %s\n", maskefp->linenumber, maskfile);
/* Get the sequence */
if (do_fetching)
{ /* If the <seqfile> is SSI indexed, try to reposition it and read <source> seq by random access */
status = esl_sqio_Fetch(sqfp, source, sq);
if (status == eslENOTFOUND) esl_fatal("seq %s not found in SSI index for file %s\n", source, sqfp->filename);
else if (status == eslEINVAL) esl_fatal("No SSI index or can't reposition in file %s\n", sqfp->filename);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected failure in fetching %s from file %s\n", source, sqfp->filename);
}
else
{ /* else, assume we're reading sequentially; <sqfile> and <maskfile> have seqs in same order */
status = esl_sqio_Read(sqfp, sq);
if (status == eslEOF) esl_fatal("File %s ended prematurely; didn't find %s\n", sqfp->filename, source);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected error reading sequence file %s\n", sqfp->filename);
if ((strcmp(sq->name, source) != 0) && (strcmp(sq->acc, source) != 0))
esl_fatal("Sequences in <sqfile> and <maskfile> aren't in same order; try -R");
}
/* If we're masking by lowercase, first make sure everything's uppercase */
if (do_lowercase)
for (pos = 0; pos < sq->n; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = toupper(sq->seq[pos]);
/* Next two fields are <start>, <end> for the masking */
/* possible future extension: wrap loop around this, enable multiple masked regions */
if (esl_fileparser_GetTokenOnLine(maskefp, &p1, NULL) != eslOK)
esl_fatal("Failed to read start coord on line %d of file %s\n", maskefp->linenumber, maskfile);
start = strtoll(p1, &p2, 0) - 1;
if (esl_fileparser_GetTokenOnLine(maskefp, &p2, NULL) != eslOK)
esl_fatal("Failed to read end coord on line %d of file %s\n", maskefp->linenumber, maskfile);
end = strtoll(p2, &p1, 0) - 1;
/* Do the masking */
if (esl_opt_GetBoolean(go, "-r")) /* Reverse masking */
{ /* leave start..end unmasked; mask prefix 0..start-1, end+1..L-1 */
i = 0;
j = ESL_MIN(sq->n-1, start - 1 + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
i = ESL_MAX(0, end + 1 - overmask);
j = sq->n-1;
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
else
{ /* normal: mask start..end */
i = ESL_MAX(0, start - overmask);
j = ESL_MIN(sq->n-1, end + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
esl_sqio_Write(ofp, sq, outfmt, FALSE);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
esl_fileparser_Close(maskefp);
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
if (ofp != stdout) fclose(ofp);
return 0;
}