int main(int argc, char* argv[]) {
FILE* F;
MSA *msa;
int *msa_gap_patterns = NULL;
HMM *hmm = NULL;
TreeNode *tree = NULL;
int i, input_format = SS, msa_idx, quiet_mode = FALSE,
ncats, nmsas, ncats_unspooled, indel_nseqs = -1;
String *msa_fname, *gff_fname;
List *gff_fname_list = NULL, *msa_fname_list = NULL,
*msa_length_list = NULL, *model_indels_str = NULL;
Matrix *traincounts = NULL;
Vector *begcounts = NULL, *statecounts = NULL;
CategoryMap *cm = NULL;
char c;
GapPatternMap *gpm = NULL;
GFF_Set *gff;
char *reverse_groups_tag = NULL;
while ((c = getopt(argc, argv, "i:g:c:m:M:R:I:n:t:P:G:qh")) != -1) {
switch(c) {
case 'i':
input_format = msa_str_to_format(optarg);
if (input_format == -1)
die("ERROR: bad alignment format.\n");
break;
case 'g':
gff_fname_list = get_arg_list(optarg);
break;
case 'c':
cm = cm_new_string_or_file(optarg);
break;
case 'm':
msa_fname_list = get_arg_list(optarg);
break;
case 'M':
msa_length_list = str_list_as_int(get_arg_list(optarg));
break;
case 'R':
reverse_groups_tag = optarg;
break;
case 'I':
model_indels_str = get_arg_list(optarg);
break;
case 'n':
indel_nseqs = get_arg_int(optarg);
break;
case 't':
if (optarg[0] == '(') /* in this case, assume topology given
at command line */
tree = tr_new_from_string(optarg);
else
tree = tr_new_from_file(phast_fopen(optarg, "r"));
break;
case 'q':
quiet_mode = TRUE;
break;
case 'h':
print_usage();
exit(0);
case '?':
die("ERROR: unrecognized option.\n\nType 'hmm_train -h' for usage.\n");
}
}
if (msa_fname_list == NULL)
die("ERROR: -m required. Type 'hmm_train -h' for usage.\n");
if (gff_fname_list == NULL)
die("ERROR: -g required in training mode. Type 'hmm_train -h' for usage.\n");
if (msa_length_list != NULL && msa_fname_list != NULL)
die("ERROR: -m and -M are mutually exclusive. Type 'hmm_train -h' for usage.\n");
if (model_indels_str != NULL && tree == NULL)
die("ERROR: -I requires -t. Type 'hmm_train -h' for usage.\n");
if (cm == NULL)
die("ERROR: category map required.\n");
set_seed(-1);
ncats = cm->ncats + 1;
ncats_unspooled = cm->unspooler != NULL ? cm->unspooler->nstates_unspooled :
ncats;
nmsas = (msa_length_list != NULL ? lst_size(msa_length_list) :
lst_size(msa_fname_list));
if (model_indels_str != NULL) {
if (tree == NULL)
die("ERROR: tree is NULL\n"); /*FIXME: indel_ncats broken */
gpm = gp_create_gapcats(cm, model_indels_str, tree, FALSE);
ncats = cm->ncats + 1; /* numbers will change */
ncats_unspooled = cm->unspooler == NULL ? ncats :
cm->unspooler->nstates_unspooled;
}
/* allocate memory for storage of "training paths" */
traincounts = mat_new(ncats_unspooled, ncats_unspooled);
statecounts = vec_new(ncats_unspooled);
begcounts = vec_new(ncats_unspooled);
mat_zero(traincounts);
vec_zero(statecounts);
vec_zero(begcounts);
/* create skeleton of new HMM. */
hmm = hmm_new_nstates(ncats_unspooled, 0, 0);
/* Main loop: consider each MSA in turn */
for (msa_idx = 0; msa_idx < nmsas; msa_idx++) {
if (msa_fname_list != NULL) {
msa_fname = (String*)lst_get_ptr(msa_fname_list, msa_idx);
F = phast_fopen(msa_fname->chars, "r");
if (!quiet_mode)
fprintf(stderr, "Reading alignment from %s ...\n",
F == stdin ? "stdin" : msa_fname->chars);
msa = msa_new_from_file(F, NULL);
phast_fclose(F);
}
else { /* only lengths of alignments specified */
msa = msa_new(NULL, NULL, 0, lst_get_int(msa_length_list, msa_idx), NULL);
/* just a shell in this case */
}
gff_fname = (String*)lst_get_ptr(gff_fname_list, msa_idx);
if (!quiet_mode)
fprintf(stderr, "Reading annotations from %s ...\n", gff_fname->chars);
gff = gff_read_set(phast_fopen(gff_fname->chars, "r"));
/* convert GFF to coordinate frame of alignment */
if (msa_length_list == NULL) {
if (!quiet_mode)
fprintf(stderr, "Mapping annotations to alignment ...\n");
msa_map_gff_coords(msa, gff, 1, 0, 0); /* assume seq 1 is ref */
}
if (model_indels_str != NULL) {
if (!quiet_mode)
fprintf(stderr, "Obtaining gap patterns ...\n");
msa_gap_patterns = smalloc(msa->length * sizeof(int));
gp_set_phylo_patterns(gpm, msa_gap_patterns, msa);
}
/* at this point, we don't actually need the alignment anymore;
if using ordered suff stats (likely with large data sets),
can free them now, to avoid running out of memory */
if (msa->ss != NULL) {
ss_free(msa->ss);
msa->ss = NULL;
}
if (reverse_groups_tag != NULL) {
if (!quiet_mode)
fprintf(stderr, "Reverse complementing features on negative strand (group by '%s') ...\n",
reverse_groups_tag);
/* we don't need to reverse complement the whole alignment --
just the gff and possibly the gap pattern array (pass a
NULL msa) */
gff_group(gff, reverse_groups_tag);
msa_reverse_compl_feats(NULL, gff, msa_gap_patterns);
}
if (!quiet_mode)
fprintf(stderr, "Labeling sites by category ...\n");
msa_label_categories(msa, gff, cm);
gff_free_set(gff);
if (model_indels_str != NULL) {
if (!quiet_mode)
fprintf(stderr, "Remapping categories according to gap patterns ...\n");
if (indel_nseqs > 0 && indel_nseqs != msa->nseqs) {
/* in this case, we'll simply reassign non-trivial gap
patterns randomly. This will achieve the desired
effect with minimal coding, as long as the number of
sites is not too small (the indel model is probably
useless anyway if the number is small) */
int pat, newpat;
int npatterns = 4 * indel_nseqs - 5;
int complex_allowed[cm->ncats+1];
List *no_complex_names, *no_complex_nums;
if (!quiet_mode)
fprintf(stderr, "(target number of sequences: %d)\n", indel_nseqs);
/* set up index indicating by cat no. whether complex gaps
are allowed */
for (i = 0; i < ncats; i++) complex_allowed[i] = 1;
no_complex_names = lst_new_ptr(10);
str_split(str_new_charstr(NO_COMPLEX), ",", no_complex_names);
no_complex_nums = cm_get_category_list(cm, no_complex_names, 1);
for (i = 0; i < lst_size(no_complex_nums); i++)
complex_allowed[lst_get_int(no_complex_nums, i)] = 0;
lst_free(no_complex_nums);
lst_free_strings(no_complex_names);
lst_free(no_complex_names);
/* now reassign all non-null numbers */
for (i = 0; i < msa->length; ) {
if ((pat = msa_gap_patterns[i]) != 0) {
if (complex_allowed[msa->categories[i]])
newpat = 1 + ((double)npatterns * unif_rand());
/* random number in interval [1, npatterns] */
else
newpat = 1 + ((double)(npatterns-1) * unif_rand());
/* random number in interval [1,npatterns-1]
(excludes complex gap pattern) */
for (; i < msa->length && msa_gap_patterns[i] == pat; i++)
msa_gap_patterns[i] = newpat; /* change for whole sequence */
}
else i++;
}
}
/* obtain gapped category number for each site */
for (i = 0; i < msa->length; i++)
if (gpm->cat_x_pattern_to_gapcat[msa->categories[i]] != NULL)
msa->categories[i] = gpm->cat_x_pattern_to_gapcat[msa->categories[i]][msa_gap_patterns[i]];
}
if (!quiet_mode)
fprintf(stderr, "Unspooling categories ...\n");
cm_spooled_to_unspooled(cm, msa->categories, msa->length);
if (!quiet_mode)
fprintf(stderr, "Collecting training data ...\n");
hmm_train_update_counts(traincounts, statecounts, begcounts,
msa->categories, msa->length,
ncats_unspooled);
if (msa_gap_patterns != NULL) sfree(msa_gap_patterns);
msa_free(msa);
}
/* now train HMM, using cumulative data */
hmm_train_from_counts(hmm, traincounts, NULL, statecounts, NULL,
begcounts, NULL);
/* if modeling indels, adjust begin transitions so probability is
distributed among different "gap pattern" states that all
correspond to the same ungapped state (category); this helps
avoid problems that occur when training on a few large sequences
(e.g., whole chromosomes) and then testing on many shorter ones */
if (model_indels_str != NULL) {
double tprob[gpm->ncats];
int nst[gpm->ncats]; /* total prob and number of states per
spooled, ungapped category */
for (i = 0; i < gpm->ncats; i++) tprob[i] = nst[i] = 0;
for (i = 0; i < hmm->nstates; i++) {
if (vec_get(hmm->begin_transitions, i) > 0)
/* have to go from unspooled space to spooled space, then to
ungapped space (HMM states correspond to unspooled,
gapped categories). Note that states with nonzero begin
probs shouldn't be conditioned on other states. */
tprob[gpm->gapcat_to_cat[cm_unspooled_to_spooled_cat(cm, i)]] +=
vec_get(hmm->begin_transitions, i);
nst[gpm->gapcat_to_cat[cm_unspooled_to_spooled_cat(cm, i)]]++;
}
for (i = 0; i < hmm->nstates; i++)
if (tprob[gpm->gapcat_to_cat[cm_unspooled_to_spooled_cat(cm, i)]] > 0)
vec_set(hmm->begin_transitions, i,
tprob[gpm->gapcat_to_cat[cm_unspooled_to_spooled_cat(cm, i)]] /
nst[gpm->gapcat_to_cat[cm_unspooled_to_spooled_cat(cm, i)]]);
/* (uniform prior) */
}
/* write trained HMM */
hmm_print(stdout, hmm);
if (!quiet_mode) fprintf(stderr, "Done.\n");
return 0;
}
int main(int argc, char* argv[]) {
FILE* F;
MSA *msa;
msa_format_type format = UNKNOWN_FORMAT;
int src_ref = -1, dest_ref = 0, offset = 0;
char *msa_fname = NULL, *feat_fname = NULL;
GFF_Set *gff;
char c;
while ((c = (char)getopt(argc, argv, "hm:f:s:d:i:p:n:")) != -1) {
switch(c) {
case 'm':
msa_fname = optarg;
break;
case 'f':
feat_fname = optarg;
break;
case 's':
src_ref = get_arg_int(optarg);
break;
case 'd':
dest_ref = get_arg_int(optarg);
break;
case 'i':
format = msa_str_to_format(optarg);
if (format == UNKNOWN_FORMAT) die("ERROR: bad alignment format.\n");
break;
case 'p':
offset = get_arg_int(optarg);
break;
case 'n':
offset = -1 * get_arg_int(optarg);
break;
case 'h':
print_usage();
exit(1);
case '?':
print_usage();
exit(1);
}
}
if (msa_fname == NULL || feat_fname == NULL) {
print_usage();
exit(1);
}
set_seed(-1);
F = phast_fopen(feat_fname, "r");
if ((gff = gff_read_set(F)) == NULL) {
die("ERROR: error reading %s.\n", feat_fname);
}
phast_fclose(F);
/* handle case of local alignment specially -- avoid representing
the alignment explicitly */
F = phast_fopen(msa_fname, "r");
if (format == UNKNOWN_FORMAT)
format = msa_format_for_content(F, 1);
if (format == LAV) {
LocalPwAlignment *lpwa = NULL;
/* int i; */
fprintf(stderr, "WARNING: in local alignment mode, coordinates may only be mapped from query (reference) sequence to target (aligned) sequence.\n");
lpwa = la_read_lav(F, 0);
la_gff_transform(lpwa, gff);
/* for (i = 0; i < lst_size(gff->features); i++) { */
/* GFF_Feature *feat = lst_get_ptr(gff->features, i); */
/* feat->start = la_get_target_coord(lpwa, feat->start); */
/* feat->end = la_get_target_coord(lpwa, feat->end); */
/* } */
}
else { /* normal alignment */
msa = msa_new_from_file_define_format(F, format, NULL);
phast_fclose(F);
msa_map_gff_coords(msa, gff, src_ref, dest_ref, offset);
msa_free(msa);
}
gff_print_set(stdout, gff);
gff_free_set(gff);
return 0;
}
int main(int argc, char *argv[]) {
char c;
int opt_idx, node;
FILE *out_f = NULL, *msa_f, *mod_f;
char *out_root;
TreeModel *mod;
MSA *msa;
char out_fname[STR_MED_LEN];
struct option long_opts[] = {
{"refseq", 1, 0, 'r'},
{"msa-format", 1, 0, 'i'},
{"seqs", 1, 0, 's'},
{"exclude", 0, 0, 'x'},
{"no-probs", 0, 0, 'n'},
{"suff-stats", 0, 0, 'S'},
{"encode", 1, 0, 'e'},
{"keep-gaps", 0, 0, 'k'},
{"gibbs", 1, 0, 'G'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
/* arguments and defaults for options */
FILE *refseq_f = NULL;
msa_format_type msa_format = UNKNOWN_FORMAT;
int suff_stats = FALSE, exclude = FALSE, keep_gaps = FALSE, do_probs = TRUE;
List *seqlist = NULL;
PbsCode *code = NULL;
int gibbs_nsamples = -1;
while ((c = (char)getopt_long(argc, argv, "r:i:s:e:knxSh", long_opts, &opt_idx)) != -1) {
switch (c) {
case 'r':
refseq_f = phast_fopen(optarg, "r");
break;
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == UNKNOWN_FORMAT)
die("ERROR: unrecognized alignment format.\n");
break;
case 'S':
suff_stats = TRUE;
break;
case 'e':
code = pbs_new_from_file(phast_fopen(optarg, "r"));
break;
case 's':
seqlist = get_arg_list(optarg);
break;
case 'x':
exclude = TRUE;
break;
case 'n':
do_probs = FALSE;
break;
case 'k':
keep_gaps = TRUE;
break;
case 'G':
gibbs_nsamples = get_arg_int_bounds(optarg, 1, INFTY);
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("Bad argument. Try 'prequel -h'.\n");
}
}
if (optind != argc - 3)
die("Three arguments required. Try 'prequel -h'.\n");
set_seed(-1);
if (!do_probs && (suff_stats || code != NULL))
die("ERROR: --no-probs can't be used with --suff-stats or --encode.\n");
msa_f = phast_fopen(argv[optind], "r");
if (msa_format == UNKNOWN_FORMAT)
msa_format = msa_format_for_content(msa_f, 1);
fprintf(stderr, "Reading alignment from %s...\n", argv[optind]);
if (msa_format == MAF) {
msa = maf_read(msa_f, refseq_f, 1, NULL, NULL, NULL, -1, !suff_stats, NULL,
NO_STRIP, FALSE);
/* (no need to store order if suff_stats mode) */
}
else
msa = msa_new_from_file_define_format(msa_f, msa_format, NULL);
if (msa->ss == NULL) {
fprintf(stderr, "Extracting sufficient statistics...\n");
ss_from_msas(msa, 1, TRUE, NULL, NULL, NULL, -1, 0);
}
else if (msa->ss->tuple_idx == NULL && !suff_stats)
die("ERROR: ordered representation of alignment required unless --suff-stats.\n");
mod_f = phast_fopen(argv[optind+1], "r");
out_root = argv[optind+2];
mod = tm_new_from_file(mod_f, 1);
/* MH prune just like in phastcons */
int old_nnodes = mod->tree->nnodes;
List *pruned_names = lst_new_ptr(msa->nseqs);
tm_prune(mod, msa, pruned_names);
if (lst_size(pruned_names) == (old_nnodes + 1) / 2)
die("ERROR: no match for leaves of tree in alignment (leaf names must match alignment names).\n");
if (lst_size(pruned_names) > 0) {
fprintf(stderr, "WARNING: pruned away leaves of tree with no match in alignment (");
int j;
for (j = 0; j < lst_size(pruned_names); j++)
fprintf(stderr, "%s%s", ((String*)lst_get_ptr(pruned_names, j))->chars, j < lst_size(pruned_names) - 1 ? ", " : ").\n");
}
lst_free_strings(pruned_names);
tr_name_ancestors(mod->tree);
if (mod->order != 0)
die("ERROR: Only single nucleotide models are supported.\n");
if (mod->nratecats > 1)
die("ERROR: Rate variation not supported.\n");
mod->tree_posteriors = tl_new_tree_posteriors(mod, msa, TRUE, FALSE,
FALSE, FALSE, FALSE, FALSE,
FALSE);
fprintf(stderr, "Computing posterior probabilities...\n");
if (gibbs_nsamples > 0)
die("ERROR: --gibbs not implemented yet.");
/* gb_sample_ancestral_seqs(mod, msa, mod->tree_posteriors, gibbs_nsamples); */
else
tl_compute_log_likelihood(mod, msa, NULL, NULL, -1, mod->tree_posteriors);
fprintf(stderr, "Reconstructing indels by parsimony...\n");
do_indels(msa, mod);
for (node = 0; node < mod->tree->nnodes; node++) {
int i, j;
TreeNode *n = lst_get_ptr(mod->tree->nodes, node);
if (n->lchild == NULL || n->rchild == NULL) continue;
if (seqlist != NULL) {
int in_list = str_in_list_charstr(n->name, seqlist);
if ((in_list && exclude) || (!in_list && !exclude))
continue;
}
fprintf(stderr, "Writing output for ancestral node '%s'...\n",
n->name);
if (suff_stats) {
if (out_f == NULL) {
sprintf(out_fname, "%s.stats", out_root);
out_f = phast_fopen(out_fname, "w+");
fprintf(out_f, "#count\t");
for (j = 0; j < mod->rate_matrix->size; j++)
fprintf(out_f, "p(%c)%c", mod->rate_matrix->states[j],
j == mod->rate_matrix->size - 1 ? '\n' : '\t');
}
for (i = 0; i < msa->ss->ntuples; i++) {
if (mod->tree_posteriors->base_probs[0][0][node][i] == -1)
continue; /* no base this node */
fprintf(out_f, "%.0f\t", msa->ss->counts[i]);
for (j = 0; j < mod->rate_matrix->size; j++) {
fprintf(out_f, "%f%c",
mod->tree_posteriors->base_probs[0][j][node][i],
j == mod->rate_matrix->size - 1 ? '\n' : '\t');
}
}
}
else if (code == NULL && do_probs) { /* ordinary sequence-by-sequence
output */
sprintf(out_fname, "%s.%s.probs", out_root, n->name);
out_f = phast_fopen(out_fname, "w+");
fprintf(out_f, "#");
for (j = 0; j < mod->rate_matrix->size; j++)
fprintf(out_f, "p(%c)%c", mod->rate_matrix->states[j],
j == mod->rate_matrix->size - 1 ? '\n' : '\t');
for (i = 0; i < msa->length; i++) {
if (mod->tree_posteriors->base_probs[0][0][node][msa->ss->tuple_idx[i]] == -1) {
/* no base */
if (keep_gaps) fprintf(out_f, "-\n");
/* otherwise do nothing */
}
else
for (j = 0; j < mod->rate_matrix->size; j++)
fprintf(out_f, "%f%c",
mod->tree_posteriors->base_probs[0][j][node][msa->ss->tuple_idx[i]],
j == mod->rate_matrix->size - 1 ? '\n' : '\t');
}
phast_fclose(out_f);
}
else if (code == NULL && !do_probs) { /* write point estimates
to FASTA file */
char *outseq = smalloc((msa->length + 1) * sizeof(char));
int len = 0;
for (i = 0; i < msa->length; i++) {
if (mod->tree_posteriors->base_probs[0][0][node][msa->ss->tuple_idx[i]] == -1) {
/* no base */
if (keep_gaps) outseq[len++] = GAP_CHAR;
/* otherwise do nothing */
}
else {
double maxprob = 0;
int maxidx = -1;
for (j = 0; j < mod->rate_matrix->size; j++) {
if (mod->tree_posteriors->base_probs[0][j][node][msa->ss->tuple_idx[i]] > maxprob) {
maxprob = mod->tree_posteriors->base_probs[0][j][node][msa->ss->tuple_idx[i]];
maxidx = j;
}
}
outseq[len++] = mod->rate_matrix->states[maxidx];
}
}
outseq[len] = '\0';
/* print in FASTA format */
sprintf(out_fname, "%s.%s.fa", out_root, n->name);
out_f = phast_fopen(out_fname, "w+");
print_seq_fasta(out_f, outseq, n->name, len);
phast_fclose(out_f);
sfree(outseq);
}
else { /* encoded sequence-by-sequence
output */
double error, tot_error = 0;
int ngaps = 0;
Vector *v;
unsigned *encoded;
/* first encode tuple by tuple */
v = vec_new(mod->rate_matrix->size);
encoded = smalloc(msa->ss->ntuples * sizeof(unsigned));
for (i = 0; i < msa->ss->ntuples; i++) {
if (mod->tree_posteriors->base_probs[0][0][node][i] == -1) {
encoded[i] = code->gap_code;
ngaps += msa->ss->counts[i];
}
else {
for (j = 0; j < mod->rate_matrix->size; j++)
vec_set(v, j, mod->tree_posteriors->base_probs[0][j][node][i]);
encoded[i] = pbs_get_index(code, v, &error);
tot_error += error * msa->ss->counts[i];
}
}
vec_free(v);
/* now write site by site */
sprintf(out_fname, "%s.%s.bin", out_root, n->name);
out_f = phast_fopen(out_fname, "w+");
for (i = 0; i < msa->length; i++) {
if (keep_gaps || encoded[msa->ss->tuple_idx[i]] != code->gap_code)
pbs_write_binary(code, encoded[msa->ss->tuple_idx[i]], out_f);
}
fprintf(stderr, "Average approximation error ('%s'): %f bits\n",
n->name, tot_error/(msa->length - ngaps));
sfree(encoded);
}
}
fprintf(stderr, "Done.\n");
return 0;
}
int main(int argc, char *argv[]) {
int check_start = 0, check_stop = 0, check_splice = 0, check_nonsense = 0,
offset5 = 0, offset3 = 0, opt_idx, i, j, indel_strict = 0, no_output = 0,
check_alignment = 0, splice_strict = 0;
int ncons_tested, nkept, nconserved_exons;
int nce_gap_type[NGAP_TYPES], nconsid[NTYPES], nfail[NTYPES];
double Nfrac = 0.05;
char c;
MSA *msa;
GFF_Set *gff;
msa_format_type msa_format = UNKNOWN_FORMAT;
List *keepers, *problems = lst_new_ptr(10),
*ends_adjusted = lst_new_ptr(1), *starts_adjusted = lst_new_ptr(1),
*discards=NULL, *intron_splice = lst_new_ptr(10);
char *rseq_fname = NULL;
FILE *logf = NULL, *mlogf = NULL, *statsf = NULL, *discardf = NULL;
cds_gap_type fshift_mode = FSHIFT_BAD;
char *groupby = "transcript_id";
msa_coord_map *map;
int *countNs, *countCDSs;
FILE *infile;
char *msa_fname;
struct option long_opts[] = {
{"start", 0, 0, 's'},
{"stop", 0, 0, 't'},
{"splice", 0, 0, 'l'},
{"nonsense", 0, 0, 'n'},
{"fshift", 0, 0, 'f'},
{"conserved", 0, 0, 'c'},
{"N-limit", 1, 0, 'N'},
{"clean-gaps", 0, 0, 'e'},
{"indel-strict", 0, 0, 'I'},
{"splice-strict", 0, 0, 'C'},
{"groupby", 1, 0, 'g'},
{"msa-format", 1, 0, 'i'},
{"refseq", 1, 0, 'r'},
{"offset5", 1, 0, 'o'},
{"offset3", 1, 0, 'p'},
{"no-output", 0, 0, 'x'},
{"discards", 1, 0, 'd'},
{"log", 1, 0, 'L'},
{"machine-log", 1, 0, 'M'},
{"stats", 1, 0, 'S'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
while ((c = (char)getopt_long(argc, argv, "N:i:r:L:M:S:g:d:stlnfceICxh",
long_opts, &opt_idx)) != -1) {
switch(c) {
case 's':
check_alignment = check_start = 1;
break;
case 't':
check_alignment = check_stop = 1;
break;
case 'l':
check_alignment = check_splice = 1;
break;
case 'n':
check_alignment = check_nonsense = 1;
break;
case 'f':
check_alignment = 1;
fshift_mode = FSHIFT_OK;
break;
case 'c':
check_alignment = check_start = check_stop = check_splice = check_nonsense = 1;
if (fshift_mode < FSHIFT_OK) fshift_mode = FSHIFT_OK;
break;
case 'N':
Nfrac = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'e':
check_alignment = 1;
if (fshift_mode < CLN_GAPS) fshift_mode = CLN_GAPS;
break;
case 'I':
check_alignment = 1;
fshift_mode = NOVRLP_CLN_GAPS;
indel_strict = 1;
break;
case 'C':
check_alignment = check_splice = splice_strict = 1;
break;
case 'g':
groupby = optarg;
break;
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == UNKNOWN_FORMAT) die("Bad alignment format.\n");
break;
case 'r':
rseq_fname = optarg;
break;
case 'o':
offset5 = get_arg_int(optarg);
break;
case 'p':
offset3 = get_arg_int(optarg);
break;
case 'L':
logf = phast_fopen(optarg, "w+");
break;
case 'M':
mlogf = phast_fopen(optarg, "w+");
break;
case 'S':
statsf = phast_fopen(optarg, "w+");
break;
case 'd':
discardf = phast_fopen(optarg, "w+");
break;
case 'x':
no_output = 1;
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("ERROR: Bad argument. Try the --help option.\n");
}
}
if (optind + 1 >= argc ) {
die("ERROR: Missing required arguments. Try the --help option.\n");
}
set_seed(-1);
gff = gff_read_set(phast_fopen(argv[optind], "r"));
msa_fname = argv[optind+1];
infile = phast_fopen(msa_fname, "r");
if (msa_format == UNKNOWN_FORMAT)
msa_format = msa_format_for_content(infile, 1);
if (msa_format == MAF) {
msa = maf_read(infile,
rseq_fname == NULL ? NULL : phast_fopen(rseq_fname, "r"),
1, NULL, NULL, NULL, -1, TRUE, NULL, NO_STRIP, FALSE);
}
else {
msa = msa_new_from_file_define_format(infile,
msa_format, NULL);
if (msa->ss == NULL)
ss_from_msas(msa, 1, 1, NULL, NULL, NULL, -1, 0);
}
if (!msa->ss->tuple_idx)
die("ERROR: need ordered tuples\n");
msa_remove_N_from_alph(msa); /* for backward compatibility (old SS files) */
if (msa->idx_offset != 0) { /* avoids offset problem */
for (i = 0; i < lst_size(gff->features); i++) {
GFF_Feature *f = lst_get_ptr(gff->features, i);
f->start -= msa->idx_offset;
f->end -= msa->idx_offset;
}
}
/* set up coordinate map; assume GFF is for sequence 1 */
map = msa_build_coord_map(msa, 1);
/* convert all features */
for (i = 0; i < lst_size(gff->features); i++) {
GFF_Feature *f = lst_get_ptr(gff->features, i);
int newstart, newend;
if (f->start < 0 || f->end < f->start)
die("ERROR: bad feature in GFF (start=%d, end=%d).\n",
f->start, f->end);
newstart = msa_map_seq_to_msa(map, f->start);
newend = msa_map_seq_to_msa(map, f->end);
if (newstart < 0 || newend < newstart)
die("ERROR: unable to map coordinates for feature (start=%d, end=%d).\n",
f->start, f->end);
f->start = newstart;
f->end = newend;
}
gff_group(gff, groupby); /* do this after coord conversion, or
group coords and feature coords
will be out of sync */
keepers = lst_new_ptr(lst_size(gff->features));
if (discardf != NULL) discards = lst_new_ptr(lst_size(gff->features));
ncons_tested = nkept = nconserved_exons = 0;
for (i = 0; i < NTYPES; i++) nconsid[i] = 0;
for (i = 0; i < NTYPES; i++) nfail[i] = 0;
for (i = 0; i < NGAP_TYPES; i++) nce_gap_type[i] = 0;
countNs = smalloc(msa->nseqs * sizeof(int));
countCDSs = smalloc(msa->nseqs * sizeof(int));
for (i = 0; i < lst_size(gff->groups); i++) {
GFF_FeatureGroup *group = lst_get_ptr(gff->groups, i);
List *gfeatures = group->features;
GFF_Feature *feat;
status_type status = OKAY;
cds_gap_type gt = FSHIFT_BAD;
problems_clear(problems);
/* make sure have frame info for CDSs */
for (j = 0; j < lst_size(gfeatures); j++) {
feat = lst_get_ptr(gfeatures, j);
if (str_equals_charstr(feat->feature, GFF_CDS_TYPE) &&
feat->frame == GFF_NULL_FRAME)
die("ERROR: Missing frame info for CDS.\n");
}
/* First, exclude stop codons from cds's, if necessary (simplifies
the detection of nonsense mutations). */
exclude_stops(group, starts_adjusted, ends_adjusted);
/* In all cases, discard any group for which the reference sequence
doesn't have valid splice sites or start/stop codons, or has a
premature stop codon */
if (!ref_seq_okay(gfeatures, msa, offset3, indel_strict, splice_strict,
problems)) {
status = BAD_REF;
nfail[BAD_REF]++;
}
else
/* Everything else counts as a potentially valid group */
ncons_tested++;
if (status == OKAY && check_alignment) {
/* only bother with below if
interested in cross-species
conservation */
/* Check first to make sure there's alignment across species in
the cds; if not, there's no need to look at individual
features. */
for (j = 0; j < lst_size(gfeatures); j++) {
feat = lst_get_ptr(gfeatures, j);
if (str_equals_charstr(feat->feature, GFF_CDS_TYPE) &&
is_incomplete_alignment(feat, msa)) {
status = NO_ALN;
nfail[NO_ALN]++;
problem_add(problems, feat, NO_ALN, -1, -1);
break;
}
}
if (status == OKAY) { /* we have alignment and agreement
with the ref seq; now check feature
by feature */
lst_clear(intron_splice);
for (j = 0; j < msa->nseqs; j++) countNs[j] = countCDSs[j] = 0;
for (j = 0; j < lst_size(gfeatures); j++) {
feat = lst_get_ptr(gfeatures, j);
if (feat->end - 1 >= msa->length)
die("ERROR: feature extends beyond alignment (%d >= %d).\n",
feat->end - 1, msa->length);
if (check_start && str_equals_charstr(feat->feature, GFF_START_TYPE)) {
nconsid[BAD_START]++;
if (!is_conserved_start(feat, msa)) {
status = BAD_START;
problem_add(problems, feat, BAD_START, -1, -1);
}
}
else if (check_stop && str_equals_charstr(feat->feature, GFF_STOP_TYPE)) {
nconsid[BAD_STOP]++;
if (!is_conserved_stop(feat, msa)) {
status = BAD_STOP;
problem_add(problems, feat, BAD_STOP, -1, -1);
}
}
else if (check_splice &&
str_equals_charstr(feat->feature, SPLICE_5)) {
nconsid[BAD_5_SPLICE]++;
if (!is_conserved_5splice(feat, msa, offset5, splice_strict)) {
status = BAD_5_SPLICE;
problem_add(problems, feat, BAD_5_SPLICE, -1, -1);
}
else lst_push_ptr(intron_splice, feat);
}
else if (check_splice &&
str_equals_charstr(feat->feature, SPLICE_5_UTR)) {
nconsid[BAD_5_SPLICE_UTR]++;
if (!is_conserved_5splice(feat, msa, offset5, splice_strict)) {
status = BAD_5_SPLICE_UTR;
problem_add(problems, feat, BAD_5_SPLICE_UTR, -1, -1);
}
else lst_push_ptr(intron_splice, feat);
}
else if (check_splice && str_equals_charstr(feat->feature, SPLICE_3)) {
nconsid[BAD_3_SPLICE]++;
if (!is_conserved_3splice(feat, msa, offset3, splice_strict)) {
status = BAD_3_SPLICE;
problem_add(problems, feat, BAD_3_SPLICE, -1, -1);
}
else lst_push_ptr(intron_splice, feat);
}
else if (check_splice && str_equals_charstr(feat->feature, SPLICE_3)) {
nconsid[BAD_3_SPLICE_UTR]++;
if (!is_conserved_3splice(feat, msa, offset3, splice_strict)) {
status = BAD_3_SPLICE_UTR;
problem_add(problems, feat, BAD_3_SPLICE_UTR, -1, -1);
}
else lst_push_ptr(intron_splice, feat);
}
else if (str_equals_charstr(feat->feature, GFF_CDS_TYPE)) {
if (fshift_mode > FSHIFT_BAD
&& (gt = get_cds_gap_type(feat, msa, problems)) < fshift_mode) {
if (status == OKAY || status == NONSENSE) status = FSHIFT;
}
if (check_nonsense && !is_nonsense_clean(feat, msa, problems)) {
if (status == OKAY) status = NONSENSE;
}
if (Nfrac < 1)
get_N_counts(countNs, countCDSs, feat, msa);
}
} /* end loop through features in group */
/* still have to make sure splice sites are paired correctly
(GT-AG, GC-AG, AT-AC) */
if (status == OKAY && !splice_strict && lst_size(intron_splice) >= 2 &&
!are_introns_okay(intron_splice, msa, problems, offset5, offset3))
status = BAD_INTRON;
/* also check fraction of Ns */
if (Nfrac < 1) {
enum {MY_OKAY, MY_FAIL, MY_WARN} Nstatus = MY_OKAY;
for (j = 0; j < msa->nseqs; j++) {
if ((double)countNs[j] / countCDSs[j] > Nfrac) Nstatus = MY_FAIL;
if (Nstatus == MY_OKAY && countNs[j] > 0) Nstatus = MY_WARN;
}
if (Nstatus == MY_FAIL) {
problem_add(problems, NULL, TOO_MANY_Ns, -1, -1);
if (status == OKAY) status = TOO_MANY_Ns;
}
else if (Nstatus == MY_WARN)
problem_add(problems, NULL, WARN_Ns, -1, -1);
}
/* if collecting stats, record counts for failures */
if (statsf != NULL) {
if (status != OKAY) {
for (j = 0; j < lst_size(problems); j++) {
struct Problem *problem = lst_get_ptr(problems, j);
status_type ftype = problem->status;
if ((ftype == FSHIFT || ftype == NONSENSE) &&
status != FSHIFT && status != NONSENSE)
continue; /* don't count secondary frame shifts
and nonsense mutations */
if (ftype == BAD_INTRON && j % 2 == 0)
continue; /* only count one of every pair of these */
nfail[ftype]++;
}
}
/* also keep track of the total number of "conserved exons", and
the number having each kind of gap */
if ((status == OKAY || (status == FSHIFT && gt >= FSHIFT_OK))) {
nconserved_exons++;
nce_gap_type[gt]++; /* number of conserved exons having
given type of gaps */
}
}
} /* end if (status == OKAY) [checks for conserved features] */
} /* end if (status == OKAY && check_alignment) [all cross-species
checks] */
/* now we have looked at the whole group; we just need to do some
final accounting and logging */
if (status == OKAY) {
nkept++;
if (!no_output) {
restore_stops(group, starts_adjusted, ends_adjusted);
for (j = 0; j < lst_size(gfeatures); j++)
lst_push_ptr(keepers, lst_get_ptr(gfeatures, j));
}
if (logf != NULL && lst_size(problems) > 0) /* warnings only */
write_log(logf, group, status, problems, msa, map);
if (mlogf != NULL) {
/* no problem, need to add an okay status to log */
problem_add(problems, NULL, OKAY, -1, -1);
write_machine_log(mlogf, group, problems, map); /* may include
warnings */
}
}
else {
if (discardf != NULL) {
restore_stops(group, starts_adjusted, ends_adjusted);
for (j = 0; j < lst_size(gfeatures); j++)
lst_push_ptr(discards, lst_get_ptr(gfeatures, j));
}
if (logf != NULL)
write_log(logf, group, status, problems, msa, map);
if (mlogf != NULL)
write_machine_log(mlogf, group, problems, map);
}
} /* end loop over groups */
/* write main output and discards */
if (!no_output || discardf != NULL) {
/* first map features back to coord frame of reference seq. */
for (i = 0; i < lst_size(gff->features); i++) {
GFF_Feature *f = lst_get_ptr(gff->features, i);
f->start = msa_map_msa_to_seq(map, f->start) + msa->idx_offset;
f->end = msa_map_msa_to_seq(map, f->end) + msa->idx_offset;
}
if (!no_output) {
gff->features = keepers;
gff_print_set(stdout, gff);
}
if (discardf != NULL) {
gff->features = discards;
gff_print_set(discardf, gff);
}
}
/* dump counts to stats file */
if (statsf != NULL) {
fprintf(statsf, "#%11s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s %12s\n",
"total", "nbad_ref", "nconsid", "nkept", "nno_aln",
"nbad_starts", "(out of)", "nbad_stops", "(out of)",
"nbad_5spl", "(out of)", "nbad_3spl", "(out of)",
"nbad_5utr", "(out of)", "nbad_3utr", "(out of)",
"nbad_intron", "nnons", "nfshifts", "nNs", "ncons_exons",
"nce_ngaps", "nce_nov_cln", "nce_clean", "nce_fshftok");
fprintf(statsf, "%12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d %12d\n",
nfail[BAD_REF]+ncons_tested, nfail[BAD_REF], ncons_tested, nkept,
nfail[NO_ALN], nfail[BAD_START], nconsid[BAD_START],
nfail[BAD_STOP], nconsid[BAD_STOP], nfail[BAD_5_SPLICE],
nconsid[BAD_5_SPLICE], nfail[BAD_3_SPLICE], nconsid[BAD_3_SPLICE],
nfail[BAD_5_SPLICE_UTR], nconsid[BAD_5_SPLICE_UTR],
nfail[BAD_3_SPLICE_UTR], nconsid[BAD_3_SPLICE_UTR],
nfail[BAD_INTRON], nfail[NONSENSE], nfail[FSHIFT],
nfail[TOO_MANY_Ns], nconserved_exons, nce_gap_type[NGAPS],
nce_gap_type[NOVRLP_CLN_GAPS], nce_gap_type[CLN_GAPS],
nce_gap_type[FSHIFT_OK]);
fprintf(statsf, "%s", STATS_DESCRIPTION);
}
if (logf != NULL) phast_fclose(logf);
if (mlogf != NULL) phast_fclose(mlogf);
if (statsf != NULL) phast_fclose(statsf);
if (discardf != NULL) phast_fclose(discardf);
return 0;
}
int main(int argc, char *argv[]) {
char *msa_fname = NULL, *alph = "ACGT";
msa_format_type input_format = UNKNOWN_FORMAT;
char c;
int opt_idx, seed=-1;
String *optstr;
List *tmplist = NULL;
struct phyloFit_struct *pf;
FILE *infile;
struct option long_opts[] = {
{"msa", 1, 0, 'm'},
{"tree", 1, 0, 't'},
{"subst-mod", 1, 0, 's'},
{"msa-format", 1, 0, 'i'},
{"nrates", 1, 0, 'k'},
{"alpha", 1, 0, 'a'},
{"features", 1, 0, 'g'},
{"catmap", 1, 0, 'c'},
{"log", 1, 0, 'l'},
{"out-root", 1, 0, 'o'},
{"EM", 0, 0, 'E'},
{"error", 1, 0, 'e'},
{"precision", 1, 0, 'p'},
{"do-cats", 1, 0, 'C'},
{"non-overlapping", 0, 0, 'V'},
{"markov", 0, 0, 'N'},
{"reverse-groups", 1, 0, 'R'},
{"init-model", 1, 0, 'M'},
{"init-random", 0, 0, 'r'},
{"init-parsimony", 0, 0, 'y'},
{"print-parsimony", 1, 0, 'Y'},
{"lnl", 0, 0, 'L'},
{"scale-only", 0, 0, 'B'},
{"scale-subtree", 1, 0, 'S'},
{"estimate-freqs", 0, 0, 'F'},
{"sym-freqs", 0, 0, 'W'},
{"no-freqs", 0, 0, 'f'},
{"no-rates", 0, 0, 'n'},
{"no-opt", 1, 0, 'O'},
{"min-informative", 1, 0, 'I'},
{"gaps-as-bases", 0, 0, 'G'},
{"quiet", 0, 0, 'q'},
{"help", 0, 0, 'h'},
{"windows", 1, 0, 'w'},
{"windows-explicit", 1, 0, 'v'},
{"ancestor", 1, 0, 'A'},
{"post-probs", 0, 0, 'P'},
{"expected-subs", 0, 0, 'X'},
{"expected-total-subs", 0, 0, 'Z'},
{"expected-subs-col", 0, 0, 'J'},
{"column-probs", 0, 0, 'U'},
{"rate-constants", 1, 0, 'K'},
{"ignore-branches", 1, 0, 'b'},
{"clock", 0, 0, 'z'},
{"alt-model", 1, 0, 'd'},
{"label-branches", 1, 0, 0},
{"label-subtree", 1, 0, 0},
{"selection", 1, 0, 0},
{"bound", 1, 0, 'u'},
{"seed", 1, 0, 'D'},
{0, 0, 0, 0}
};
// NOTE: remaining shortcuts left: HjQx
pf = phyloFit_struct_new(0);
while ((c = (char)getopt_long(argc, argv, "m:t:s:g:c:C:i:o:k:a:l:w:v:M:p:A:I:K:S:b:d:O:u:Y:e:D:GVENRqLPXZUBFfnrzhWyJ", long_opts, &opt_idx)) != -1) {
switch(c) {
case 'm':
msa_fname = optarg;
break;
case 't':
if (optarg[0] == '(') /* in this case, assume topology given
at command line */
pf->tree = tr_new_from_string(optarg);
else
pf->tree = tr_new_from_file(phast_fopen(optarg, "r"));
break;
case 's':
pf->subst_mod = tm_get_subst_mod_type(optarg);
if (pf->subst_mod == UNDEF_MOD)
die("ERROR: illegal substitution model. Type \"phyloFit -h\" for usage.\n");
break;
case 'g':
pf->gff = gff_read_set(phast_fopen(optarg, "r"));
break;
case 'c':
pf->cm = cm_new_string_or_file(optarg);
break;
case 'C':
pf->cats_to_do_str = get_arg_list(optarg);
break;
case 'V':
pf->nonoverlapping = TRUE;
break;
case 'o':
pf->output_fname_root = optarg;
break;
case 'k':
pf->nratecats = get_arg_int_bounds(optarg, 0, INFTY);
break;
case 'a':
pf->alpha = get_arg_dbl(optarg);
break;
case 'R':
pf->reverse_group_tag = optarg;
break;
case 'i':
input_format = msa_str_to_format(optarg);
if (input_format == UNKNOWN_FORMAT)
die("ERROR: unrecognized alignment format. Type 'phyloFit -h' for usage.\n");
break;
case 'l':
if (!strcmp(optarg, "-"))
pf->logf = stderr;
else pf->logf = phast_fopen(optarg, "w+");
break;
case 'N':
pf->use_conditionals = 1;
break;
case 'w':
tmplist = get_arg_list(optarg);
if (lst_size(tmplist) != 2 ||
str_as_int(lst_get_ptr(tmplist, 0), &(pf->window_size)) != 0 ||
str_as_int(lst_get_ptr(tmplist, 1), &(pf->window_shift)) != 0)
die("ERROR: illegal arguments to --windows.\n");
lst_free_strings(tmplist);
lst_free(tmplist);
break;
case 'v':
tmplist = get_arg_list(optarg);
if (lst_size(tmplist) % 2 != 0)
die("ERROR: argument to --windows-explicit must be a list of even length.\n");
pf->window_coords = str_list_as_int(tmplist);
lst_free(tmplist);
break;
case 'E':
pf->use_em = TRUE;
break;
case 'e':
pf->error_fname=optarg;
break;
case 'p':
if (!strcmp(optarg, "LOW")) pf->precision = OPT_LOW_PREC;
else if (!strcmp(optarg, "MED")) pf->precision = OPT_MED_PREC;
else if (!strcmp(optarg, "HIGH")) pf->precision = OPT_HIGH_PREC;
else if (!strcmp(optarg, "VERY_HIGH")) pf->precision = OPT_VERY_HIGH_PREC;
else die("ERROR: --precision must be LOW, MED, or HIGH.\n\n");
break;
case 'M':
pf->input_mod = tm_new_from_file(phast_fopen(optarg, "r"), 1);
break;
case 'r':
pf->random_init = TRUE;
break;
case 'y':
pf->init_parsimony = TRUE;
break;
case 'Y':
pf->init_parsimony = TRUE;
pf->parsimony_cost_fname = optarg;
pf->parsimony_only=TRUE;
break;
case 'L':
pf->likelihood_only = TRUE;
break;
case 'q':
pf->quiet = TRUE;
break;
case 'P':
pf->do_bases = TRUE;
break;
case 'X':
pf->do_expected_nsubst = TRUE;
break;
case 'Z':
pf->do_expected_nsubst_tot = TRUE;
break;
case 'J':
pf->do_expected_nsubst_col = TRUE;
break;
case 'U':
pf->likelihood_only = TRUE; /* force -L */
pf->nsites_threshold = 0; /* also force this; typical use is
with small number of tuples, no
tuple_idx */
pf->do_column_probs = TRUE;
break;
case 'A':
pf->root_seqname = optarg;
break;
case 'I':
pf->nsites_threshold = get_arg_int(optarg);
break;
case 'G':
pf->gaps_as_bases = TRUE;
alph = "ACGT-";
break;
case 'B':
pf->estimate_scale_only = TRUE;
break;
case 'S':
pf->subtree_name = optarg;
break;
case 'F':
pf->estimate_backgd = TRUE;
break;
case 'W':
pf->estimate_backgd = TRUE;
pf->symfreq = TRUE;
break;
case 'f':
pf->no_freqs = TRUE;
break;
case 'n':
pf->no_rates = TRUE;
break;
case 'K':
tmplist = get_arg_list(optarg);
pf->rate_consts = str_list_as_dbl(tmplist);
pf->nratecats = lst_size(pf->rate_consts);
pf->use_em = 1;
lst_free_strings(tmplist); lst_free(tmplist);
break;
case 'b':
pf->ignore_branches = get_arg_list(optarg);
break;
case 'z':
pf->assume_clock = TRUE;
break;
case 'O':
if (pf->nooptstr == NULL)
pf->nooptstr = str_new_charstr(optarg);
else die("ERROR: no-opt argument can only be used once! parameters can be comma-separated list.");
break;
case 'd':
if (pf->alt_mod_str == NULL) {
pf->alt_mod_str = lst_new_ptr(1);
}
optstr = str_new_charstr(optarg);
lst_push_ptr(pf->alt_mod_str, optstr);
break;
case 0:
if (strcmp(long_opts[opt_idx].name, "label-branches") == 0 ||
strcmp(long_opts[opt_idx].name, "label-subtree") == 0) {
optstr = str_new_charstr(optarg);
if (pf->label_str == NULL) {
pf->label_str = lst_new_ptr(3);
pf->label_type = lst_new_int(3);
}
lst_push_ptr(pf->label_str, optstr);
lst_push_int(pf->label_type,
strcmp(long_opts[opt_idx].name, "label-branches") == 0 ?
BRANCH_TYPE : SUBTREE_TYPE);
}
else if (strcmp(long_opts[opt_idx].name, "selection") == 0) {
pf->selection = get_arg_dbl(optarg);
pf->use_selection = TRUE;
}
else {
die("ERROR: unknown option. Type 'phyloFit -h' for usage.\n");
}
break;
case 'u':
if (pf->bound_arg == NULL)
pf->bound_arg = lst_new_ptr(1);
optstr = str_new_charstr(optarg);
lst_push_ptr(pf->bound_arg, optstr);
break;
case 'D':
seed = get_arg_int_bounds(optarg, 1, INFTY);
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("ERROR: illegal argument. Type 'phyloFit -h' for usage.\n");
}
}
set_seed(seed);
if (msa_fname == NULL) {
if (optind >= argc)
die("ERROR: missing alignment filename. Type 'phyloFit -h' for usage.\n");
msa_fname = argv[optind];
pf->msa_fname = msa_fname;
}
infile = phast_fopen(msa_fname, "r");
if (input_format == UNKNOWN_FORMAT)
input_format = msa_format_for_content(infile, 1);
if (pf->nonoverlapping && (pf->use_conditionals || pf->gff != NULL ||
pf->cats_to_do_str || input_format == SS))
die("ERROR: cannot use --non-overlapping with --markov, --features,\n--msa-format SS, or --do-cats.\n");
/* read alignment */
if (!pf->quiet) fprintf(stderr, "Reading alignment from %s ...\n", msa_fname);
if (input_format == MAF) {
pf->msa = maf_read(infile, NULL,
tm_order(pf->subst_mod) + 1,
NULL, pf->gff, pf->cm,
pf->nonoverlapping ? tm_order(pf->subst_mod) + 1 : -1,
FALSE, pf->reverse_group_tag, NO_STRIP, FALSE);
if (pf->gaps_as_bases)
msa_reset_alphabet(pf->msa, alph);
}
else
pf->msa = msa_new_from_file_define_format(infile,
input_format, alph);
/* set up for categories */
/* first label sites, if necessary */
pf->label_categories = (input_format != MAF);
run_phyloFit(pf);
if (pf->logf != NULL && pf->logf != stderr && pf->logf != stdout)
phast_fclose(pf->logf);
if (!pf->quiet) fprintf(stderr, "Done.\n");
sfree(pf);
return 0;
}
int main(int argc, char *argv[]) {
struct phastCons_struct *p = phastCons_struct_new(0);
struct option long_opts[] = {
{"states", 1, 0, 'S'},
{"hmm", 1, 0, 'H'},
{"viterbi", 1, 0, 'V'},
{"most-conserved", 1, 0, 'V'}, /* same as --viterbi */
{"no-post-probs", 0, 0, 'n'},
{"msa-format", 1, 0, 'i'},
{"FC", 0, 0, 'X'},
{"lambda", 1, 0, 'l'},
{"target-coverage", 1, 0, 'C'},
{"transitions", 1, 0, 't'},
{"expected-length", 1, 0, 'E'},
{"expected-lengths", 1, 0, 'E'}, /* for backward compatibility */
{"estimate-trees", 1, 0, 'T'},
{"estimate-rho", 1, 0, 'O'},
{"rho", 1, 0, 'R'},
{"gc", 1, 0, 'G'},
{"ignore-missing", 0, 0, 'z'},
{"nrates", 1, 0, 'k'},
{"log", 1, 0, 'g'},
{"refidx", 1, 0, 'r'},
{"suppress-missing", 0, 0, 'x'}, /* for backward compatibility */
{"reflect-strand", 1, 0, 'U'},
{"catmap", 1, 0, 'c'},
{"extrapolate", 1, 0, 'e'},
{"indels", 0, 0, 'I'},
{"max-micro-indel", 1, 0, 'Y'},
{"indel-params", 1, 0, 'D'},
{"min-informative-types", 1, 0, 'M'}, /* for backward compatibility */
{"require-informative", 1, 0, 'M'},
{"not-informative", 1, 0, 'F'},
{"lnl", 1, 0, 'L'},
{"seqname", 1, 0, 'N'},
{"idpref", 1, 0, 'P'},
{"score", 0, 0, 's'},
{"coding-potential", 0, 0, 'p'},
{"indels-only", 0, 0, 'J'},
{"alias", 1, 0, 'A'},
{"quiet", 0, 0, 'q'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
/* other vars */
FILE *infile;
char *msa_fname;
char c;
int opt_idx, i, coding_potential=FALSE;
List *tmpl = NULL;
String *tmpstr;
char *mods_fname = NULL;
List *mod_fname_list;
msa_format_type msa_format = UNKNOWN_FORMAT;
while ((c = getopt_long(argc, argv,
"S:H:V:ni:k:l:C:G:zt:E:R:T:O:r:xL:sN:P:g:U:c:e:IY:D:JM:F:pA:Xqh",
long_opts, &opt_idx)) != -1) {
switch (c) {
case 'S':
p->states = get_arg_list(optarg);
break;
case 'H':
p->hmm = hmm_new_from_file(phast_fopen(optarg, "r"));
p->two_state = FALSE;
break;
case 'V':
p->viterbi_f = phast_fopen(optarg, "w+");
tmpstr = str_new_charstr(optarg);
if (str_ends_with_charstr(tmpstr, ".gff"))
p->gff = TRUE;
str_free(tmpstr);
break;
case 'n':
p->post_probs = FALSE;
break;
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == UNKNOWN_FORMAT)
die("ERROR: bad argument to --msa-format\n");
break;
case 'X':
p->FC = TRUE;
p->two_state = FALSE;
break;
case 'l':
if (optarg[0] != '~')
p->estim_lambda = FALSE;
else optarg = &optarg[1];
p->lambda = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'C':
p->gamma = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'G':
p->gc = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 't':
p->set_transitions = TRUE;
if (optarg[0] != '~')
p->estim_transitions = FALSE;
else optarg = &optarg[1];
tmpl = get_arg_list_dbl(optarg);
if (lst_size(tmpl) != 2)
die("ERROR: bad argument to --transitions.\n");
p->mu = lst_get_dbl(tmpl, 0);
p->nu = lst_get_dbl(tmpl, 1);
if (p->mu <= 0 || p->mu >= 1 || p->nu <= 0 || p->nu >= 1)
die("ERROR: bad argument to --transitions.\n");
lst_free(tmpl);
break;
case 'E':
if (optarg[0] != '~')
p->estim_transitions = FALSE;
else optarg = &optarg[1];
p->omega = get_arg_dbl_bounds(optarg, 1, INFTY);
p->mu = 1/p->omega;
break;
case 'T':
p->estim_trees = TRUE;
p->estim_trees_fname_root = optarg;
break;
case 'O':
p->estim_rho = TRUE;
p->estim_trees_fname_root = optarg;
break;
case 'z':
p->ignore_missing = TRUE;
break;
case 'k':
tmpl = get_arg_list_int(optarg);
if (lst_size(tmpl) > 2)
die("ERROR: too many arguments with --nrates.\n");
p->nrates = lst_get_int(tmpl, 0);
if (p->nrates <= 0)
die("ERROR: bad argument to --nrates (%d).\n", p->nrates);
if (lst_size(tmpl) == 2) {
p->nrates2 = lst_get_int(tmpl, 1);
if (p->nrates2 <= 0)
die("ERROR: bad argument to --nrates (%d).\n", p->nrates2);
}
lst_free(tmpl);
break;
case 'R':
p->rho = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'g':
if (!strcmp(optarg, "-"))
p->log_f = stderr;
else p->log_f = phast_fopen(optarg, "w+");
break;
case 'r':
p->refidx = get_arg_int_bounds(optarg, 0, INFTY);
break;
case 'x':
/* do nothing; left in for backward compatibility */
break;
case 'U':
p->pivot_states = get_arg_list(optarg); /* we want strings not ints
for phmm_new */
break;
case 'e':
p->extrapolate_tree_fname = optarg;
break;
case 'I':
p->indels = TRUE;
break;
case 'Y':
p->max_micro_indel = get_arg_int_bounds(optarg, 1, INFTY);
break;
case 'D':
if (optarg[0] != '~')
p->estim_indels = FALSE;
else optarg = &optarg[1];
tmpl = get_arg_list_dbl(optarg);
if (lst_size(tmpl) != 6) die("ERROR: bad argument to --indel-params.\n");
p->alpha_0 = lst_get_dbl(tmpl, 0);
p->beta_0 = lst_get_dbl(tmpl, 1);
p->tau_0 = lst_get_dbl(tmpl, 2);
p->alpha_1 = lst_get_dbl(tmpl, 3);
p->beta_1 = lst_get_dbl(tmpl, 4);
p->tau_1 = lst_get_dbl(tmpl, 5);
if (p->alpha_0 < 0 || p->beta_0 < 0 || p->tau_0 < 0 ||
p->alpha_1 < 0 || p->beta_1 < 0 || p->tau_1 < 0)
die("ERROR: bad argument to --indel-params.\n");
lst_free(tmpl);
break;
case 'J':
p->indels_only = TRUE;
p->two_state = FALSE;
p->indels = TRUE;
p->post_probs = FALSE;
break;
case 'M':
p->inform_reqd = get_arg_list(optarg);
break;
case 'F':
p->not_informative = get_arg_list(optarg);
break;
case 'c':
p->cm = cm_new_string_or_file(optarg);
break;
case 'L':
p->lnl_f = phast_fopen(optarg, "w+");
break;
case 'N':
p->seqname = optarg;
break;
case 'P':
p->idpref = optarg;
break;
case 's':
p->score = TRUE;
break;
case 'p':
coding_potential = TRUE;
break;
case 'A':
p->alias_hash = make_name_hash(optarg);
break;
case 'q':
p->results_f = NULL;
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("Bad argument. Try '%s -h'.\n", argv[0]);
}
}
if ((!coding_potential && optind != argc - 2) ||
(coding_potential && optind != argc - 2 && optind != argc - 1))
die("ERROR: extra or missing arguments. Try '%s -h'.\n", argv[0]);
set_seed(-1);
if (p->extrapolate_tree_fname != NULL &&
!strcmp(p->extrapolate_tree_fname, "default")) {
p->extrapolate_tree_fname = smalloc((strlen(PHAST_HOME)+100)*sizeof(char));
#if defined(__MINGW32__)
sprintf(p->extrapolate_tree_fname,
"%s\\data\\exoniphy\\mammals\\cftr25_hybrid.nh", PHAST_HOME);
#else
sprintf(p->extrapolate_tree_fname,
"%s/data/exoniphy/mammals/cftr25_hybrid.nh", PHAST_HOME);
#endif
}
if (p->extrapolate_tree_fname != NULL)
p->extrapolate_tree = tr_new_from_file(phast_fopen(p->extrapolate_tree_fname, "r"));
mods_fname = (optind == argc - 2 ? argv[argc - 1] : NULL);
/* if there are two args, mods are the second one; otherwise will
use default mods for coding potential (see below) */
/* set defaults for coding-potential mode */
if (coding_potential) {
char tmp[5000];
p->two_state = FALSE;
if (p->cm == NULL)
p->cm = cm_new_string_or_file("NCATS=4; CNS 1; CDS 2-4");
if (p->hmm == NULL) {
#if defined(__MINGW32__)
sprintf(tmp, "%s\\data\\phastCons\\%s", PHAST_HOME,
p->indels ? "simple-coding-indels.hmm" : "simple-coding.hmm");
#else
sprintf(tmp, "%s/data/phastCons/%s", PHAST_HOME,
p->indels ? "simple-coding-indels.hmm" : "simple-coding.hmm");
#endif
if (p->results_f!=NULL)
fprintf(p->results_f, "Reading HMM from %s...\n", tmp);
p->hmm = hmm_new_from_file(phast_fopen(tmp, "r"));
}
if (mods_fname == NULL) {
#if defined(__MINGW32__)
sprintf(tmp, "%s\\data\\exoniphy\\mammals\\r3.ncns.mod, %s\\data\\exoniphy\\mammals\\r3.cns.mod, %s\\data\\exoniphy\\mammals\\r3.cds-1.mod, %s\\data\\exoniphy\\mammals\\r3.cds-2.mod, %s\\data\\exoniphy\\mammals\\r3.cds-3.mod", PHAST_HOME, PHAST_HOME, PHAST_HOME, PHAST_HOME, PHAST_HOME);
#else
sprintf(tmp, "\
%s/data/exoniphy/mammals/r3.ncns.mod,\
%s/data/exoniphy/mammals/r3.cns.mod,\
%s/data/exoniphy/mammals/r3.cds-1.mod,\
%s/data/exoniphy/mammals/r3.cds-2.mod,\
%s/data/exoniphy/mammals/r3.cds-3.mod",
PHAST_HOME, PHAST_HOME, PHAST_HOME, PHAST_HOME, PHAST_HOME);
#endif
mods_fname = tmp;
}
if (p->states == NULL)
p->states = get_arg_list("CDS");
if (p->pivot_states == NULL)
p->pivot_states = get_arg_list("background,CNS");
}
int main(int argc, char* argv[]) {
char *maf_fname = NULL, *out_root_fname = "maf_parse", *masked_fn = NULL;
String *refseq = NULL, *currRefseq;
int opt_idx, startcol = 1, endcol = -1, include = 1, splitInterval = -1;
char c, outfilename[1000], splitFormat[100]="%s%.1i.maf", *group_tag = NULL;
List *order_list = NULL, *seqlist_str = NULL, *cats_to_do_str=NULL, *cats_to_do=NULL;
MafBlock *block;
FILE *mfile, *outfile=NULL, *masked_file=NULL;
int useRefseq=TRUE, currLen=-1, blockIdx=0, currSize, sortWarned=0;
int lastIdx = 0, currStart=0, by_category = FALSE, i, pretty_print = FALSE;
int lastStart = -1, gffSearchIdx=0;
GFF_Set *gff = NULL, *gffSub;
GFF_Feature *feat;
CategoryMap *cm = NULL;
int base_mask_cutoff = -1, stripILines=FALSE, stripELines=FALSE;//, numspec=0;
List *outfileList=NULL;
Hashtable *outfileHash=NULL;//, *specNameHash=NULL;
msa_format_type output_format = MAF;
MSA *msa = NULL;//, **catMsa;
char *mask_features_spec_arg=NULL;
List *mask_features_spec=NULL;
struct option long_opts[] = {
{"start", 1, 0, 's'},
{"end", 1, 0, 'e'},
{"seqs", 1, 0, 'l'},
{"exclude", 0, 0, 'x'},
{"order", 1, 0, 'O'},
{"split", 1, 0, 'S'},
{"out-root", 1, 0, 'r'},
{"out-root-digits", 1, 0, 'd'},
{"no-refseq", 0, 0, 'n'},
{"features", 1, 0, 'g'},
{"by-category", 0, 0, 'L'},
{"do-cats", 1, 0, 'C'},
{"catmap", 1, 0, 'c'},
{"by-group", 1, 0, 'P'},
{"mask-bases", 1, 0, 'b'},
{"masked-file", 1, 0, 'm'},
{"strip-i-lines", 0, 0, 'I'},
{"strip-e-lines", 0, 0, 'E'},
{"mask-features", 1, 0, 'M'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
while ((c = getopt_long(argc, argv, "s:e:l:O:r:S:d:g:c:P:b:o:m:M:pLnxEIh", long_opts, &opt_idx)) != -1) {
switch(c) {
case 's':
startcol = get_arg_int(optarg);
break;
case 'e':
endcol = get_arg_int(optarg);
break;
case 'l':
seqlist_str = get_arg_list(optarg);
break;
case 'O':
order_list = get_arg_list(optarg);
break;
case 'x':
include = FALSE;
break;
case 'S':
splitInterval = atoi(optarg);
break;
case 'r':
out_root_fname = optarg;
break;
case 'd':
sprintf(splitFormat, "%%s%%.%si.%%s", optarg);
break;
case 'n':
useRefseq = FALSE;
break;
case 'g':
gff = gff_read_set(phast_fopen(optarg, "r"));
gff_sort(gff);
stripILines=TRUE;
stripELines=TRUE;
break;
case 'c':
cm = cm_new_string_or_file(optarg);
break;
case 'C':
cats_to_do_str = get_arg_list(optarg);
break;
case 'L':
by_category = TRUE;
break;
case 'P':
group_tag = optarg;
break;
case 'b':
base_mask_cutoff = atoi(optarg);
break;
case 'm':
masked_fn = optarg;
break;
case 'M':
mask_features_spec_arg = optarg;
break;
case 'E':
stripELines=TRUE;
break;
case 'I':
stripILines=TRUE;
break;
case 'o':
output_format = msa_str_to_format(optarg);
if (output_format == UNKNOWN_FORMAT)
die("ERROR: bad output format. Try \"maf_parse -h\" for help.\n");
if (output_format != MAF)
die("Sorry, only MAF format output has been implemented right now.\n");
break;
case 'p':
pretty_print = TRUE;
break;
case 'h':
print_usage();
exit(0);
case '?':
die("Bad argument. Try 'maf_parse -h' for help.\n");
}
}
if (optind >= argc)
die("Missing alignment filename. Try 'maf_parse -h' for help.\n");
else if (optind == argc - 1)
maf_fname = argv[optind];
else
die("ERROR: Too many arguments. Try 'maf_parse -h' for help.\n");
set_seed(-1);
if (startcol < 1 || (endcol != -1 && endcol < startcol))
die("ERROR: must have 1 <= start <= end <= [msa_length]\n");
if ((group_tag != NULL || by_category) && gff == NULL)
die("ERROR: --by-category and --by-group require --features. Try \"maf_parse -h\""
" for help.\n");
if (group_tag != NULL && by_category)
die("ERROR: --by-category and --by-group cannot be used together. Try \"maf_parse -h\""
" for help.\n");
if (splitInterval != -1 && gff != NULL)
die("ERROR: can't use --split and --features together. Try \"maf_parse -h\""
"for help\n");
if (group_tag != NULL || by_category) {
outfileList = lst_new_ptr(10);
outfileHash = hsh_new(100);
}
if (gff != NULL && cm == NULL)
cm = cm_new_from_features(gff);
if (cats_to_do_str != NULL) {
cats_to_do = cm_get_category_str_list(cm, cats_to_do_str, FALSE);
if (gff != NULL)
gff_filter_by_type(gff, cats_to_do, 0, NULL);
}
if (masked_fn != NULL) {
if (base_mask_cutoff == -1)
die("ERROR: need to use --mask-bases with --masked-file");
masked_file = phast_fopen(masked_fn, "w");
}
if (mask_features_spec_arg != NULL) {
if (gff==NULL)
die("ERROR: need --features with --mask-features");
mask_features_spec = lst_new_ptr(10);
str_split(str_new_charstr(mask_features_spec_arg), ",", mask_features_spec);
for (i=0; i < lst_size(mask_features_spec); i++) {
fprintf(stderr, "masking species %s within features\n",
((String*)lst_get_ptr(mask_features_spec, i))->chars);
}
}
/* Check to see if --do-cats names a feature which is length 1.
If so, set output_format to SS ? or FASTA ? */
mfile = phast_fopen(maf_fname, "r");
block = mafBlock_read_next(mfile, NULL, NULL);
if (splitInterval == -1 && gff==NULL) {
//TODO: do we want to copy header from original MAF in this case?
mafBlock_open_outfile(NULL, argc, argv);
}
while (block != NULL) {
if (order_list != NULL)
mafBlock_reorder(block, order_list);
if (seqlist_str != NULL)
mafBlock_subSpec(block, seqlist_str, include);
if (mafBlock_numSpec(block)==0 || mafBlock_all_gaps(block))
goto get_next_block;
if (stripILines)
mafBlock_strip_iLines(block);
if (stripELines)
mafBlock_strip_eLines(block);
if (base_mask_cutoff != -1)
mafBlock_mask_bases(block, base_mask_cutoff, masked_file);
//TODO: still need to implement (either here or elsewhere)
// if (indel_mask_cutoff != -1)
// mafBlock_mask_indels(block, indel_mask_cutoff, mfile);
if (useRefseq) { //get refseq and check that it is consistent in MAF file
currRefseq = mafBlock_get_refSpec(block);
if (refseq == NULL)
refseq = str_new_charstr(currRefseq->chars);
else if (str_compare(refseq, currRefseq)!=0)
die("Error: refseq not consistent in MAF (got %s, %s)\n",
refseq->chars, currRefseq->chars);
}
if (startcol != 1 || endcol != -1)
if (0 == mafBlock_trim(block, startcol, endcol, refseq, useRefseq ? 0 : lastIdx))
goto get_next_block;
currSize = mafBlock_get_size(block, refseq);
if (useRefseq) {
currStart = mafBlock_get_start(block, refseq);
if (currStart < lastIdx && sortWarned == 0) {
fprintf(stderr, "Warning: input MAF not sorted with respect to refseq. Output files may not represent contiguous alignments. (%i, %i)\n", lastIdx, currStart);
sortWarned = 1;
}
}
else currStart = lastIdx;
if (currStart < lastStart) gffSearchIdx = 0;
lastStart = currStart;
lastIdx = currStart + currSize;
//split by length
if (splitInterval != -1) {
if (currLen == -1 || currLen+currSize > splitInterval) {
sprintf(outfilename, splitFormat, out_root_fname, ++blockIdx,
msa_suffix_for_format(output_format));
if (output_format == MAF) {
if (outfile != NULL) mafBlock_close_outfile(outfile);
outfile = mafBlock_open_outfile(outfilename, argc, argv);
}
else if (output_format != MAF && msa != NULL) {
// msa_print_to_filename(msa, outfilename, output_format, pretty_print);
msa_free(msa);
msa = NULL;
}
currLen = 0;
}
currLen += currSize;
}
else outfile = stdout;
if (gff != NULL && mask_features_spec != NULL) {
gffSub = gff_subset_range_overlap_sorted(gff, currStart+1, lastIdx,
&gffSearchIdx);
if (gffSub != NULL) {
mafBlock_mask_region(block, gffSub, mask_features_spec);
gff_free_set(gffSub);
}
mafBlock_print(outfile, block, pretty_print);
} else if (gff != NULL) {
gffSub = gff_subset_range_overlap_sorted(gff, currStart+1, lastIdx,
&gffSearchIdx);
if (gffSub != NULL) {
if (by_category) gff_group_by_feature(gffSub);
else if (group_tag != NULL) gff_group(gffSub, group_tag);
gff_sort(gffSub);
gff_flatten_within_groups(gffSub);
for (i=0; i<lst_size(gffSub->features); i++) {
feat = (GFF_Feature*)lst_get_ptr(gffSub->features, i);
MafBlock *subBlock = mafBlock_copy(block);
mafBlock_trim(subBlock, feat->start, feat->end, refseq, 0);
if (by_category)
outfile = get_outfile(outfileList, outfileHash, feat->feature, out_root_fname,
argc, argv);
else if (group_tag != NULL)
outfile = get_outfile(outfileList, outfileHash,
gff_group_name(gffSub, feat), out_root_fname,
argc, argv);
else outfile = stdout;
if (output_format == MAF)
mafBlock_print(outfile, subBlock, pretty_print);
// else msa_add_mafBlock(msa);
mafBlock_free(subBlock);
}
gff_free_set(gffSub);
}
}
else {
if (output_format == MAF)
mafBlock_print(outfile, block, pretty_print);
// else msa = msa_add_mafBlock(mafBlock, msa, );
}
get_next_block:
mafBlock_free(block);
block = mafBlock_read_next(mfile, NULL, NULL);
}
if (masked_file != NULL) fclose(masked_file);
if (output_format == MAF) {
if (by_category || group_tag != NULL)
close_outfiles(outfileList, outfileHash);
else if (outfile!=NULL) mafBlock_close_outfile(outfile);
} else {
msa_print(stdout, msa, output_format, pretty_print);
msa_free(msa);
}
if (gff != NULL) gff_free_set(gff);
phast_fclose(mfile);
return 0;
}
int main(int argc, char *argv[]) {
List *pruned_names = lst_new_ptr(5);
TreeModel *source_mod;
MSA *msa = NULL, *out_msa;
IndelHistory *ih;
char *read_hist_fname = NULL;
char c;
int opt_idx, old_nnodes, i;
msa_format_type msa_format = UNKNOWN_FORMAT;
int output_alignment = FALSE, ia_names = FALSE;
struct option long_opts[] = {
{"msa-format", 1, 0, 'i'},
{"output-alignment", 0, 0, 'A'},
{"read-history", 1, 0, 'H'},
{"ia-names", 0, 0, 'I'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
while ((c = getopt_long(argc, argv, "i:H:AIh", long_opts, &opt_idx)) != -1) {
switch (c) {
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == -1)
die("ERROR: unrecognized alignment format.\n");
break;
case 'A':
output_alignment = TRUE;
break;
case 'H':
read_hist_fname = optarg;
break;
case 'I':
ia_names = TRUE;
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("Bad argument. Try 'indelHistory -h'.\n");
}
}
set_seed(-1);
if (read_hist_fname != NULL) {
fprintf(stderr, "Reading indel history from %s...\n", read_hist_fname);
ih = ih_new_from_file(phast_fopen(read_hist_fname, "r"));
}
else {
FILE *mfile;
if (optind != argc - 2)
die("Two arguments required. Try 'indelHistory -h'.\n");
fprintf(stderr, "Reading alignment from %s...\n", argv[optind]);
mfile = phast_fopen(argv[optind], "r");
if (msa_format == UNKNOWN_FORMAT)
msa_format = msa_format_for_content(mfile, 1);
msa = msa_new_from_file_define_format(mfile, msa_format, "ACGTNB^.-");
phast_fclose(mfile);
if (msa->seqs == NULL && (msa->ss == NULL || msa->ss->tuple_idx == NULL))
die("ERROR: ordered representation of alignment required.\n");
fprintf(stderr, "Reading tree from %s...\n", argv[optind+1]);
source_mod = tm_new_from_file(phast_fopen(argv[optind+1], "r"), 1);
/* prune tree, if necessary */
old_nnodes = source_mod->tree->nnodes;
tm_prune(source_mod, msa, pruned_names);
if (lst_size(pruned_names) == (old_nnodes + 1) / 2)
die("ERROR: no match for leaves of tree in alignment (leaf names must match alignment names).\n");
if (lst_size(pruned_names) > 0) {
fprintf(stderr, "WARNING: pruned away leaves of tree with no match in alignment (");
for (i = 0; i < lst_size(pruned_names); i++)
fprintf(stderr, "%s%s", ((String*)lst_get_ptr(pruned_names, i))->chars,
i < lst_size(pruned_names) - 1 ? ", " : ").\n");
}
lst_free(pruned_names);
tr_name_ancestors(source_mod->tree);
if (msa->nseqs > (source_mod->tree->nnodes + 1) / 2) { /* assume ancestral
seqs specified
in this case */
if (ia_names) {
fprintf(stderr, "Converting sequence names...\n");
ih_convert_ia_names(msa, source_mod->tree);
}
fprintf(stderr, "Extracting indel history from alignment...\n");
ih = ih_extract_from_alignment(msa, source_mod->tree);
}
else { /* infer by parsimony */
if (msa->ss == NULL) {
fprintf(stderr, "Extracting sufficient statistics...\n");
ss_from_msas(msa, 1, TRUE, NULL, NULL, NULL, -1, 0);
}
fprintf(stderr, "Inferring indel history by parsimony...\n");
ih = ih_reconstruct(msa, source_mod->tree);
}
}
if (output_alignment) {
out_msa = ih_as_alignment(ih, msa);
msa_print(stdout, out_msa, FASTA, FALSE);
}
else
ih_print(ih, stdout,
read_hist_fname != NULL ? read_hist_fname : argv[optind],
"indelHistory");
fprintf(stderr, "Done.\n");
return 0;
}
int main(int argc, char *argv[]) {
char c;
List *l;
int i, j, strand, bed_output = 0, backgd_nmods = -1, feat_nmods = -1,
winsize = -1, verbose = 0, max_nmods, memblocksize, old_nleaves,
refidx = 1, base_by_base = FALSE, windowWig = FALSE;
TreeModel **backgd_mods = NULL, **feat_mods = NULL;
HMM *backgd_hmm = NULL, *feat_hmm = NULL;
msa_format_type inform = UNKNOWN_FORMAT;
GFF_Set *features = NULL;
MSA *msa, *msa_compl=NULL;
double **backgd_emissions, **feat_emissions, **mem, **dummy_emissions,
*winscore_pos=NULL, *winscore_neg=NULL;
int *no_alignment=NULL;
List *pruned_names;
char *msa_fname;
FILE *infile;
int opt_idx;
struct option long_opts[] = {
{"background-mods", 1, 0, 'b'},
{"background-hmm", 1, 0, 'B'},
{"feature-mods", 1, 0, 'f'},
{"feature-hmm", 1, 0, 'F'},
{"features", 1, 0, 'g'},
{"window", 1, 0, 'w'},
{"window-wig", 1, 0, 'W'},
{"base-by-base", 0, 0, 'y'},
{"msa-format", 1, 0, 'i'},
{"refidx", 1, 0, 'r'},
{"output-bed", 0, 0, 'd'},
{"verbose", 0, 0, 'v'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
while ((c = getopt_long(argc, argv, "B:b:F:f:r:g:w:W:i:ydvh", long_opts, &opt_idx)) != -1) {
switch (c) {
case 'B':
backgd_hmm = hmm_new_from_file(phast_fopen(optarg, "r"));
break;
case 'b':
l = get_arg_list(optarg);
backgd_nmods = lst_size(l);
backgd_mods = smalloc(backgd_nmods * sizeof(void*));
for (i = 0; i < backgd_nmods; i++)
backgd_mods[i] = tm_new_from_file(phast_fopen(((String*)lst_get_ptr(l, i))->chars, "r"), 1);
lst_free_strings(l);
lst_free(l);
break;
case 'F':
feat_hmm = hmm_new_from_file(phast_fopen(optarg, "r"));
break;
case 'f':
l = get_arg_list(optarg);
feat_nmods = lst_size(l);
feat_mods = smalloc(feat_nmods * sizeof(void*));
for (i = 0; i < feat_nmods; i++)
feat_mods[i] = tm_new_from_file(phast_fopen(((String*)lst_get_ptr(l, i))->chars, "r"), 1);
lst_free_strings(l);
lst_free(l);
break;
case 'g':
features = gff_read_set(phast_fopen(optarg, "r"));
break;
case 'w':
winsize = get_arg_int(optarg);
if (winsize <= 0) die("ERROR: window size must be positive.\n");
break;
case 'W':
winsize = get_arg_int(optarg);
if (winsize <= 0) die("ERROR: window size must be positive.\n");
windowWig = TRUE;
break;
case 'y':
base_by_base = TRUE;
break;
case 'i':
inform = msa_str_to_format(optarg);
if (inform == UNKNOWN_FORMAT) die("Bad argument to -i.\n");
break;
case 'r':
refidx = get_arg_int_bounds(optarg, 0, INFTY);
break;
case 'd':
bed_output = 1;
break;
case 'h':
printf("%s", HELP);
exit(0);
case 'v':
verbose = 1;
break;
case '?':
die("Bad argument. Try '%s -h'.\n", argv[0]);
}
}
set_seed(-1);
if (backgd_mods == NULL || feat_mods == NULL)
die("ERROR: -b and -f required. Try '%s -h'.\n", argv[0]);
if (backgd_nmods == 1 && backgd_hmm == NULL)
backgd_hmm = hmm_create_trivial();
else if (backgd_hmm == NULL)
die("ERROR: -B required. Try '%s -h'.\n", argv[0]);
if (feat_nmods == 1 && feat_hmm == NULL)
feat_hmm = hmm_create_trivial();
else if (feat_hmm == NULL)
die("ERROR: -F required. Try '%s -h'.\n", argv[0]);
if ((winsize == -1 && features == NULL && !base_by_base) ||
(winsize != -1 && features != NULL) ||
(winsize != -1 && base_by_base) ||
(features != NULL && base_by_base))
die("ERROR: must specify exactly one of -g, -w, and -y. Try '%s -h'.\n", argv[0]);
if (backgd_hmm->nstates != backgd_nmods)
die("ERROR: number of states must equal number of tree models for background.\n");
if (feat_hmm->nstates != feat_nmods)
die("ERROR: number of states must equal number of tree models for features.\n");
if (features != NULL && lst_size(features->features) == 0)
die("ERROR: empty features file.\n");
if (base_by_base && (backgd_nmods > 1 || feat_nmods > 1))
die("ERROR: only single phylogenetic models (not HMMs) are supported with --base-by-base.\n");
if (optind != argc - 1)
die("ERROR: too few arguments. Try '%s -h'.\n", argv[0]);
if (verbose) fprintf(stderr, "Reading alignment ...\n");
msa_fname = argv[optind];
infile = phast_fopen(msa_fname, "r");
if (inform == UNKNOWN_FORMAT)
inform = msa_format_for_content(infile, 1);
if (inform == MAF)
msa = maf_read(infile, NULL, 1, NULL, NULL,
NULL, -1, TRUE, NULL, NO_STRIP, FALSE);
else
msa = msa_new_from_file_define_format(infile, inform, NULL);
if (msa_alph_has_lowercase(msa)) msa_toupper(msa);
msa_remove_N_from_alph(msa);
/* need ordered representation of alignment */
if (msa->seqs == NULL && (msa->ss == NULL || msa->ss->tuple_idx == NULL) )
die("ERROR: ordered sufficient statistics are required.\n");
pruned_names = lst_new_ptr(msa->nseqs);
for (i = 0; i < backgd_nmods; i++) {
old_nleaves = (backgd_mods[i]->tree->nnodes + 1) / 2;
tm_prune(backgd_mods[i], msa, pruned_names);
if (lst_size(pruned_names) >= old_nleaves)
die("ERROR: no match for leaves of tree in alignment (background model #%d)\n", i+1);
else if (lst_size(pruned_names) > 0) {
fprintf(stderr, "WARNING: pruned away leaves in background model (#%d) with no match in alignment (", i+1);
for (j = 0; j < lst_size(pruned_names); j++)
fprintf(stderr, "%s%s", ((String*)lst_get_ptr(pruned_names, j))->chars,
j < lst_size(pruned_names) - 1 ? ", " : ").\n");
}
lst_free_strings(pruned_names);
}
for (i = 0; i < feat_nmods; i++) {
old_nleaves = (feat_mods[i]->tree->nnodes + 1) / 2;
tm_prune(feat_mods[i], msa, pruned_names);
if (lst_size(pruned_names) >= old_nleaves)
die("ERROR: no match for leaves of tree in alignment (features model #%d)\n", i+1);
else if (lst_size(pruned_names) > 0) {
fprintf(stderr, "WARNING: pruned away leaves in features model (#%d) with no match in alignment (", i+1);
for (j = 0; j < lst_size(pruned_names); j++)
fprintf(stderr, "%s%s", ((String*)lst_get_ptr(pruned_names, j))->chars,
j < lst_size(pruned_names) - 1 ? ", " : ").\n");
}
lst_free_strings(pruned_names);
}
lst_free(pruned_names);
/* first have to subtract offset from features, if necessary */
if (msa->idx_offset != 0 && features != NULL) {
for (i = 0; i < lst_size(features->features); i++) {
GFF_Feature *f = lst_get_ptr(features->features, i);
f->start -= msa->idx_offset;
f->end -= msa->idx_offset;
}
}
/* convert to coord frame of alignment */
if (features != NULL && refidx != 0) {
if (verbose) fprintf(stderr, "Mapping coordinates ...\n");
msa_map_gff_coords(msa, features, refidx, 0, 0);
if (lst_size(features->features) == 0)
die("ERROR: no features within coordinate range of alignment.\n");
}
/* Make a reverse complemented copy of the alignment. The two
strands will be processed separately, to avoid problems with
overlapping features, etc. */
if (!base_by_base) { /* skip in base by base case */
if (verbose) fprintf(stderr, "Creating reverse complemented alignment ...\n");
msa_compl = msa_create_copy(msa, 0);
/* temporary workaround: make sure reverse complement not based on
sufficient stats */
if (msa_compl->seqs == NULL) ss_to_msa(msa_compl);
if (msa_compl->ss != NULL) {
ss_free(msa_compl->ss);
msa_compl->ss = NULL;
}
msa_reverse_compl(msa_compl);
}
/* allocate memory for computing scores */
backgd_emissions = smalloc(backgd_nmods * sizeof(void*));
for (i = 0; i < backgd_nmods; i++)
backgd_emissions[i] = smalloc(msa->length * sizeof(double));
feat_emissions = smalloc(feat_nmods * sizeof(void*));
for (i = 0; i < feat_nmods; i++)
feat_emissions[i] = smalloc(msa->length * sizeof(double));
max_nmods = max(backgd_nmods, feat_nmods);
dummy_emissions = smalloc(max_nmods * sizeof(void*));
mem = smalloc(max_nmods * sizeof(void*));
/* memory for forward algorithm -- each block must be as large as
the largest feature */
if (features != NULL) {
for (i = 0, memblocksize = -1; i < lst_size(features->features); i++) {
GFF_Feature *f = lst_get_ptr(features->features, i);
if (f->end - f->start + 1 > memblocksize)
memblocksize = f->end - f->start + 1;
}
}
else memblocksize = winsize; /* -1 if base-by-base mode */
if (memblocksize > 0)
for (i = 0; i < max_nmods; i++)
mem[i] = smalloc(memblocksize * sizeof(double));
if (winsize != -1) {
winscore_pos = smalloc(msa->length * sizeof(double));
winscore_neg = smalloc(msa->length * sizeof(double));
no_alignment = smalloc(msa->length * sizeof(int));
for (i = 0; i < msa->length; i++) {
winscore_pos[i] = winscore_neg[i] = NEGINFTY;
if (refidx == 0)
no_alignment[i] = FALSE;
else
no_alignment[i] = msa_missing_col(msa, refidx, i);
}
}
/* the rest will be repeated for each strand */
for (strand = 1; strand <= 2; strand++) {
MSA *thismsa = strand == 1 ? msa : msa_compl;
double *winscore = strand == 1 ? winscore_pos : winscore_neg;
if (base_by_base && strand == 2) break; /* don't do second pass in
base_by_base case */
if (verbose) fprintf(stderr, "Processing %c strand ...\n",
strand == 1 ? '+' : '-');
/* set up dummy categories array, so that emissions are only
computed where needed */
thismsa->categories = smalloc(thismsa->length * sizeof(int));
thismsa->ncats = 1;
if (winsize != -1) {
if (strand == 1)
for (i = 0; i < thismsa->length; i++)
thismsa->categories[i] = no_alignment[i] ? 0 : 1;
else
for (i = 0; i < thismsa->length; i++)
thismsa->categories[i] = no_alignment[thismsa->length - i - 1] ? 0 : 1;
}
else if (features != NULL) {
for (i = 0; i < thismsa->length; i++) thismsa->categories[i] = 0;
for (i = 0; i < lst_size(features->features); i++) {
GFF_Feature *f = lst_get_ptr(features->features, i);
if (f->start <= 0 || f->end <= 0) {
fprintf(stderr, "WARNING: feature out of range ('");
gff_print_feat(stderr, f);
fprintf(stderr, "')\n");
continue;
}
if (strand == 1 && f->strand != '-')
for (j = f->start - 1; j < f->end; j++)
thismsa->categories[j] = 1;
else if (strand == 2 && f->strand == '-')
for (j = thismsa->length - f->end;
j < thismsa->length - f->start + 1; j++)
thismsa->categories[j] = 1;
}
}
else { /* base-by-base scores */
for (i = 0; i < thismsa->length; i++) thismsa->categories[i] = 1;
}
if (thismsa->ss != NULL) ss_update_categories(thismsa);
/* compute emissions */
for (i = 0; i < backgd_nmods; i++) {
if (verbose)
fprintf(stderr, "Computing emissions for background model #%d ...\n", i+1);
tl_compute_log_likelihood(backgd_mods[i], thismsa,
backgd_emissions[i], NULL, 1, NULL);
}
for (i = 0; i < feat_nmods; i++) {
if (verbose)
fprintf(stderr, "Computing emissions for features model #%d ...\n", i+1);
tl_compute_log_likelihood(feat_mods[i], thismsa,
feat_emissions[i], NULL, 1, NULL);
}
/* now compute scores */
if (winsize != -1) { /* windows case */
int winstart;
if (verbose) fprintf(stderr, "Computing scores ...\n");
for (winstart = 0; winstart <= thismsa->length - winsize; winstart++) {
int centeridx = winstart + winsize/2;
if (strand == 2) centeridx = thismsa->length - centeridx - 1;
if (no_alignment[centeridx]) continue;
for (j = 0; j < feat_nmods; j++)
dummy_emissions[j] = &(feat_emissions[j][winstart]);
winscore[centeridx] = hmm_forward(feat_hmm, dummy_emissions,
winsize, mem);
if (winscore[centeridx] <= NEGINFTY) {
winscore[centeridx] = NEGINFTY;
continue;
}
for (j = 0; j < backgd_nmods; j++)
dummy_emissions[j] = &(backgd_emissions[j][winstart]);
winscore[centeridx] -= hmm_forward(backgd_hmm, dummy_emissions,
winsize, mem);
if (winscore[centeridx] < NEGINFTY) winscore[centeridx] = NEGINFTY;
}
}
else if (features != NULL) { /* features case */
if (verbose) fprintf(stderr, "Computing scores ...\n");
for (i = 0; i < lst_size(features->features); i++) {
GFF_Feature *f = lst_get_ptr(features->features, i);
int s, e;
if ((strand == 1 && f->strand == '-') ||
(strand == 2 && f->strand != '-') ||
f->start <= 0 || f->end <= 0 || f->end - f->start < 0)
continue;
/* effective coords */
if (f->strand == '-') {
s = thismsa->length - f->end + 1;
e = thismsa->length - f->start + 1;
}
else {
s = f->start;
e = f->end;
}
f->score_is_null = 0;
for (j = 0; j < feat_nmods; j++)
dummy_emissions[j] = &(feat_emissions[j][s-1]);
f->score = hmm_forward(feat_hmm, dummy_emissions, e - s + 1, mem);
if (f->score <= NEGINFTY) {
f->score = NEGINFTY;
continue;
}
for (j = 0; j < backgd_nmods; j++)
dummy_emissions[j] = &(backgd_emissions[j][s-1]);
f->score -= hmm_forward(backgd_hmm, dummy_emissions, e - s + 1, mem);
if (f->score < NEGINFTY) f->score = NEGINFTY;
}
}
}
if (verbose) fprintf(stderr, "Generating output ...\n");
if (winsize != -1 && windowWig == FALSE) { /* standard windows output */
for (i = 0, j = 0; i < msa->length; i++) {
if (no_alignment[i] == FALSE)
printf("%d\t%.3f\t%.3f\n", j + msa->idx_offset + 1, winscore_pos[i],
winscore_neg[i]);
if (ss_get_char_pos(msa, i, 0, 0) != GAP_CHAR) j++;
}
}
else if (windowWig == TRUE) { /* windows with wig output */
int last = NEGINFTY;
for (i = 0, j = 0; i < msa->length; i++) {
if (refidx == 0 || msa_get_char(msa, refidx-1, i) != GAP_CHAR) {
if (no_alignment[i] == FALSE && winscore_pos[i] > NEGINFTY) {
if (j > last + 1)
printf("fixedStep chrom=%s start=%d step=1\n",
refidx > 0 ? msa->names[refidx-1] : "alignment",
j + msa->idx_offset + 1);
printf("%.3f\n", winscore_pos[i]);
last = j;
}
j++;
}
}
}
else if (features != NULL) { /* features output */
/* return to coord frame of reference seq (also, replace offset) */
if (refidx != 0)
msa_map_gff_coords(msa, features, 0, refidx, msa->idx_offset);
else if (msa->idx_offset != 0) {
for (i = 0; i < lst_size(features->features); i++) {
GFF_Feature *f = lst_get_ptr(features->features, i);
f->start += msa->idx_offset;
f->end += msa->idx_offset;
}
}
if (bed_output)
gff_print_bed(stdout, features, FALSE);
else
gff_print_set(stdout, features);
}
else { /* base-by-base scores */
/* in this case, we can just output the difference between the emissions */
printf("fixedStep chrom=%s start=%d step=1\n",
refidx > 0 ? msa->names[refidx-1] : "alignment",
msa->idx_offset + 1);
for (i = 0, j = 0; i < msa->length; i++) {
if (refidx == 0 || msa_get_char(msa, refidx-1, i) != GAP_CHAR) {
printf("%.3f\n", feat_emissions[0][i] - backgd_emissions[0][i]);
j++;
}
}
}
if (verbose) fprintf(stderr, "\nDone.\n");
return 0;
}
int main(int argc, char *argv[]) {
char c;
char *msa_fname = NULL;
int opt_idx, i, old_nnodes;
MSA *msa;
List *pruned_names = lst_new_ptr(5), *tmpl;
BDPhyloHmm *bdphmm;
GFF_Set *predictions;
int found = FALSE;
List *ignore_types = lst_new_ptr(1);
struct option long_opts[] = {
{"refseq", 1, 0, 'M'},
{"msa-format", 1, 0, 'i'},
{"refidx", 1, 0, 'r'},
{"rho", 1, 0, 'R'},
{"phi", 1, 0, 'p'},
{"transitions", 1, 0, 't'},
{"expected-length", 1, 0, 'E'},
{"target-coverage", 1, 0, 'C'},
{"seqname", 1, 0, 'N'},
{"idpref", 1, 0, 'P'},
{"indel-model", 1, 0, 'I'},
{"indel-history", 1, 0, 'H'},
{"help", 0, 0, 'h'},
{0, 0, 0, 0}
};
/* arguments and defaults for options */
FILE *refseq_f = NULL, *msa_f = NULL;
msa_format_type msa_format = UNKNOWN_FORMAT;
TreeModel *source_mod;
double rho = DEFAULT_RHO, mu = DEFAULT_MU, nu = DEFAULT_NU,
phi = DEFAULT_PHI, gamma = -1, omega = -1,
alpha_c = -1, beta_c = -1, tau_c = -1,
alpha_n = -1, beta_n = -1, tau_n = -1;
int set_transitions = FALSE, refidx = 1, estim_phi = TRUE,
estim_gamma = TRUE, estim_omega = TRUE;
char *seqname = NULL, *idpref = NULL;
IndelHistory *ih = NULL;
while ((c = getopt_long(argc, argv, "R:t:p:E:C:r:M:i:N:P:I:H:h", long_opts, &opt_idx)) != -1) {
switch (c) {
case 'R':
rho = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 't':
if (optarg[0] != '~') estim_gamma = estim_omega = FALSE;
else optarg = &optarg[1];
set_transitions = TRUE;
tmpl = get_arg_list_dbl(optarg);
if (lst_size(tmpl) != 2)
die("ERROR: bad argument to --transitions.\n");
mu = lst_get_dbl(tmpl, 0);
nu = lst_get_dbl(tmpl, 1);
if (mu <= 0 || mu >= 1 || nu <= 0 || nu >= 1)
die("ERROR: bad argument to --transitions.\n");
lst_free(tmpl);
break;
case 'p':
if (optarg[0] != '~') estim_phi = FALSE;
else optarg = &optarg[1];
phi = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'E':
if (optarg[0] != '~') estim_omega = FALSE;
else optarg = &optarg[1];
omega = get_arg_dbl_bounds(optarg, 1, INFTY);
mu = 1/omega;
break;
case 'C':
if (optarg[0] != '~') estim_gamma = FALSE;
else optarg = &optarg[1];
gamma = get_arg_dbl_bounds(optarg, 0, 1);
break;
case 'r':
refidx = get_arg_int_bounds(optarg, 0, INFTY);
break;
case 'M':
refseq_f = phast_fopen(optarg, "r");
break;
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == UNKNOWN_FORMAT)
die("ERROR: unrecognized alignment format.\n");
break;
case 'N':
seqname = optarg;
break;
case 'P':
idpref = optarg;
break;
case 'I':
tmpl = get_arg_list_dbl(optarg);
if (lst_size(tmpl) != 3 && lst_size(tmpl) != 6)
die("ERROR: bad argument to --indel-model.\n");
alpha_n = lst_get_dbl(tmpl, 0);
beta_n = lst_get_dbl(tmpl, 1);
tau_n = lst_get_dbl(tmpl, 2);
if (lst_size(tmpl) == 6) {
alpha_c = lst_get_dbl(tmpl, 3);
beta_c = lst_get_dbl(tmpl, 4);
tau_c = lst_get_dbl(tmpl, 5);
}
else {
alpha_c = alpha_n; beta_c = beta_n; tau_c = tau_n;
}
if (alpha_c <= 0 || alpha_c >= 1 || beta_c <= 0 || beta_c >= 1 ||
tau_c <= 0 || tau_c >= 1 || alpha_n <= 0 || alpha_n >= 1 ||
beta_n <= 0 || beta_n >= 1 || tau_n <= 0 || tau_n >= 1)
die("ERROR: bad argument to --indel-model.\n");
break;
case 'H':
fprintf(stderr, "Reading indel history from %s...\n", optarg);
ih = ih_new_from_file(phast_fopen(optarg, "r"));
break;
case 'h':
printf("%s", HELP);
exit(0);
case '?':
die("Bad argument. Try 'dless -h'.\n");
}
}
if (optind != argc - 1)
die("Missing alignment file or model file. Try 'dless -h'.\n");
if (set_transitions && (gamma != -1 || omega != -1))
die("ERROR: --transitions and --target-coverage/--expected-length cannot be used together.\n");
if ((gamma != -1 && omega == -1) || (gamma == -1 && omega != -1))
die("ERROR: --target-coverage and --expecteed-length must be used together.\n");
set_seed(-1);
if (gamma != -1)
nu = gamma/(1-gamma) * mu;
fprintf(stderr, "Reading tree model from %s...\n", argv[optind]);
source_mod = tm_new_from_file(phast_fopen(argv[optind], "r"), 1);
if (source_mod->nratecats > 1)
die("ERROR: rate variation not currently supported.\n");
if (source_mod->order > 0)
die("ERROR: only single nucleotide models are currently supported.\n");
if (!tm_is_reversible(source_mod))
phast_warning("WARNING: p-value computation assumes reversibility and your model is non-reversible.\n");
/* read alignment */
msa_f = phast_fopen(argv[optind], "r");
fprintf(stderr, "Reading alignment from %s...\n", argv[optind]);
if (msa_format == UNKNOWN_FORMAT)
msa_format = msa_format_for_content(msa_f, 1);
if (msa_format == MAF) {
msa = maf_read(msa_f, refseq_f, 1, NULL, NULL, NULL, -1, TRUE, NULL,
NO_STRIP, FALSE);
}
else
msa = msa_new_from_file_define_format(msa_f, msa_format, NULL);
if (msa_alph_has_lowercase(msa)) msa_toupper(msa);
msa_remove_N_from_alph(msa);
if (msa->ss == NULL) {
fprintf(stderr, "Extracting sufficient statistics...\n");
ss_from_msas(msa, 1, TRUE, NULL, NULL, NULL, -1, 0);
}
else if (msa->ss->tuple_idx == NULL)
die("ERROR: ordered representation of alignment required unless --suff-stats.\n");
/* prune tree, if necessary */
old_nnodes = source_mod->tree->nnodes;
tm_prune(source_mod, msa, pruned_names);
if (lst_size(pruned_names) == (old_nnodes + 1) / 2)
die("ERROR: no match for leaves of tree in alignment (leaf names must match alignment names).\n");
if (lst_size(pruned_names) > 0) {
fprintf(stderr, "WARNING: pruned away leaves of tree with no match in alignment (");
for (i = 0; i < lst_size(pruned_names); i++)
fprintf(stderr, "%s%s", ((String*)lst_get_ptr(pruned_names, i))->chars,
i < lst_size(pruned_names) - 1 ? ", " : ").\n");
}
/* this has to be done after pruning tree */
tr_name_ancestors(source_mod->tree);
/* also make sure match for reference sequence in tree */
if (refidx > 0) {
for (i = 0, found = FALSE; !found && i < source_mod->tree->nnodes; i++) {
TreeNode *n = lst_get_ptr(source_mod->tree->nodes, i);
if (!strcmp(n->name, msa->names[refidx-1]))
found = TRUE;
}
if (!found) die("ERROR: no match for reference sequence in tree.\n");
}
/* checks for indel model */
if (alpha_c > 0) {
if (ih == NULL) {
fprintf(stderr, "Reconstructing indel history by parsimony...\n");
ih = ih_reconstruct(msa, source_mod->tree);
}
else {
if (ih->ncols != msa->length)
die("ERROR: indel history doesn't seem to match alignment.\n");
if (ih->tree->nnodes != source_mod->tree->nnodes)
die("ERROR: indel history doesn't seem to match tree model.\n");
}
}
bdphmm = bd_new(source_mod, rho, mu, nu, phi, alpha_c, beta_c, tau_c,
alpha_n, beta_n, tau_n, estim_gamma, estim_omega,
estim_phi);
/* compute emissions */
phmm_compute_emissions(bdphmm->phmm, msa, FALSE);
/* add emissions for indel model, if necessary */
if (alpha_c > 0) {
fprintf(stderr, "Adjusting emissions for indels...\n");
bd_add_indel_emissions(bdphmm, ih);
}
/* postprocess for missing data (requires special handling) */
fprintf(stderr, "Adjusting emissions for missing data...\n");
bd_handle_missing_data(bdphmm, msa);
if (estim_gamma || estim_omega || estim_phi) {
fprintf(stderr, "Estimating free parameters...\n");
bd_estimate_transitions(bdphmm, msa);
}
/* set seqname and idpref, if necessary */
if (seqname == NULL || idpref == NULL) {
/* derive default from file name root */
String *tmp = str_new_charstr(msa_fname);
if (!str_equals_charstr(tmp, "-")) {
str_remove_path(tmp);
str_root(tmp, '.');
if (idpref == NULL) idpref = copy_charstr(tmp->chars);
str_root(tmp, '.'); /* apply one more time for double suffix */
if (seqname == NULL) seqname = tmp->chars;
}
else if (seqname == NULL) seqname = "refseq";
}
/* obtain predictions */
fprintf(stderr, "Running Viterbi algorithm...\n");
predictions = phmm_predict_viterbi(bdphmm->phmm, seqname, NULL, idpref, NULL);
lst_push_ptr(ignore_types, str_new_charstr("nonconserved"));
gff_filter_by_type(predictions, ignore_types, TRUE, NULL);
/* score predictions */
fprintf(stderr, "Scoring predictions...\n");
bd_score_predictions(bdphmm, predictions);
/* can free emissions now */
for (i = 0; i < bdphmm->phmm->hmm->nstates; i++)
sfree(bdphmm->phmm->emissions[i]);
sfree(bdphmm->phmm->emissions);
bdphmm->phmm->emissions = NULL;
/* convert GFF to coord frame of reference sequence and adjust
coords by idx_offset, if necessary */
if (refidx != 0 || msa->idx_offset != 0)
msa_map_gff_coords(msa, predictions, 0, refidx, msa->idx_offset);
if (refidx != 0)
gff_flatten(predictions);
/* necessary because coord conversion might create overlapping
features (can happen in deletions in reference sequence) */
/* now output predictions */
fprintf(stderr, "Writing GFF to stdout...\n");
gff_print_set(stdout, predictions);
fprintf(stderr, "Done.\n");
return 0;
}
int main(int argc, char *argv[]) {
TreeNode *tree = NULL;
TreeModel *backgd_mod = NULL;
int i, j,
size = DEFAULT_SIZE, meme_mode = 0, profile_mode = 0,
nrestarts = 10, npseudocounts = 5, nsamples = -1,
nmostprevalent = -1, tuple_size = -1, nbest = -1, sample_parms = 0,
nmotifs = DEFAULT_NUMBER, nseqs = -1, do_html = 0, do_bed = 0,
suppress_stdout = 0;
List *msa_name_list = NULL, *pos_examples = NULL, *init_list = NULL, *tmpl;
List *msas, *motifs;
SeqSet *seqset = NULL;
PooledMSA *pmsa = NULL;
msa_format_type msa_format = UNKNOWN_FORMAT;
Vector *backgd_mnmod = NULL;
Hashtable *hash=NULL;
String *output_prefix = str_new_charstr("phastm.");
double *has_motif = NULL;
double prior = PRIOR;
char c;
GFF_Set *bedfeats = NULL;
while ((c = getopt(argc, argv, "t:i:b:sk:md:pn:I:R:P:w:c:SB:o:HDxh")) != -1) {
switch (c) {
case 't':
tree = tr_new_from_file(phast_fopen(optarg, "r"));
break;
case 'i':
msa_format = msa_str_to_format(optarg);
if (msa_format == UNKNOWN_FORMAT)
die("ERROR: bad input format.\n");
break;
case 'b':
backgd_mod = tm_new_from_file(phast_fopen(optarg, "r"), 1);
break;
case 's':
break;
case 'k':
size = get_arg_int(optarg);
break;
case 'm':
meme_mode = 1;
break;
case 'd':
pos_examples = get_arg_list(optarg);
break;
case 'p':
profile_mode = 1;
break;
case 'n':
nrestarts = get_arg_int(optarg);
break;
case 'I':
init_list = get_arg_list(optarg);
break;
case 'P':
tmpl = str_list_as_int(get_arg_list(optarg));
if (lst_size(tmpl) != 2) die("ERROR: bad argument to -P.\n");
nmostprevalent = lst_get_int(tmpl, 0);
tuple_size = lst_get_int(tmpl, 1);
if (!(nmostprevalent > 0 && tuple_size > 0))
die("ERROR: bad argument nmostprevalent=%i tuple_size=%i\n",
nmostprevalent, tuple_size);
lst_free(tmpl);
break;
case 'R':
tmpl = str_list_as_int(get_arg_list(optarg));
if (lst_size(tmpl) != 2) die("ERROR: bad argument to -R.\n");
nsamples = lst_get_int(tmpl, 0);
tuple_size = lst_get_int(tmpl, 1);
if (!(nsamples > 0 && tuple_size > 0))
die("ERROR nsamples=%i tuple_sizse=%i\n", nsamples, tuple_size);
lst_free(tmpl);
break;
case 'c':
npseudocounts = get_arg_int(optarg);
break;
case 'w':
nbest = get_arg_int(optarg);
break;
case 'S':
sample_parms = 1;
break;
case 'B':
nmotifs = get_arg_int(optarg);
break;
case 'o':
str_free(output_prefix);
output_prefix = str_new_charstr(optarg);
str_append_char(output_prefix, '.');
break;
case 'H':
do_html = 1;
break;
case 'D':
do_bed = 1;
break;
case 'x':
suppress_stdout = 1;
break;
case 'h':
usage(argv[0]);
case '?':
die("Bad argument. Try '%s -h'.\n", argv[0]);
}
}
if (optind != argc - 1)
die("ERROR: List of alignment files required. Try '%s -h'.\n", argv[0]);
if ((nsamples > 0 && nmostprevalent > 0) ||
(nsamples > 0 && init_list != NULL) ||
(nmostprevalent > 0 && init_list != NULL))
die("ERROR: -I, -P, and -R are mutually exclusive.");
set_seed(-1);
msa_name_list = get_arg_list(argv[optind]);
if (backgd_mod != NULL && tree == NULL) tree = backgd_mod->tree;
if (tree == NULL && !meme_mode && !profile_mode)
die("ERROR: Must specify -t, -m, or -p.\n");
if ((init_list != NULL || nsamples > 0 || nmostprevalent > 0) &&
!sample_parms)
nrestarts = 1;
if (pos_examples != NULL) {
hash = hsh_new(lst_size(pos_examples));
for (i = 0; i < lst_size(pos_examples); i++)
hsh_put_int(hash, ((String*)lst_get_ptr(pos_examples, i))->chars, 1);
has_motif = smalloc(lst_size(msa_name_list) * sizeof(double));
}
/* open all MSAs */
msas = lst_new_ptr(lst_size(msa_name_list));
fprintf(stderr, "Reading alignment(s) ...\n");
for (i = 0, j = 0; i < lst_size(msa_name_list); i++) {
String *name = lst_get_ptr(msa_name_list, i);
FILE *mfile = phast_fopen(name->chars, "r");
msa_format_type temp_format;
MSA *msa;
if (msa_format == UNKNOWN_FORMAT)
temp_format = msa_format_for_content(mfile, 1);
else temp_format = msa_format;
msa = msa_new_from_file_define_format(mfile, temp_format, NULL);
phast_fclose(mfile);
if (nseqs == -1) nseqs = msa->nseqs;
if (!meme_mode &&
(msa->length - msa_num_gapped_cols(msa, STRIP_ANY_GAPS, -1, -1) < 300 ||
msa->nseqs != nseqs)) {
fprintf(stderr, "WARNING: ignoring alignment '%s' -- too few informative sites.\n", name->chars);
msa_free(msa);
continue;
}
if (msa_alph_has_lowercase(msa)) msa_toupper(msa);
msa_remove_N_from_alph(msa); /* Ns can be a problem */
lst_push_ptr(msas, msa);
if (has_motif != NULL) {
int k, hm = (hsh_get_int(hash, name->chars) == 1);
if (meme_mode) { /* here need to record at individ seq level */
has_motif = srealloc(has_motif,
(j + msa->nseqs + 1) * sizeof(double)); /* FIXME */
for (k = 0; k < msa->nseqs; k++) has_motif[j++] = hm;
}
else has_motif[j++] = hm;
}
}
if (!meme_mode) {
fprintf(stderr, "Extracting and pooling sufficient statistics ...\n");
pmsa = ss_pooled_from_msas(msas, 1, size, NULL, 0);
msa_remove_N_from_alph(pmsa->pooled_msa);
}
/* obtain individual sequences, if necessary */
if (nmostprevalent > 0 || nsamples > 0 || meme_mode) {
if (meme_mode) fprintf(stderr, "Converting to individual sequences ...\n");
else fprintf(stderr, "Obtaining reference sequences for pre-processing ...\n");
seqset = mtf_get_seqset(msas, meme_mode ? -1 : 1, 10 * size);
/* for now, assume 1st seq is reference */
msa_remove_N_from_alph(seqset->set);
}
if (nmostprevalent > 0) {
fprintf(stderr, "Obtaining %d most prevalent %d-tuples ...\n",
nmostprevalent, tuple_size);
init_list = lst_new_ptr(nmostprevalent);
mtf_get_common_ntuples(seqset, init_list, tuple_size, nmostprevalent);
}
else if (nsamples > 0) {
fprintf(stderr, "Sampling %d %d-tuples ...\n", nsamples, tuple_size);
init_list = lst_new_ptr(nsamples);
mtf_sample_ntuples(seqset, init_list, tuple_size, nsamples);
}
/* in meme_mode, backgd model can be specified as eq freqs in a .mod file */
if (meme_mode && backgd_mod != NULL && has_motif == NULL)
backgd_mnmod = backgd_mod->backgd_freqs;
/* estimate background model, if necessary */
else if (backgd_mod == NULL && (!meme_mode || has_motif == NULL)) {
fprintf(stderr, "Fitting background model%s ...\n",
has_motif == NULL ? "" : " (for use in initialization)");
/* if discriminative, be clear
backgd isn't really part of the
estimation procedure */
if (meme_mode) {
backgd_mnmod = vec_new(strlen(seqset->set->alphabet));
mtf_estim_backgd_mn(seqset, backgd_mnmod);
}
else {
backgd_mod = tm_new(tr_create_copy(tree), NULL, NULL, F81,
pmsa->pooled_msa->alphabet, 1, 0, NULL, -1);
tm_fit(backgd_mod, pmsa->pooled_msa,
tm_params_init(backgd_mod, .1, 5, 0),
-1, OPT_MED_PREC, NULL, 0, NULL);
}
}
/* select subset of init strings, if necessary */
if (nbest > 0 && init_list != NULL) {
fprintf(stderr, "Winnowing candidate start strings ...\n");
tmpl = lst_new_ptr(nbest);
mtf_winnow_starts(meme_mode ? (void*)seqset : (void*)pmsa,
init_list, nbest, tmpl, !meme_mode, size, tree,
meme_mode ? (void*)backgd_mnmod : (void*)backgd_mod,
has_motif);
lst_free(init_list);
init_list = tmpl;
}
/* Now find motifs */
motifs = mtf_find(meme_mode ? (void*)seqset : (void*)pmsa,
!meme_mode, size, nmotifs, tree,
meme_mode ? (void*)backgd_mnmod : (void*)backgd_mod,
has_motif, prior, nrestarts, init_list, sample_parms,
npseudocounts);
fprintf(stderr, "\n\n");
if (do_bed)
bedfeats = gff_new_set_init("phast_motif", "0.1b");
/* generate output */
for (i = 0; i < lst_size(motifs); i++) {
Motif *m = lst_get_ptr(motifs, i);
if (!suppress_stdout) {
if (lst_size(motifs) > 1)
printf("\n**********\nMOTIF #%d\n**********\n\n", i+1);
mtf_print(stdout, m);
}
if (do_html) {
String *fname = str_dup(output_prefix);
str_append_int(fname, i+1);
str_append_charstr(fname, ".html");
mtf_print_html(phast_fopen(fname->chars, "w+"), m);
str_free(fname);
}
if (do_bed)
mtf_add_features(m, bedfeats);
}
if (do_html) {
String *fname = str_dup(output_prefix);
str_append_charstr(fname, "index.html");
mtf_print_summary_html(phast_fopen(fname->chars, "w+"),
motifs, output_prefix);
str_free(fname);
}
if (do_bed) {
String *fname = str_dup(output_prefix);
str_append_charstr(fname, "bed");
gff_print_bed(phast_fopen(fname->chars, "w+"),
bedfeats, FALSE);
str_free(fname);
}
return 0;
}
