def __init__(self, seed, **kwargs):
self.seed = seed
defaults = {
# PhyloFit
"subst_model": "REV",
"use_em": True,
# PhyloP
"method": "SPH",
"mode": "CONACC",
"aln_alphabet": ["A", "T", "C", "G", "N"],
}
self.phast = Phast()
self.__dict__.update(defaults)
self._routine_done = False
class mmEvolution(object):
"""
Compute Evolutionary features
Args:
aln_fname (str): Alignment filename.
aln (str): Alignment it-self.
aln_format (str): Alignment format. Currently supported is FASTA.
aln_alphabet (list): List of nucleotides to consider in the aligned
sequences (others get filtered).
subst_model (str): PhyloFit substitution model (REV...).
tree (str): Tree in the Newick format.
fitting_tree (bool): Fitting or not the tree on the alignment.
use_em (bool): Fitting or not the tree with Expectation-Maximization algorithm.
motif_def (str): 'seed' or 'seed_extended' or 'site'.
motif_upstream_extension (int): Upstream extension length.
motif_downstream_extension (int): Downstream extension length.
"""
def __init__(self, seed, **kwargs):
self.seed = seed
defaults = {
# PhyloFit
"subst_model": "REV",
"use_em": True,
# PhyloP
"method": "SPH",
"mode": "CONACC",
"aln_alphabet": ["A", "T", "C", "G", "N"],
}
self.phast = Phast()
self.__dict__.update(defaults)
self._routine_done = False
def _eval_routine(self, setup, worker, **kwargs):
# Parameters
if "aln_fname" not in kwargs and "aln" not in kwargs:
raise IOError("An alignment is required")
if "tree" not in kwargs and "fitting_tree" not in kwargs:
raise IOError("A tree is required")
if "aln_format" not in kwargs and "aln_fname" in kwargs:
aln_format = kwargs.get("aln_fname").split(".")[-1].upper()
if aln_format == "FA":
aln_format = "FASTA"
if aln_format is None:
raise ValueError("Alignment format undetected")
# Load alignment
if "aln_fname" in kwargs:
seqs = utils.load_fasta(kwargs["aln_fname"])
else:
seqs = utils.load_fasta(kwargs["aln"], as_string=True)
seqs_cleaned = {}
seqs_coords = {}
for seq_name, seq in seqs.items():
seqs_cleaned[seq_name] = utils.clean_seq(seq, self.aln_alphabet)
seqs_coords[seq_name] = get_coord_vec(seq, self.aln_alphabet)
kwargs.update({"seqs": seqs})
args = setup(**kwargs)
args.update({"aln_format": aln_format})
# Reset
out = []
# Compute
for its in range(len(self.seed.end_sites)):
end_site = self.seed.end_sites[its]
# Motif
# start_motif and end_motif are sequence coordinates => 1-based
start_motif, end_motif = seed.get_motif_coordinates(
end_site,
self.motif_def,
self.seed.pairings[its],
self.motif_upstream_extension,
self.motif_downstream_extension,
self.seed.min_target_length,
)
motif = self.seed.target_seq[start_motif - 1 : end_motif].replace("U", "T")
# Species with seed(s)
species_with_seed = []
for seq_name, seq in seqs_cleaned.items():
if seq.find(motif) != -1:
species_with_seed.append(seq_name)
args.update(
{
"species_with_seed": species_with_seed,
"start_motif": start_motif,
"end_motif": end_motif,
"seqs_coords": seqs_coords,
}
)
out.append(worker(**args))
return out
def _eval_cons_bls(self, **kwargs):
def setup(**kwargs):
return {
"subst_model": kwargs.get("subst_model", self.subst_model),
"fitting_tree": kwargs.get("fitting_tree", True),
"use_em": kwargs.get("use_em", self.use_em),
"fitting_tree_done": False,
}
def worker(species_with_seed, fitting_tree_done, fitting_tree, subst_model, aln_format, use_em):
if len(species_with_seed) > 1:
# Fitting tree if necessary
if fitting_tree_done is False:
if fitting_tree:
if kwargs.get("aln_fname"):
fitted_tree = self.phast.phylofit(
subst_model=subst_model,
aln_fname=kwargs["aln_fname"],
aln_format=aln_format,
tree=kwargs["tree"],
use_em=use_em,
)["tree"]
elif kwargs("aln"):
fitted_tree = self.phast.phylofit(
subst_model=subst_model,
aln=kwargs["aln"],
aln_format=aln_format,
tree=kwargs["tree"],
use_em=use_em,
)["tree"]
else:
fitted_tree = kwargs["tree"]
fitting_tree_done = True
# Compute BLS
dtree = dendropy.Tree.get_from_string(fitted_tree, schema="newick", preserve_underscores=True)
dtree.retain_taxa_with_labels(species_with_seed)
return sum([edge.length for edge in dtree.postorder_edge_iter()][:-1])
else:
return 0.0
self.cons_blss = self._eval_routine(setup, worker, **kwargs)
return self.cons_blss
def _eval_selec_phylop(self, **kwargs):
def setup(**kwargs):
return {
"method": kwargs.get("method", self.method),
"mode": kwargs.get("mode", self.mode),
"motif_upstream_extension": kwargs.get("motif_upstream_extension", 0),
"motif_downstream_extension": kwargs.get("motif_downstream_extension", 0),
"ref_species": kwargs["seq"].keys()[0],
"mod_fname": kwargs.get("mod_fname", None),
}
def worker(species_with_seed, ref_species, start_motif, end_motif, seqs_coords, seqs, method, mode, mod_fname):
pval = 1.0
if len(species_with_seed) > 1:
# Extract alignment
start_motif_in_aln = seqs_coords[ref_species][start_motif - 1]
end_motif_in_aln = seqs_coords[ref_species][end_motif - 1]
partial_seqs = collections.OrderedDict()
for seq_name, seq in seqs.items():
if seq_name in species_with_seed:
a = start_motif_in_aln - 1
b = end_motif_in_aln
partial_seqs[seq_name] = seq[a:b]
partial_seqs = remove_gap_column(partial_seqs)
aln = "\n".join(["> %s\n%s" % (k, v) for k, v in partial_seqs.items()])
pval = self.phast.phylop(method=method, mode=mode, mod_fname=mod_fname, aln=aln, aln_format="FASTA")
return pval
self.selec_phylops = self._eval_routine(setup, worker, **kwargs)
return self.selec_phylops
def routine(self, **kwargs):
try:
self._eval_cons_bls(**kwargs)
self._eval_selec_phylop(**kwargs)
except IOError:
self.cons_blss = [0 for _ in range(len(self.seed.end_sites))]
self.selec_phylops = [0 for _ in range(len(self.seed.end_sites))]
self._routine_done = True
@property
def cons_bls(self):
try:
return min(self.cons_blss)
except AttributeError:
return min(self._eval_cons_bls())
@property
def selec_phylop(self):
try:
return min(self.selec_phylops)
except AttributeError:
return min(self._eval_selec_phylop())
