def compute_true_negative_hmm_scores(self):
count = 1
print("\n")
for group, file_list in self.group_to_file_list.items():
for file_name in file_list:
hmm = self.group_by_file_to_hmm[group][file_name]
fasta_hash = self.make_cluster_specific_TN_hash(
group, file_name)
if fasta_hash and len(fasta_hash) >= 10:
scoring_obj = ScoreObject(fasta_hash, hmm)
score_hash = scoring_obj.bulk_score_computation()
self.group_by_file_to_twin_score_obj[group][
file_name] = scoring_obj
if len(fasta_hash) >= 20:
self.group_by_file_to_twin_hmm[group][
file_name] = scoring_obj.compute_full_phmm(
os.path.join(output_dir,
file_name + ".TN_hmm"))
print_progress(
count,
self.valid_input_scope,
prefix='\tTrue Negative Score Distributions:\t',
suffix='Complete',
bar_length=30)
if keep:
keep_file = write_hash_to_fasta(
os.path.join(
keep_dir,
"{}_{}_TrueNegativeScores.txt".format(
group, file_name)), score_hash, "{}\t{}\n")
else:
self.group_by_file_to_twin_score_obj[group][
file_name] = None
count += 1
return self.group_by_file_to_twin_score_obj
def compute_all_hmm_scores(self):
count = 1
print("\n")
for group, file_list in self.group_to_file_list.items():
for file_name in file_list:
fasta_hash = self.group_by_file_to_cluster_hash[group][
file_name]
self.group_by_file_to_score_obj[group][
file_name] = ScoreObject(
fasta_hash,
self.group_by_file_to_hmm[group][file_name])
if len(fasta_hash) < 20:
score_hash = self.group_by_file_to_score_obj[group][
file_name].iterative_score_computation()
else:
score_hash = self.group_by_file_to_score_obj[group][
file_name].bulk_score_computation()
print_progress(count,
self.valid_input_scope,
prefix='\tComputing HMM Score Distributions:\t',
suffix='Complete',
bar_length=30)
count += 1
if keep:
keep_file = write_hash_to_fasta(
os.path.join(
keep_dir,
"{}_{}_scores.txt".format(group, file_name)),
score_hash, "{}\t{}\n")
return self.group_by_file_to_score_obj
def write_length_binned_fasta(fasta_dict, cluster_name, location):
"""takes a dictionary in in style of {'fasta_header':sequence} and outputs
a fasta file with header in style of {'>cluster_name_bin': sequence} at the specified 'location'."""
length_hash = {header: len(seq) for header, seq in fasta_dict.items()}
if not length_hash:
raise AttributeError("! No length hash")
if subset:
bin_to_header = bin_sequence_lengths(length_hash)
header_to_sequence = {
header: fasta_dict[header]
for bin_x, header in bin_to_header.items()
}
file_path = write_hash_to_fasta(location, header_to_sequence,
"{}_" + cluster_name + "\n" + "{}\n")
else:
file_path = write_hash_to_fasta(location, fasta_dict,
"{}_" + cluster_name + "\n" + "{}\n")
return length_hash
def generate_hmm_and_filtered_fasta(self, directory):
count = 1
removed_group_to_file_list = defaultdict(list)
for group, file_list in self.group_to_file_list.items():
for file_name in file_list:
msa_list = generate_msa(
self.group_by_file_to_filepath[group][file_name])
msa_obj = MsaObject(msa_list, file_name, directory)
msa_obj.msa_to_fasta()
msa_obj.trim_remove()
if msa_obj.check_msa_size_and_length() is True:
if msa_obj.size_history[0] != msa_obj.size_history[-1]:
self.group_by_file_to_cluster_hash[group][
file_name] = clean_fasta_hash(
self.group_by_file_to_cluster_hash[group]
[file_name], msa_obj.all_header(), file_name)
same_msa_path = write_hash_to_fasta(
msa_obj.file_path,
self.group_by_file_to_cluster_hash[group]
[file_name], ">{}\n{}\n")
msa_obj.re_align(same_msa_path)
self.group_by_file_to_msa_obj[group][file_name] = msa_obj
length_hash = write_length_binned_fasta(
self.group_by_file_to_cluster_hash[group][file_name],
file_name,
os.path.join(output_dir, file_name + ".fasta"))
self.group_by_file_to_length_range[group][
file_name] = calculate_length_range(length_hash)
self.group_by_file_to_hmm[group][file_name] = generate_hmm(
os.path.join(output_dir, file_name + ".hmm"),
msa_obj.file_path)
else:
removed_group_to_file_list[group].append(file_name)
logger_Filtered.warning(
"Filtered due to MSA benchmarking {}".format(
file_name))
print_progress(count,
self.valid_input_scope,
prefix='\tGenerating Hidden Markov Models:\t',
suffix='Complete',
bar_length=30)
count += 1
return self.remove_filtered_files(removed_group_to_file_list)
def bulk_score_computation(self):
with tmp.NamedTemporaryFile() as q_tmp:
write_hash_to_fasta(q_tmp.name, self.fasta_hash)
self.score_dict = get_phmm_score(self.hmm_path, q_tmp.name)
return self.score_dict
def test_formatting_None_if_more_then_two_brackets(self):
with tmp.NamedTemporaryFile() as hash_tmp:
file_path = shared_code_box.write_hash_to_fasta(hash_tmp.name, {">1" : "ATGSAD", ">2": "ADFAT"}, line_style="{}\n{}\n{}\n")
self.assertEqual(file_path, None)
def test_formatting_without_dictionary_value(self):
with tmp.NamedTemporaryFile() as hash_tmp:
file_path = shared_code_box.write_hash_to_fasta(hash_tmp.name, {">1" : "ATGSAD", ">2": "ADFAT"}, line_style="{}")
self.assertSetEqual(set(open(file_path).readlines()[0].split(">")), {'', '2', '1'})
def test_rigth_optional_formatting(self):
with tmp.NamedTemporaryFile() as hash_tmp:
file_path = shared_code_box.write_hash_to_fasta(hash_tmp.name, {">1" : "ATGSAD", ">2": "ADFAT"}, line_style="{}\t{}\t")
self.assertSetEqual(set(open(file_path).readlines()[0].split("\t")), {'>2', 'ADFAT', '>1', 'ATGSAD', ''})
def test_empty_hash(self):
with tmp.NamedTemporaryFile() as hash_tmp:
file_path = shared_code_box.write_hash_to_fasta(hash_tmp.name, {})
self.assertEqual(file_path, None)
def test_fasta_dictionary(self):
with tmp.NamedTemporaryFile() as hash_tmp:
file_path = shared_code_box.write_hash_to_fasta(hash_tmp.name, {">1" : "ATGSAD", ">2": "ADFAT"})
self.assertSetEqual(set(open(file_path).readlines()), {'>2\n', 'ADFAT\n', '>1\n', 'ATGSAD\n'})