def test_codon_optimize_as_hard_constraint():
numpy.random.seed(123)
problem = DnaOptimizationProblem(
sequence=random_dna_sequence(2000, seed=123),
constraints=[
EnforceTranslation(location=Location(1000, 1300)),
CodonOptimize(location=Location(1000, 1300), species='e_coli')
]
)
assert not problem.all_constraints_pass()
problem.resolve_constraints()
assert problem.all_constraints_pass()
def test_EnforceTranslation_bacterial_valine_antisense():
table_name = "Bacterial"
protein = "LLTMMVTTTTVMVL"
protein_sequence = reverse_translate(protein, table=table_name)
for first_codon_before, first_codon_after in [
("ATG", "ATG"), # methionine stays the only methionine codon
("GTG", "GTG"), # valine-start-codon stays the only valine-start-codon
]:
sequence = first_codon_before + protein_sequence
cds_constraint = EnforceTranslation(
genetic_table="Bacterial",
start_codon="keep",
location=Location(0, len(sequence), -1),
)
problem = DnaOptimizationProblem(
sequence=reverse_complement(sequence),
constraints=[cds_constraint],
objectives=[EnforceChanges()],
logger=None,
)
assert problem.constraints[0].translation == "MLLTMMVTTTTVMVL"
problem.optimize()
problem_sequence_rv = reverse_complement(problem.sequence)
protein_after = translate(
problem_sequence_rv, table_name, assume_start_codon=True
)
assert protein_after == "M" + protein
assert problem_sequence_rv[:3] == first_codon_after
def test_enforce_pattern_basics():
numpy.random.seed(123)
for seed in [2, 3, 123456]:
# The seeds cover various cases:
# 2: the problem has no occurences instead of 1 wanted
# 3: the pattern has no occurences instead of 1 wanted
# 123456: the pattern is over-represented (4 times instead of 1)
sequence = random_dna_sequence(5000, seed=seed)
constraints = [
EnforceTranslation(location=Location(1000, 2500)),
EnforceTranslation(location=Location(3000, 4500)),
EnforcePatternOccurence("ANANANANTT",
location=Location(1100, 2150)),
]
problem = DnaOptimizationProblem(sequence=sequence,
constraints=constraints,
logger=None)
assert not problem.all_constraints_pass()
problem.resolve_constraints()
assert problem.all_constraints_pass()
def evaluate(self, problem):
"""Return Gen9's ninemer score for the problem' sequence"""
sequence = problem.sequence
all_9mers = [sequence[i:i + 9] for i in range(len(sequence) - 9)]
number_of_non_unique_9mers = sum([
count for ninemer, count in Counter(all_9mers).items() if count > 1
])
score = -(9.0 * number_of_non_unique_9mers) / len(sequence)
return SpecEvaluation(self,
problem,
score=score,
locations=[Location(0, len(sequence))],
message="Score: %.02f (%d non-unique ninemers)" %
(score, number_of_non_unique_9mers))
def __init__(
self,
left_overhang,
right_overhang,
left_addition="",
right_addition="",
enzyme="BsmBI",
extra_avoided_sites=(),
description="Golden Gate domesticator",
name="unnamed_domesticator",
cds_by_default=False,
constraints=(),
objectives=(),
):
self.enzyme = enzyme
self.left_overhang = left_overhang
left_overhang = sequence_to_biopython_record(left_overhang)
self.right_overhang = right_overhang
right_overhang = sequence_to_biopython_record(right_overhang)
for seq in [left_overhang, right_overhang]:
annotate_record(seq, label=str(seq.seq))
enzyme_seq = Restriction.__dict__[enzyme].site
enzyme_seq = sequence_to_biopython_record(enzyme_seq)
annotate_record(enzyme_seq, label=enzyme)
self.enzyme_seq = enzyme_seq
left_flank = self.enzyme_seq + "A" + left_overhang + left_addition
right_flank = (right_addition + right_overhang +
(self.enzyme_seq + "A").reverse_complement())
self.extra_avoided_sites = extra_avoided_sites
constraints = list(constraints) + [(lambda seq: AvoidPattern(
EnzymeSitePattern(enzyme),
location=Location(len(left_flank),
len(left_flank) + len(seq)),
)) for enz in ([enzyme] + list(extra_avoided_sites))]
PartDomesticator.__init__(
self,
left_flank=left_flank,
right_flank=right_flank,
constraints=constraints,
objectives=objectives,
description=description,
name=name,
cds_by_default=cds_by_default,
)
def test_EnforceSequence():
# Two enzymes, BsmBI(CGTCTC) is GC-rich, EcoRI(GAATTC) is GC-poor, which
# enzyme will be chosen and inserted in the sequence depends on the other
# constraint on GC content
numpy.random.seed(1234)
for symbol, nucleotides in [("W", "AT"), ("S", "GC")]:
n_nucleotides = 15
start = 50
location = (start, start + n_nucleotides)
problem = DnaOptimizationProblem(
sequence=25 * "ATGC",
constraints=[
AvoidPattern("ATGC"),
AvoidPattern("AAA"),
AvoidPattern("GGG"),
EnforceSequence(n_nucleotides * symbol, location=location),
],
)
problem.max_random_iters = 10000
problem.resolve_constraints()
s, e = start, start + n_nucleotides
assert all([n in nucleotides for n in problem.sequence[s:e]])
# Test -1 strand:
seq = "ATG" + "CAG" + "AGCAAGGTGCTGCT"
problem = DnaOptimizationProblem(
sequence=seq,
constraints=[
EnforcePatternOccurence(
pattern="CTG", # CAG on strand +1
occurences=2,
strand=-1,
location=Location(start=0, end=50),
)
],
)
assert not problem.all_constraints_pass()
problem.resolve_constraints()
assert problem.all_constraints_pass()
print("DOWNLOADING AND PARSING THE GENBANK DATA...")
url = (
"https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?"
+ "db=nucleotide&id=48994873&rettype=gb&retmode=txt"
)
genbank_data = request.urlopen(url).read().decode("utf-8")
genbank_record = load_record(StringIO(genbank_data), file_format="genbank")
print("INITIALIZING THE PROBLEM WITH CONSTRAINTS FOR EACH GENE...")
constraints = []
for feature in genbank_record.features:
if feature.type == "gene" and len(feature.location.parts) == 1:
location = Location.from_biopython_location(feature.location)
if (len(location) % 3 == 0) and len(location) > 100:
gene_constraints = [
EnforceTranslation(location = location),
AvoidPattern("BsmBI_site", location),
EnforceGCContent(
mini=0.40, maxi=0.60, window=150, location=location
),
]
constraints.extend(gene_constraints)
problem = DnaOptimizationProblem(genbank_record, constraints)
print("RESOLVING THE CONSTRAINTS...")
problem.logger.ignore_bars_under = 50
problem.resolve_constraints()
from dnachisel import (EnforceTranslation, DnaOptimizationProblem,
random_dna_sequence, Location, EnforcePattern)
# sequence = random_dna_sequence(5000, seed=123456)
# sequence = random_dna_sequence(5000, seed=2)
# sequence = random_dna_sequence(5000, seed=3)
# for seed in [2, 3, 123456]:
sequence = random_dna_sequence(5000, seed=123)
constraints = [
EnforceTranslation(Location(1000, 2500)),
EnforceTranslation(Location(3000, 4500)),
EnforcePattern("ANANANANTT", location=Location(1100, 2150)),
EnforcePattern("ATGATGCCTK", location=Location(2700, 2800))
]
problem = DnaOptimizationProblem(sequence=sequence, constraints=constraints)
print(problem.constraints_text_summary())
assert not problem.all_constraints_pass()
problem.resolve_constraints()
assert problem.all_constraints_pass()
print(problem.constraints_text_summary())
"""Example of use of the AvoidPAttern specification"""
from dnachisel import (DnaOptimizationProblem, random_dna_sequence,
CodonOptimize, Location, EnforceTranslation)
problem = DnaOptimizationProblem(
sequence=random_dna_sequence(2000, seed=123),
constraints=[
EnforceTranslation(location=Location(1000, 1300)),
CodonOptimize(location=Location(1000, 1300), species='e_coli')
])
print("\nBefore resolution:\n")
print(problem.constraints_text_summary())
problem.resolve_constraints()
print("\nAfter resolution:\n")
print(problem.constraints_text_summary())
base, ext = os.path.splitext(os.path.basename(args.vector)) output_filename = base + "_" + destination + ext naive_construct, objectives, constraints = load_template(args.vector, placeholder, destination) else: output_filename = destination + ".gb" objectives = [] constraints = [] naive_construct = placeholder whole_seq_feat = SeqFeature() whole_seq_feat.type = "misc_feature" whole_seq_feat.qualifiers['label'] = [destination] whole_seq_feat.location = FeatureLocation(0,len(placeholder),strand=1) naive_construct.features.append(whole_seq_feat) dest_feat = find_annotation(naive_construct, placeholder.name) dest_loc = Location.from_biopython_location(dest_feat.location) user_objectives, user_constraints = load_user_options(args, dest_loc) objectives += user_objectives constraints += user_constraints problem = DnaOptimizationProblem(str(naive_construct.seq), constraints=constraints, objectives=objectives) domesticator_record = problem.to_record() mature_construct = naive_construct mature_construct.features.extend(domesticator_record.features)
def domesticate(
self,
dna_sequence=None,
protein_sequence=None,
is_cds="default",
codon_optimization=None,
extra_constraints=(),
extra_objectives=(),
final_record_target=None,
edit=False,
barcode="",
barcode_spacer="AA",
report_target=None,
):
"""Domesticate a sequence.
Parameters
----------
dna_sequence
The DNA sequence string to domesticate.
protein_sequence
Amino-acid sequence of the protein, which will be converted into
a DNA sequence string.
is_cds
If True, sequence edits are restricted to synonymous mutations.
codon_optimization
Either None for no codon optimization or the name of an organism
supported by DnaChisel.
extra_constraints
List of extra constraints to apply to the domesticated sequences.
Each constraint is either a DnaChisel constraint or a function
(dna_sequence => DnaChisel constraint).
extra_objectives
List of extra optimization objectives to apply to the domesticated
sequences. Each objective is either a DnaChisel constraint or a
function (dna_sequence => DnaChisel constraint).
final_record_target
Path to the file where to write the final genbank.
edit
Turn to True to allow sequence edits (if it is false and no all
constraints are originally satisfied, a failed domestication result
(i.e. with attribute ``success`` set to False) will be returned.
report_target
Target for the sequence optimization report (a folder path, or a zip
path).
barcode
A sequence of DNA that will be added to the left of the sequence once
the domestication is done.
barcode_spacer
Nucleotides to be added between the barcode and the enzyme (optional,
the idea here is that they will make sure to avoid the creation of
unwanted cutting sites).
Returns
-------
final_record, edits_record, report_data, success, msg
"""
if is_cds == "default":
is_cds = self.cds_by_default
if isinstance(dna_sequence, SeqRecord):
problem = DnaOptimizationProblem.from_record(dna_sequence)
for spec in problem.constraints + problem.objectives:
spec.location += len(self.left_flank)
extra_constraints = list(extra_constraints) + problem.constraints
extra_objectives = list(extra_constraints) + problem.objectives
if protein_sequence is not None:
is_cds = True
dna_sequence = reverse_translate(protein_sequence)
constraints = [
c(dna_sequence) if hasattr(c, "__call__") else c
for c in list(extra_constraints) + self.constraints
]
location = Location(len(self.left_flank),
len(self.left_flank) + len(dna_sequence))
if is_cds:
constraints.append(EnforceTranslation(location=location))
objectives = [
o(dna_sequence) if hasattr(o, "__call__") else o
for o in list(extra_objectives) + self.objectives
]
if codon_optimization:
objectives.append(
CodonOptimize(species=codon_optimization, location=location))
if self.minimize_edits:
objectives.append(AvoidChanges())
extended_sequence = self.left_flank + dna_sequence + self.right_flank
if (not is_cds) and (not edit):
constraints.append(AvoidChanges())
problem = DnaOptimizationProblem(
extended_sequence,
constraints=constraints,
objectives=objectives,
logger=self.logger,
)
all_constraints_pass = problem.all_constraints_pass()
no_objectives = (len(problem.objectives) - self.minimize_edits) == 0
report_data = None
optimization_successful = True
message = ""
# print (all_constraints_pass, no_objectives)
if not (all_constraints_pass and no_objectives):
problem.n_mutations = self.simultaneous_mutations
if report_target is not None:
(success, message, report_data) = problem.optimize_with_report(
target=report_target, project_name=self.name)
optimization_successful = success
else:
report_data = None
try:
problem.resolve_constraints()
problem.optimize()
except Exception as err:
message = str(err)
optimization_successful = False
report_data = None
final_record = problem.to_record(
with_original_features=True,
with_original_spec_features=False,
with_constraints=False,
with_objectives=False,
)
edits_record = problem.to_record(
with_original_features=True,
with_original_spec_features=False,
with_constraints=False,
with_objectives=False,
with_sequence_edits=True,
)
if final_record_target is not None:
SeqIO.write(final_record, final_record_target, "genbank")
return DomesticationResult(
problem.sequence_before,
final_record,
edits_record,
report_data,
optimization_successful,
message,
)
def test_enforce_pattern_options():
# Checks for Github issue #53
# Test 6 cases: location yes/no, 3 strand options
sequence = "A" * 10
pattern = "C" * 4
# location=None
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(pattern,
occurences=1,
strand="from_location"),
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern in problem.sequence
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(pattern, occurences=1, strand="both")
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern in problem.sequence
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(pattern, occurences=1, strand=-1)
],
logger=None,
)
assert problem.constraints[0].evaluate(problem).score == -1
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert dc.reverse_complement(
pattern) in problem.sequence # other strand used
# location specificed
# Use -1 strand from location:
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(
pattern,
occurences=1,
strand="from_location",
location=Location(1, 6, strand=-1),
)
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert dc.reverse_complement(pattern) in problem.sequence
# Overwrite -1 strand to "both":
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(
pattern,
occurences=1,
strand="both",
location=Location(1, 6, strand=-1),
)
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern in problem.sequence # uses +1 strand by default
# Overwrite -1 strand to +1:
problem = dc.DnaOptimizationProblem(
sequence=sequence,
constraints=[
dc.EnforcePatternOccurence(
pattern,
occurences=1,
strand=1,
location=Location(1, 6, strand=-1),
)
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern in problem.sequence # uses +1 strand
def test_avoid_pattern_options():
# Checks Github issue #53
pattern = "C" * 4
sequence = "A" * 6 + pattern
# location=None
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[AvoidPattern(pattern, strand="from_location")],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern not in problem.sequence
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[AvoidPattern(pattern, strand="both")],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern not in problem.sequence
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[AvoidPattern(pattern, strand=-1)],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
assert pattern in problem.sequence
# location specified
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[
AvoidPattern(pattern,
location=Location(0, 10, -1),
strand="from_location")
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
# sequence not changed because location strand is -1:
assert pattern in problem.sequence
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[
AvoidPattern(pattern, location=Location(0, 10, -1), strand="both")
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
# sequence changed because strand option overwrites location:
assert pattern not in problem.sequence
problem = DnaOptimizationProblem(
sequence=sequence,
constraints=[
AvoidPattern(pattern, location=Location(0, 10, 1), strand=-1)
],
logger=None,
)
problem.resolve_constraints()
assert problem.all_constraints_pass()
# sequence not changed because strand option overwrites location strand:
assert pattern in problem.sequence
def test_location_strand_gets_conserved():
cst = AvoidPattern("AACAAAT", Location(4, 1624, -1))
location = Location(9, 10)
new_cst = cst.localized(location)
assert new_cst.location.to_tuple() == (4, 16, -1)