def get_reaction(pi_structure=pi_structure, ec='1.1.1.1'):
return data_model.Reaction(
participants=[
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=atp_structure)),
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=h2o_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=adp_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=pi_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=h_structure)),
],
cross_references=[
data_model.Resource(namespace='ec-code', id=ec)
])
def test_normalize(self):
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='c'),
coefficient=-1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='c'),
coefficient=-1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='c'),
coefficient=1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='c'),
coefficient=0),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='e'),
coefficient=-1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.adp, id='adp'),
compartment=data_model.Compartment(id='c'),
coefficient=1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.adp, id='adp'),
compartment=data_model.Compartment(id='c'),
coefficient=3),
])
ordered_participants = rxn.get_ordered_participants()
self.assertEqual(len(ordered_participants), 4)
part = list(
filter(
lambda p: p.specie.id == 'atp' and p.compartment.id == 'c' and
p.coefficient < 0, ordered_participants))
self.assertEqual(len(part), 1)
self.assertEqual(part[0].coefficient, -2)
part = list(
filter(
lambda p: p.specie.id == 'atp' and p.compartment.id == 'c' and
p.coefficient > 0, ordered_participants))
self.assertEqual(len(part), 1)
self.assertEqual(part[0].coefficient, 1)
part = list(
filter(lambda p: p.specie.id == 'atp' and p.compartment.id == 'e',
ordered_participants))
self.assertEqual(len(part), 1)
self.assertEqual(part[0].coefficient, -1)
part = list(
filter(lambda p: p.specie.id == 'adp', ordered_participants))
self.assertEqual(len(part), 1)
self.assertEqual(part[0].coefficient, 4)
def setUpClass(cls):
cls.cache_dirname = tempfile.mkdtemp()
cls.flk = common_schema.CommonSchema(cache_dirname=cls.cache_dirname)
cls.q = reaction_kinetics.ReactionKineticsQuery(
cache_dirname=cls.cache_dirname, include_variants=True)
cls.reaction = data_model.Reaction(
participants = [
data_model.ReactionParticipant(
specie = data_model.Specie(
id = 'Dihydrofolate',
structure = 'InChI=1S/C19H21N7O6/c20-19-25-15-14(17(30)26-19)23-11(8-22-15)'+\
'7-21-10-3-1-9(2-4-10)16(29)24-12(18(31)32)5-6-13(27)28/h1-4,12,21H,5-8H2,'+\
'(H,24,29)(H,27,28)(H,31,32)(H4,20,22,25,26,30)'),
coefficient = -1),
data_model.ReactionParticipant(
specie = data_model.Specie(
id = 'NADPH',
structure = 'InChI=1S/C21H30N7O17P3/c22-17-12-19(25-7-24-17)28(8-26-12)21-16'+\
'(44-46(33,34)35)14(30)11(43-21)6-41-48(38,39)45-47(36,37)40-5-10-13(29)15(31)'+\
'20(42-10)27-3-1-2-9(4-27)18(23)32/h1,3-4,7-8,10-11,13-16,20-21,29-31H,2,5-6H2,'+\
'(H2,23,32)(H,36,37)(H,38,39)(H2,22,24,25)(H2,33,34,35)'),
coefficient = -1),
data_model.ReactionParticipant(
specie = data_model.Specie(
id = 'H+',
structure = 'InChI=1S/H'),
coefficient = -1),
data_model.ReactionParticipant(
specie = data_model.Specie(
id = 'NADP+',
structure = 'InChI=1S/C21H28N7O17P3/c22-17-12-19(25-7-24-17)28(8-26-12)21-16('+\
'44-46(33,34)35)14(30)11(43-21)6-41-48(38,39)45-47(36,37)40-5-10-13(29)15(31)20'+\
'(42-10)27-3-1-2-9(4-27)18(23)32/h1-4,7-8,10-11,13-16,20-21,29-31H,5-6H2,(H7-,22'+\
',23,24,25,32,33,34,35,36,37,38,39)/p+1'),
coefficient = 1),
data_model.ReactionParticipant(
specie = data_model.Specie(
id = '5,6,7,8-Tetrahydrofolate',
structure = 'InChI=1S/C19H23N7O6/c20-19-25-15-14(17(30)26-19)23-11(8-22-15)7-21-'+\
'10-3-1-9(2-4-10)16(29)24-12(18(31)32)5-6-13(27)28/h1-4,11-12,21,23H,5-8H2,(H,24,29'+\
')(H,27,28)(H,31,32)(H4,20,22,25,26,30)'),
coefficient = 1)
])
def make_reaction(self):
return data_model.Reaction(participants=[
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.atp, id='atp'),
compartment=data_model.Compartment(id='c'),
coefficient=-1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.h2o, id='h2o'),
compartment=data_model.Compartment(id='c'),
coefficient=-1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.adp, id='adp'),
compartment=data_model.Compartment(id='c'),
coefficient=1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.pi, id='pi'),
compartment=data_model.Compartment(id='c'),
coefficient=1),
data_model.ReactionParticipant(
specie=data_model.Specie(structure=self.h, id='h'),
compartment=data_model.Compartment(id='c'),
coefficient=1),
])
def parse_participants(side, coefficient, reaction, errors):
for participant in side.split(' + '):
participant = participant.strip()
pubchem_compounds = pubchempy.get_compounds(participant, 'name')
if len(pubchem_compounds) == 1:
structure = pubchem_compounds[0].inchi
elif molecule_util.Molecule(structure=participant).get_format():
structure = participant
else:
structure = ''
errors.append(participant)
reaction.participants.append(data_model.ReactionParticipant(
specie=data_model.Specie(structure=structure),
coefficient=coefficient,
))
def _port(self, reaction_list):
"""
Converts SQL model reaction into a Obj Model based data_model reaction
Args:
reaction_list :obj:`list` of :obj:`models.Reaction`:): list of reaction participants
Returns:
:obj:`data_model.Reaction`: a cohesive reaction data object
"""
references = []
participants = []
for rxn_part in reaction_list:
if rxn_part._is_reactant:
coef = -1
elif rxn_part._is_product:
coef = 1
elif rxn_part._is_modifier:
coef = 0
part = data_model.ReactionParticipant(specie=data_model.Specie(
name=rxn_part.metabolite.metabolite_name,
structure=rxn_part.metabolite.structure._value_inchi
if rxn_part.metabolite.structure else None),
coefficient=coef)
participants.append(part)
if len(references) < 1:
for item in rxn_part.kinetic_law._metadata.resource:
references.append(
data_model.Resource(namespace=item.namespace,
id=item._id,
assignment_method=data_model.
ResourceAssignmentMethod.manual))
rxn = data_model.Reaction(
participants=participants,
cross_references=references,
kinetic_law_id=reaction_list[0].kinetic_law_id)
rxn.name = rxn.stringify()
return rxn
def read_species(self, ws):
""" Read species from an Excel worksheet
Args:
ws (:obj:`openpyxl.Worksheet`): worksheet
Returns:
:obj:`list` of :obj:`data_model.Specie`: list of species
"""
species = []
for i in range(2, ws.max_row + 1):
species.append(
data_model.Specie(
id=ws.cell(row=i, column=1).value,
structure=ws.cell(row=i, column=2).value,
))
return species
def test_SpecieStructuralSimilarityFilter(self):
adp = 'NC1=C2N=CN(C3OC(COP([O-])(=O)OP([O-])([O-])=O)C(O)C3O)C2=NC=N1'
atp = 'NC1=C2N=CN(C3OC(COP([O-])(=O)OP([O-])(=O)OP([O-])([O-])=O)C(O)C3O)C2=NC=N1'
h2o = 'O'
# min_similarity = 0
f = data_query.SpecieStructuralSimilarityFilter(min_similarity=0.)
ov = data_model.ObservedValue(observable=data_model.Observable(specie=data_model.Specie(structure=adp)))
numpy.testing.assert_almost_equal(f.score(data_model.Specie(structure=atp), ov), 0.955, decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(specie=data_model.Specie(structure=h2o)))
numpy.testing.assert_almost_equal(f.score(data_model.Specie(structure=atp), ov), 0, decimal=3)
# min_similarity = 0.75
f = data_query.SpecieStructuralSimilarityFilter(min_similarity=0.75)
ov = data_model.ObservedValue(observable=data_model.Observable(specie=data_model.Specie(structure=adp)))
numpy.testing.assert_almost_equal(f.score(data_model.Specie(structure=atp), ov), 0.955, decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(specie=data_model.Specie(structure=h2o)))
numpy.testing.assert_almost_equal(f.score(data_model.Specie(structure=atp), ov), -1., decimal=3)
def get_observed_result(self, metabolite):
""" Find observed concentrations for the metabolite or similar metabolites
Args:
metabolite (:obj:`models.Metabolite`): metabolite to find data for
Returns:
:obj:`list` of :obj:`data_model.ObservedValue`: list of relevant observations
"""
concentrations = self.get_concentration_by_structure(
metabolite.structure._value_inchi,
only_formula_and_connectivity=False).all()
observed_values = []
references = [
data_model.Resource(namespace=item.namespace, id=item._id)
for item in metabolite._metadata.resource
]
for c in concentrations:
metadata = self.metadata_dump(c)
observable = data_model.Observable(
specie=data_model.Specie(
name=metabolite.metabolite_name,
cross_references=references,
structure=metabolite.structure._value_inchi),
compartment=data_model.Compartment(
name=c._metadata.cell_compartment[0].name))
observed_values.append(
data_model.ObservedValue(metadata=metadata,
observable=observable,
value=c.value,
error=c.error,
units=c.units))
return observed_values
def test_to_inchi(self):
s = data_model.Specie(structure=self.pi)
self.assertEqual(s.to_inchi(), self.pi)
self.assertEqual(s.to_inchi(only_formula_and_connectivity=True),
'H3O4P/c1-5(2,3)4')
def test_ObservedResultMetadata(self):
o = data_model.ObservedResultMetadata()
o.genetics = data_model.Genetics(taxon='Mycoplasma pneumoniae',
variation='ΔMPN001')
o.environment = data_model.Environment(temperature=37,
ph=7.,
media='Hayflick')
o.reference = data_model.Reference(title='title',
author='author',
year=2017,
volume=1,
number=1,
pages='1-10')
o.method = data_model.ExperimentalMethod(
name='assay', description='description of assay')
observable = data_model.Observable(
interaction=data_model.Reaction(id='AtpSynthase'),
specie=data_model.Specie(id='ATP'),
compartment=data_model.Compartment(id='c'),
property='K_m',
)
ov = data_model.ObservedValue(
observable=observable,
value=1.0,
error=0.5,
units='U/mg',
)
o.observed_result.append(ov)
o.validate()
self.assertEqual(o.observed_result, [ov])
self.assertEqual(ov.metadata, o)
self.assertEqual(ov.observable.interaction.id, 'AtpSynthase')
self.assertEqual(ov.observable.specie.id, 'ATP')
self.assertEqual(ov.observable.compartment.id, 'c')
self.assertEqual(ov.observable.property, 'K_m')
self.assertEqual(ov.value, 1.0)
self.assertEqual(ov.error, 0.5)
self.assertEqual(ov.units, 'U/mg')
self.assertEqual(o.genetics.taxon, 'Mycoplasma pneumoniae')
self.assertEqual(o.genetics.variation, 'ΔMPN001')
self.assertEqual(o.environment.temperature, 37.)
self.assertEqual(o.environment.ph, 7.)
self.assertEqual(o.environment.media, 'Hayflick')
self.assertEqual(o.reference.title, 'title')
self.assertEqual(o.reference.author, 'author')
self.assertEqual(o.reference.year, 2017)
self.assertEqual(o.reference.volume, 1)
self.assertEqual(o.reference.number, 1)
self.assertEqual(o.reference.pages, '1-10')
self.assertEqual(o.method.name, 'assay')
self.assertEqual(o.method.description, 'description of assay')
def test_ReactionParticipantFilter(self):
atp = data_model.Specie(structure=(
'InChI=1S/C10H16N5O13P3/c11-8-5-9(13-2-12-8)15(3-14-5)10-7(17)6(16)4(26-10)1-25-30(21,22)28-31(23,24)27-29(18,19)20'
'/h2-4,6-7,10,16-17H,1H2,(H,21,22)(H,23,24)(H2,11,12,13)(H2,18,19,20)/p-4/t4-,6-,7-,10-/m1/s1'
))
h2o = data_model.Specie(structure='InChI=1S/H2O/h1H2')
adp = data_model.Specie(structure=(
'InChI=1S/C10H15N5O10P2/c11-8-5-9(13-2-12-8)15(3-14-5)10-7(17)6(16)4(24-10)1-23-27(21,22)25-26(18,19)20'
'/h2-4,6-7,10,16-17H,1H2,(H,21,22)(H2,11,12,13)(H2,18,19,20)/p-3/t4-,6-,7-,10-/m1/s1'
))
pi = data_model.Specie(structure='InChI=1S/H3O4P/c1-5(2,3)4/h(H3,1,2,3,4)/p-2')
h = data_model.Specie(structure='InChI=1S/p+1/i/hH')
glc = data_model.Specie(structure='InChI=1S/C6H12O6/c7-1-2-3(8)4(9)5(10)6(11)12-2/h2-11H,1H2/t2-,3-,4+,5-,6?/m1/s1')
glc_2 = data_model.Specie(structure='InChI=1S/C6H12O6/c7-1-2-3(8)4(9)5(10)6(11)12-2')
gtp = data_model.Specie(structure=(
'InChI=1S/C10H16N5O14P3/c11-10-13-7-4(8(18)14-10)12-2-15(7)9-6(17)5(16)3(27-9)1-26-31(22,23)29-32(24,25)28-30(19,20)21'
'/h2-3,5-6,9,16-17H,1H2,(H,22,23)(H,24,25)(H2,19,20,21)(H3,11,13,14,18)/p-4/t3-,5-,6-,9-/m1/s1'
))
lactate = data_model.Specie(structure='InChI=1S/C3H6O3/c1-2(4)3(5)6/h2,4H,1H3,(H,5,6)/p-1')
def get_reaction(ntp=atp, glc=glc):
return data_model.Reaction(
participants=[
data_model.ReactionParticipant(coefficient=-1, specie=ntp),
data_model.ReactionParticipant(coefficient=-1, specie=h2o),
data_model.ReactionParticipant(coefficient=1, specie=adp),
data_model.ReactionParticipant(coefficient=1, specie=pi),
data_model.ReactionParticipant(coefficient=1, specie=h),
data_model.ReactionParticipant(coefficient=0, specie=glc),
])
f = data_query.ReactionParticipantFilter()
rxn_atp = get_reaction(ntp=atp)
rxn_gtp = get_reaction(ntp=gtp)
rxn_lac = get_reaction(ntp=lactate)
rxn_glc_2 = get_reaction(glc=glc_2)
# identical reactant
ov = data_model.ObservedValue(observable=data_model.Observable(property='Km', specie=atp, interaction=rxn_atp))
self.assertEqual(f.score(rxn_atp, ov), 1)
# identical product
ov = data_model.ObservedValue(observable=data_model.Observable(property='Ki', specie=adp, interaction=rxn_atp))
self.assertEqual(f.score(rxn_atp, ov), 1)
# similar modifier
ov = data_model.ObservedValue(observable=data_model.Observable(property='Ki', specie=glc_2, interaction=rxn_glc_2))
self.assertEqual(f.score(rxn_atp, ov), 1)
# similar species
ov = data_model.ObservedValue(observable=data_model.Observable(property='Km', specie=gtp, interaction=rxn_gtp))
numpy.testing.assert_almost_equal(f.score(rxn_atp, ov), 0.767, decimal=3)
# different species
ov = data_model.ObservedValue(observable=data_model.Observable(property='Km', specie=lactate, interaction=rxn_lac))
self.assertEqual(f.score(rxn_atp, ov), -1)
# property without species
ov = data_model.ObservedValue(observable=data_model.Observable(property='kcat', interaction=rxn_atp))
self.assertEqual(f.score(rxn_atp, ov), 1)
def get_observed_result(self, reaction):
""" Find observed kinetics for the reaction or similar reactions
TODO: Add compartment infomrmation
1. Find kinetics observed for the reaction
a. Find the metabolite(s) of each participant
b. Find the reaction(s) which contain all of these metabolites
c. Find the kinetic laws associated with these reactions
2. Find kinetics observed for similar reactions
a. Find kinetics observed for the assigned EC number(s)
b. Find kinetics observed for EC number(s) predicted by tools such as E-zyme
Args:
reaction (:obj:`data_model.Reaction`): reaction to find data for
Returns:
:obj:`list` of :obj:`data_model.ObservedValue`: list of relevant observed values
"""
q_law = self.get_kinetic_laws_by_reaction(reaction)
observed_vals = []
for law in q_law:
common_schema_reaction_id = next(xr._id for xr in law._metadata.resource if xr.namespace == 'sabiork.reaction')
reaction = data_model.Reaction(
cross_references=[
data_model.Resource(namespace='common_schema.kinetic_law_id', id=str(law.kinetic_law_id)),
data_model.Resource(namespace='sabiork.reaction', id=common_schema_reaction_id),
],
)
species = {}
compartments = {}
cs_rxn = self.data_source.session.query(models.Reaction).filter_by(kinetic_law_id = law.kinetic_law_id)
reactants = cs_rxn.filter_by(_is_reactant = True).all()
products = cs_rxn.filter_by(_is_product = True).all()
modifiers = cs_rxn.filter_by(_is_modifier = True).all()
for reactant in reactants:
part = data_model.ReactionParticipant(coefficient=-1)
if reactant.metabolite_id not in species:
species[reactant.metabolite_id] = data_model.Specie(name=reactant.metabolite.metabolite_name)
part.specie = species[reactant.metabolite_id]
if reactant.metabolite.structure_id:
part.specie.structure = reactant.metabolite.structure._value_inchi
if reactant.compartment_id:
if reactant.compartment.name not in compartments:
compartments[reactant.compartment.name] = data_model.Compartment(name=reactant.compartment.name)
part.compartment = compartments[reactant.compartment.name]
reaction.participants.append(part)
for product in products:
part = data_model.ReactionParticipant(coefficient=1)
if product.metabolite_id not in species:
species[product.metabolite_id] = data_model.Specie(name=product.metabolite.metabolite_name)
part.specie = species[product.metabolite_id]
if product.metabolite.structure_id:
part.specie.structure = product.metabolite.structure._value_inchi
if product.compartment_id:
if product.compartment.name not in compartments:
compartments[product.compartment.name] = data_model.Compartment(name=product.compartment.name)
part.compartment = compartments[product.compartment.name]
reaction.participants.append(part)
for modifier in modifiers:
part = data_model.ReactionParticipant(coefficient=0)
if modifier.metabolite_id not in species:
species[modifier.metabolite_id] = data_model.Specie(name=modifier.metabolite.metabolite_name)
part.specie = species[modifier.metabolite_id]
if modifier.metabolite.structure_id:
part.specie.structure = modifier.metabolite.structure._value_inchi
if modifier.compartment_id:
if modifier.compartment.name not in compartments:
compartments[modifier.compartment.name] = data_model.Compartment(name=modifier.compartment.name)
part.compartment = compartments[modifier.compartment.name]
reaction.participants.append(part)
metadata = self.metadata_dump(law)
for parameter in law.parameter:
if parameter.value is None:
continue
observable = data_model.Observable(
interaction=reaction,
property=parameter.observed_name,
)
if parameter.metabolite_id:
observable.specie = species[parameter.metabolite_id]
# if parameter.compartment:
# observable.compartment = data_model.Compartment(
# id=parameter.compartment.name,
# )
observed_vals.append(data_model.ObservedValue(
metadata=metadata,
observable=observable,
value=parameter.value,
error=parameter.error,
units=parameter.units,
))
return observed_vals
def test_ReactionSimilarityFilter(self):
atp_structure = (
'InChI=1S/C10H16N5O13P3/c11-8-5-9(13-2-12-8)15(3-14-5)10-7(17)6(16)4(26-10)1-25-30(21,22)28-31(23,24)27-29(18,19)20'
'/h2-4,6-7,10,16-17H,1H2,(H,21,22)(H,23,24)(H2,11,12,13)(H2,18,19,20)/p-4/t4-,6-,7-,10-/m1/s1'
)
h2o_structure = 'InChI=1S/H2O/h1H2'
adp_structure = (
'InChI=1S/C10H15N5O10P2/c11-8-5-9(13-2-12-8)15(3-14-5)10-7(17)6(16)4(24-10)1-23-27(21,22)25-26(18,19)20'
'/h2-4,6-7,10,16-17H,1H2,(H,21,22)(H2,11,12,13)(H2,18,19,20)/p-3/t4-,6-,7-,10-/m1/s1'
)
pi_structure = 'InChI=1S/H3O4P/c1-5(2,3)4/h(H3,1,2,3,4)/p-2'
h_structure = 'InChI=1S/p+1/i/hH'
def get_reaction(pi_structure=pi_structure, ec='1.1.1.1'):
return data_model.Reaction(
participants=[
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=atp_structure)),
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=h2o_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=adp_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=pi_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=h_structure)),
],
cross_references=[
data_model.Resource(namespace='ec-code', id=ec)
])
rxn = get_reaction()
f = data_query.ReactionSimilarityFilter(min_ec_level=3, scale=1)
# same participants
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=get_reaction()))
numpy.testing.assert_almost_equal(f.score(rxn, ov), 1, decimal=3)
# similiar participants
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=get_reaction(pi_structure='InChI=1S/H3O4P/c1-5(2,3)4')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), 1, decimal=3)
# different participants, same 4-digit EC
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), math.exp(-1), decimal=3)
# different participants, same 3-digit EC
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1.2')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), math.exp(-2), decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), math.exp(-2), decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1.')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), math.exp(-2), decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1.-')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), math.exp(-2), decimal=3)
# different participants, same 2-digit EC
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.2.1')))
numpy.testing.assert_almost_equal(f.score(rxn, ov), -1, decimal=3)
# target reaction only has 3 digits
rxn_1_1_1 = get_reaction(ec='1.1.1')
rxn_1_1_1_1 = get_reaction(ec='1.1.1.1')
f1 = data_query.ReactionSimilarityFilter(min_ec_level=3, scale=1)
f2 = data_query.ReactionSimilarityFilter(min_ec_level=3, scale=1)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1')))
numpy.testing.assert_almost_equal(f1.score(rxn_1_1_1, ov), math.exp(-2), decimal=3)
numpy.testing.assert_almost_equal(f2.score(rxn_1_1_1_1, ov), math.exp(-2), decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.1.1')))
numpy.testing.assert_almost_equal(f1.score(rxn_1_1_1, ov), math.exp(-2), decimal=3)
numpy.testing.assert_almost_equal(f2.score(rxn_1_1_1_1, ov), math.exp(-1), decimal=3)
ov = data_model.ObservedValue(observable=data_model.Observable(
interaction=get_reaction(pi_structure='InChI=1S/H4O4P/c1-5(2,3)4', ec='1.1.2.1')))
numpy.testing.assert_almost_equal(f1.score(rxn_1_1_1, ov), -1, decimal=3)
numpy.testing.assert_almost_equal(f2.score(rxn_1_1_1_1, ov), -1, decimal=3)
# reverse direction, different numbers of reactants/products
f = data_query.ReactionSimilarityFilter(min_ec_level=3, scale=1)
for_rxn = get_reaction()
rev_rxn = get_reaction()
for part in rev_rxn.participants:
part.coefficient = -1 * part.coefficient
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=for_rxn))
self.assertEqual(f.score(for_rxn, ov), 1.)
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=rev_rxn))
self.assertEqual(f.score(for_rxn, ov), -1.)
# reverse direction, same numbers of reactants/products
for_rxn = data_model.Reaction(
participants=[
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=atp_structure)),
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=adp_structure)),
],
cross_references=[
data_model.Resource(namespace='ec-code', id='1.1.1.1')
])
rev_rxn = data_model.Reaction(
participants=[
data_model.ReactionParticipant(coefficient=1, specie=data_model.Specie(structure=atp_structure)),
data_model.ReactionParticipant(coefficient=-1, specie=data_model.Specie(structure=adp_structure)),
],
cross_references=[
data_model.Resource(namespace='ec-code', id='1.1.1.1')
])
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=for_rxn))
self.assertEqual(f.score(for_rxn, ov), 1.)
ov = data_model.ObservedValue(observable=data_model.Observable(interaction=rev_rxn))
self.assertEqual(f.score(for_rxn, ov), -1.)
def test_to_mol(self):
s = data_model.Specie(structure=self.pi)
self.assertIsInstance(s.to_mol(), str)
self.assertNotEqual(s.to_mol(), '')
def test_to_openbabel(self):
s = data_model.Specie(structure=self.pi)
self.assertEqual(s.to_openbabel().GetFormula(), 'HO4P--')
def test_to_pybel(self):
s = data_model.Specie(structure=self.pi)
self.assertEqual(s.to_pybel().formula, 'HO4P--')
def test_to_smiles(self):
s = data_model.Specie(structure=self.pi)
self.assertEqual(s.to_smiles(), '[O-]P(=O)(O)[O-]')
def test_run(self):
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['atp'], id='atp'),
compartment=data_model.Compartment(
id='c'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['h2o'], id='h2o'),
compartment=data_model.Compartment(
id='c'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['adp'], id='adp'),
compartment=data_model.Compartment(
id='c'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['ppi'], id='ppi'),
compartment=data_model.Compartment(
id='c'),
coefficient=1,
order=3),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['h'], id='h'),
compartment=data_model.Compartment(
id='c'),
coefficient=1,
order=4),
])
self.assertEqual(rxn.get_reactants()[0].specie.id, 'atp')
self.assertEqual(rxn.get_products()[0].specie.id, 'adp')
self.assertEqual(rxn.get_reactant_product_pairs()[0][0].specie.id,
'atp')
self.assertEqual(rxn.get_reactant_product_pairs()[0][1].specie.id,
'adp')
self.assertEqual(rxn.get_reactant_product_pairs()[1][0].specie.id,
'h2o')
self.assertEqual(rxn.get_reactant_product_pairs()[1][1].specie.id,
'ppi')
result = ezyme.Ezyme().run(rxn)
print(result)
self.assertEqual(result[0].ec_number, '3.6.1') # true EC is 3.6.1.3
# example where Ezyme predicts no EC number when the order is swapped
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['t3p1'], id='atp'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['s7p'], id='h2o'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['x5p'], id='adp'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['r5p'], id='ppi'),
coefficient=1,
order=3),
])
result = ezyme.Ezyme().run(rxn)
self.assertEqual(result[0].ec_number, "4.1.2") # true EC is 2.2.1.1
self.assertEqual(result[1].ec_number, "2.2.1")
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['s7p'], id='atp'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['t3p1'], id='h2o'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['x5p'], id='adp'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['r5p'], id='ppi'),
coefficient=1,
order=3),
])
result = ezyme.Ezyme().run(rxn)
self.assertEqual(result, [])
# example where Ezyme predicts different EC numbers when the order is swapped
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dGMP'], id='atp'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['h2o'], id='h2o'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dG'], id='adp'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['pi'], id='ppi'),
coefficient=1,
order=3),
])
result = ezyme.Ezyme().run(rxn)
self.assertEqual(result[0].ec_number, "3.1.3") # true EC is 3.1.3.89
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['h2o'], id='atp'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dGMP'], id='h2o'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dG'], id='adp'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['pi'], id='ppi'),
coefficient=1,
order=3),
])
result = ezyme.Ezyme().run(rxn)
self.assertEqual(result[0].ec_number, "3.1.3") # true EC is 3.1.3.89
# example where a structure is not defined
rxn = data_model.Reaction(participants=[
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['h2o'], id='atp'),
coefficient=-1,
order=1),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dGMP'], id='h2o'),
coefficient=-1,
order=0),
data_model.ReactionParticipant(specie=data_model.Specie(
structure=self.molecules['e1dG'], id='adp'),
coefficient=1,
order=2),
data_model.ReactionParticipant(specie=data_model.Specie(
structure='', id='ppi'),
coefficient=1,
order=3),
])
result = ezyme.Ezyme().run(rxn)
self.assertEqual(result, None)
def get_similarity(self):
adp = data_model.Specie(structure=seld.adp)
atp = data_model.Specie(structure=seld.atp)
self.assertAlmostEqual(adp.get_similarity(atp), 0.955, places=3)
def _port(self, metabolite):
structure = metabolite.structure._value_inchi if metabolite.structure else None
references = [data_model.Resource(namespace=item.namespace, id=item._id) for item in metabolite._metadata.resource]
return data_model.Specie(id=metabolite.metabolite_id, name=metabolite.metabolite_name, structure = structure, cross_references = references)
def test_parse_reaction_equation(self):
c = data_model.Compartment(id='c')
e = data_model.Compartment(id='e')
compartments = [c, e]
Complex_Acp = data_model.Specie(id='Complex_Acp')
ATP = data_model.Specie(id='ATP')
HDCA = data_model.Specie(id='HDCA')
PPI = data_model.Specie(id='PPI')
AMP = data_model.Specie(id='AMP')
Complex_Acp_hdc = data_model.Specie(id='Complex_Acp_hdc')
A = data_model.Specie(id='A')
B2 = data_model.Specie(id='B2')
B_2 = data_model.Specie(id='B_2')
C = data_model.Specie(id='C')
D = data_model.Specie(id='D')
E = data_model.Specie(id='E')
F = data_model.Specie(id='F')
species = [
Complex_Acp,
ATP,
HDCA,
PPI,
AMP,
Complex_Acp_hdc,
A,
B2,
B_2,
C,
D,
E,
F,
]
eq = "[c]: Complex_Acp + ATP + HDCA ==> PPI + AMP + Complex_Acp_hdc"
response = io.InputReader().parse_reaction_equation(eq, compartments, species)
expected_answer = data_model.Reaction(
participants=[
data_model.ReactionParticipant(specie=Complex_Acp, compartment=c, coefficient=-1, order=0),
data_model.ReactionParticipant(specie=ATP, compartment=c, coefficient=-1, order=1),
data_model.ReactionParticipant(specie=HDCA, compartment=c, coefficient=-1, order=2),
data_model.ReactionParticipant(specie=PPI, compartment=c, coefficient=1, order=3),
data_model.ReactionParticipant(specie=AMP, compartment=c, coefficient=1, order=4),
data_model.ReactionParticipant(specie=Complex_Acp_hdc, compartment=c, coefficient=1, order=5),
],
reversible=False,
)
self.assertTrue(self.are_reactions_equal(response, expected_answer))
eq = "A[c] + B2[e] + 2 C[c] ==> 2.5 D[c] + E[c] + F[c]"
response = io.InputReader().parse_reaction_equation(eq, compartments, species)
expected_answer = data_model.Reaction(
participants=[
data_model.ReactionParticipant(specie=A, compartment=c, coefficient=-1., order=0),
data_model.ReactionParticipant(specie=B2, compartment=e, coefficient=-1., order=1),
data_model.ReactionParticipant(specie=C, compartment=c, coefficient=-2., order=2),
data_model.ReactionParticipant(specie=D, compartment=c, coefficient=2.5, order=3),
data_model.ReactionParticipant(specie=E, compartment=c, coefficient=1., order=4),
data_model.ReactionParticipant(specie=F, compartment=c, coefficient=1., order=5),
],
reversible=False,
)
self.assertTrue(self.are_reactions_equal(response, expected_answer))
eq = "A[c]+B_2[e]+2 C[c]<==>2.5 D[c] + E[c] + F[c]"
response = io.InputReader().parse_reaction_equation(eq, compartments, species)
expected_answer = data_model.Reaction(
participants=[
data_model.ReactionParticipant(specie=A, compartment=c, coefficient=-1., order=0),
data_model.ReactionParticipant(specie=B_2, compartment=e, coefficient=-1., order=1),
data_model.ReactionParticipant(specie=C, compartment=c, coefficient=-2., order=2),
data_model.ReactionParticipant(specie=D, compartment=c, coefficient=2.5, order=3),
data_model.ReactionParticipant(specie=E, compartment=c, coefficient=1., order=4),
data_model.ReactionParticipant(specie=F, compartment=c, coefficient=1., order=5),
],
reversible=True,
)
self.assertTrue(self.are_reactions_equal(response, expected_answer))
# alternative separators
expected_answer = data_model.Reaction(
participants=[
data_model.ReactionParticipant(specie=A, compartment=c, coefficient=-1., order=0),
data_model.ReactionParticipant(specie=B2, compartment=e, coefficient=-1., order=1),
data_model.ReactionParticipant(specie=C, compartment=c, coefficient=-2., order=2),
data_model.ReactionParticipant(specie=D, compartment=c, coefficient=2.5, order=3),
data_model.ReactionParticipant(specie=E, compartment=c, coefficient=1., order=4),
data_model.ReactionParticipant(specie=F, compartment=c, coefficient=1., order=5),
],
reversible=False,
)
eq = "A[c] + B2[e] + 2 C[c] ==> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] => 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] --> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] -> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
expected_answer = data_model.Reaction(
participants=[
data_model.ReactionParticipant(specie=A, compartment=c, coefficient=-1., order=0),
data_model.ReactionParticipant(specie=B2, compartment=e, coefficient=-1., order=1),
data_model.ReactionParticipant(specie=C, compartment=c, coefficient=-2., order=2),
data_model.ReactionParticipant(specie=D, compartment=c, coefficient=2.5, order=3),
data_model.ReactionParticipant(specie=E, compartment=c, coefficient=1., order=4),
data_model.ReactionParticipant(specie=F, compartment=c, coefficient=1., order=5),
],
reversible=True,
)
eq = "A[c] + B2[e] + 2 C[c] <==> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] <=> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] <--> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
eq = "A[c] + B2[e] + 2 C[c] <-> 2.5 D[c] + E[c] + F[c]"
self.are_reactions_equal(io.InputReader().parse_reaction_equation(eq, compartments, species), expected_answer)
# negative examples
eq = "[c]: A[c] + B2[e] + 2 C[c] ==> 2.5 D[c] + E[c] + F[c]"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
eq = "[c]: A + B2 + 2 C[c] ==> 2.5 D + E + F"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
eq = "A + B2 + 2 C[c] ==> 2.5 D + E + F"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
eq = "A[c] + B-2[e] + 2 C[c] ==> 2.5 D[c] + E[c] + F[c]"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
eq = "A[c] + B2[e] + 2 C[c] ===> 2.5 D[c] + E[c] + F[c]"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
eq = "A[c] + B2[e] + 2 C[c] ---> 2.5 D[c] + E[c] + F[c]"
self.assertRaises(ValueError, io.InputReader().parse_reaction_equation, eq, compartments, species)
