def test_basic_gecko_adjustment():
in_model = {'P00549': 0.1, 'P31373': 0.1, 'P31382': 0.1, 'P39708': 0.1, 'P39714': 0.1, 'P39726': 0.1, 'Q01574': 0.1}
not_in_model = {'P10591': 0.1, 'P31383': 0.1, 'P32471': 0.1}
measurements = pd.concat([pd.Series(in_model), pd.Series(not_in_model)])
model = GeckoModel('multi-pool')
model.limit_proteins(fractions=pd.Series(measurements))
sol = model.optimize()
assert sol.objective_value > 0.05
assert len(model.proteins) - len(model.pool_proteins) - len(in_model) == 0
assert all(rxn.upper_bound > 0 for rxn in model.individual_protein_exchanges)
def essential_prot():
model = GeckoModel("single-pool")
model.solver = 'cplex'
print("Essential proteins with cplex, normal model (single-pool)")
for p in model.proteins:
with model as m:
r = m.reactions.get_by_id("draw_prot_" + p)
r.lower_bound = 0
r.upper_bound = 0
res = m.optimize()
if (res.objective_value <= 1e-10):
print(p, ",", res.objective_value)
def analysis_ko(resFileName):
levels = [0, 1e-10, 1e-8, 1e-6, 1e-4, 1e-2, 0.1]
model = GeckoModel('single-pool')
df = pandas.DataFrame(index=range(100), columns=["ko", "Biomass"])
for i in range(100):
proteins = random.sample(model.proteins, 10)
dic = {p: 0 for p in proteins}
model.limit_proteins(pandas.Series(dic))
res = model.optimize()
df.loc[i] = (dic, res.objective_value)
df.to_csv(resFileName)
def test_adjust_pool_bounds():
essential = {'P00498': 0., 'P00815': 0.}
in_model = {'P00549': 0.1, 'P31373': 0.1, 'P31382': 0.1, 'P39708': 0.1, 'P39714': 0.1, 'P39726': 0.1, 'Q01574': 0.1}
expected = set('prot_{}_exchange'.format(pool_id) for pool_id in essential)
measurements = pd.concat([pd.Series(in_model), pd.Series(essential)])
model = GeckoModel('multi-pool')
model.limit_proteins(fractions=pd.Series(measurements))
assert model.slim_optimize() < 1e-3
model.adjust_pool_bounds(inplace=False)
assert model.slim_optimize() < 1e-3
adjustment = model.adjust_pool_bounds(inplace=True)
observed = set(adjustment['reaction'])
assert abs(model.slim_optimize() - 0.05) < 1e-3
assert expected == observed
def prot_ko_optim(prot_measure_fractions=None,
prot_measure_ggdw=None,
constraints=None,
isMultiProc=False,
size=1):
#load model
if prot_measure_fractions is None and prot_measure_ggdw is None:
model = GeckoModel("single-pool")
else:
model = GeckoModel("multi-pool")
if prot_measure_fractions:
model.limit_proteins(fractions=prot_measure_fractions)
else:
model.limit_proteins(ggdw=prot_measure_ggdw)
fileRes = basePath + "Results/optim_KO_Gecko_Yeast_SUCC_max5.csv"
simulProb = GeckoSimulationProblem(model, constraints=constraints)
evalFunc = build_evaluation_function("BPCY", "r_2111", "r_2056",
"r_1714_REV") # max succ exchange
gecko_strain_optim(simulProb,
evaluationFunc=evalFunc,
levels=None,
isMultiProc=isMultiProc,
candidateSize=size,
resultFile=fileRes) #KO_Reaction by default
def essential_prot_scale():
model = GeckoModel("single-pool")
model.solver = 'cplex'
print("Essential proteins with cplex, scale model (single-pool)")
for r in model.reactions:
r.lower_bound = r.lower_bound * 100000
r.upper_bound = r.upper_bound * 100000
for p in model.proteins:
with model as m:
r = model.reactions.get_by_id("draw_prot_" + p)
r.lower_bound = 0
r.upper_bound = 0
res = model.optimize()
print(p, ",", res.objective_value)
def simulate_prot():
model = GeckoModel("single-pool")
model.solver = 'cplex'
with model:
# for p in ["P53685","Q01574"]:
for p in ['P33421']:
r = model.reactions.get_by_id("draw_prot_" + p)
r.lower_bound = 0
r.upper_bound = 0
res = model.optimize()
print(" --> growth " + str(res.objective_value))
print(" --> r_2111 " + str(res.fluxes["r_2111"]))
print(" --> r_2056 " + str(res.fluxes["r_2056"]))
print(" --> r_1714 " + str(res.fluxes["r_1714_REV"]))
print(" ------------ ")
def simulate_wt():
model = GeckoModel('single-pool')
res = model.optimize()
print(res)
#for p in model.proteins:
p = "P38066"
with model:
r = model.reactions.get_by_id("draw_prot_" + p)
lb = r.lower_bound
ub = r.upper_bound
r.lower_bound = 0
r.upper_bound = 0.000001
res = model.optimize()
#r.knock_out()
#res = model.optimize()
print(p + " wt simulation1 " + str(res.objective_value))
print(str(r.lower_bound) + " --> " + str(r.upper_bound))
def analysis_growth(resFileName):
levels = [0, 1e-10, 1e-8, 1e-6, 1e-4, 1e-2, 0.1]
model = GeckoModel('single-pool')
proteins = model.proteins
df = pandas.DataFrame(index=proteins, columns=levels)
for p in proteins:
print(p)
if p != "P38066":
for level in levels:
r = model.reactions.get_by_id("draw_prot_" + p)
lb = r.lower_bound
ub = r.upper_bound
r.lower_bound = 0
r.upper_bound = level
res = model.optimize()
df.loc[p][level] = res.objective_value
r.lower_bound = lb
r.upper_bound = ub
df.to_csv(resFileName)
def test_gecko_adjustment_sanchez_etal():
mmol_gdw = pd.read_csv(os.path.join(
os.path.dirname(__file__),
'../geckopy/data_files/sanchez-mmol_gdw.csv'),
index_col=0,
header=None,
squeeze=True)
ggdw = pd.Series(PROTEIN_PROPERTIES.loc[mmol_gdw.index, 'mw'] /
1000.) * pd.Series(mmol_gdw)
model = GeckoModel('multi-pool')
growth_rate_unlimited_protein = model.slim_optimize()
model.limit_proteins(ggdw=pd.Series(ggdw))
growth_rate_limited_protein = model.slim_optimize()
# should be smaller, but how much..
assert growth_rate_limited_protein < 0.8 * growth_rate_unlimited_protein
measured_in_model = set(mmol_gdw.index).intersection(model.proteins)
assert sum(model.concentrations[p] - ggdw[p]
for p in measured_in_model) < 1e-10
assert sum(
abs(rxn.upper_bound - mmol_gdw[rxn.annotation['uniprot']])
for rxn in model.individual_protein_exchanges) < 1e-6
assert sum(rxn.metabolites[model.common_protein_pool] +
PROTEIN_PROPERTIES.loc[rxn.annotation['uniprot'], 'mw'] / 1000.
for rxn in model.pool_protein_exchanges) < 1e-6
assert model.p_measured > 0.25 # With yeast 8.1.3 -> p_measured = 0.296
assert model.f_mass_fraction_measured_matched_to_total > 0.25 # With yeast 8.1.3 -> f = 0.304
assert model.protein_pool_exchange.upper_bound > 0.015 # With yeast 8.1.3 -> pool_exchange = 0.0212
def test_gecko_adjustment_sanchez_etal():
mmol_gdw = pd.Series.from_csv(
os.path.join(os.path.dirname(__file__),
'../geckopy/data_files/sanchez-mmol_gdw.csv'))
ggdw = pd.Series(PROTEIN_PROPERTIES.loc[mmol_gdw.index, 'mw'] /
1000.) * pd.Series(mmol_gdw)
model = GeckoModel('multi-pool')
growth_rate_unlimited_protein = model.slim_optimize()
model.limit_proteins(ggdw=pd.Series(ggdw))
growth_rate_limited_protein = model.slim_optimize()
# should be smaller, but how much..
assert growth_rate_limited_protein < 0.8 * growth_rate_unlimited_protein
measured_in_model = set(mmol_gdw.index).intersection(model.proteins)
assert sum(model.concentrations[p] - ggdw[p]
for p in measured_in_model) < 1e-10
assert sum(
abs(rxn.upper_bound - mmol_gdw[rxn.annotation['uniprot']])
for rxn in model.individual_protein_exchanges) < 1e-6
assert abs(model.p_measured - 0.296) < 1e-2 # section 3.2.6 reports 0.283
assert sum(rxn.metabolites[model.common_protein_pool] +
PROTEIN_PROPERTIES.loc[rxn.annotation['uniprot'], 'mw'] / 1000.
for rxn in model.pool_protein_exchanges) < 1e-6
# FIXME: section 3.2.6 reports 0.2154
assert abs(model.f_mass_fraction_measured_matched_to_total - 0.291) < 1e-2
# FIXME: provided model had 0.0168, value including p_base term is 0.0507 as in matlab..
assert abs(model.protein_pool_exchange.upper_bound - 0.0203) < 1e-2
sanchez_biomass = pd.Series.from_csv(
os.path.join(os.path.dirname(__file__),
'../geckopy/data_files/sanchez-biomass.csv'))
biomass = pd.Series(
dict((m.id, v) for m, v in model.reactions.r_4041.metabolites.items()))
# FIXME: poor match with provided model for ATP, ADP, H+, H20, P
assert sum((biomass - sanchez_biomass).abs()) < 20
def analysis_growth(resFileName, scale=False):
model = GeckoModel('single-pool')
model.solver = 'cplex'
#scale model
if scale:
for r in model.reactions:
r.upper_bound = r.upper_bound * 100000
r.lower_bound = r.lower_bound * 100000
proteins = model.proteins
df = pandas.DataFrame(index=proteins, columns=range(100))
for i in range(100):
print(i)
for p in proteins:
with model as m:
r = m.reactions.get_by_id("draw_prot_" + p)
r.lower_bound = 0
r.upper_bound = 0
res = m.optimize()
df.loc[p][i] = 0 if res.objective_value < 1e-4 else 1
df.to_csv(resFileName)
def loading_yeast_gecko(prot_measure_fractions=None, prot_measure_ggdw=None):
"""
Loads the provided yeast gecko
:param pd.Series prot_measure_fractions: measured fraction of proteins
:param pd.Series prot_measure_ggdw: measured ggdw of proteins
:return GeckoModel:
"""
if prot_measure_fractions is None and prot_measure_ggdw is None:
model = GeckoModel("single-pool")
else:
model = GeckoModel("multi-pool")
if prot_measure_fractions:
model.limit_proteins(fractions=prot_measure_fractions)
else:
model.limit_proteins(ggdw=prot_measure_ggdw)
return model
def simulate_wt_multi():
model = GeckoModel('multi-pool')
import pandas
some_measurements = pandas.Series({
'P00549': 0.1,
'P31373': 0.1,
'P31382': 0.1
})
model = GeckoModel('multi-pool')
model.limit_proteins(some_measurements)
res = model.optimize()
print(" wt simulation1 ", res.objective_value)
for r in model.reactions:
print(r.id, " --> ", res.fluxes[r.id])
def loading_any_gecko(path, biomass, protein=None, carbs=None):
"""
Default templeate function to load geckos
:param str path: string path to the sbml file
:param str biomass: biomass function id
:param str protein: protein reaction id
:param str carbs: carbohydrate function id
:return GeckoModel:
"""
if not protein:
protein = biomass
if not carbs:
carbs = biomass
any_sbml_model = read_sbml_model(path)
any_gecko = GeckoModel(model=any_sbml_model,
biomass_reaction_id=biomass,
protein_reaction_id=protein,
carbohydrate_reaction_id=carbs)
return any_gecko
pool.join()
except (OSError, RuntimeError) as e:
logger.error('failed parallel_evaluation_mp: {0}'.format(str(e)))
raise
else:
end = time.time()
print('completed parallel_evaluation_mp in {0} seconds'.format(end -
start))
logger.debug(
'completed parallel_evaluation_mp in {0} seconds'.format(end -
start))
# print("--- %s seconds ---" % (time.time() - start), 'end_pop')
return [r.get()[0] for r in results]
model = GeckoModel('single-pool')
const = {'r_1714_REV': (0, 1)}
simulProblem = GeckoSimulationProblem(model, const)
ids = [
x for x in simulProblem.model.proteins
if x not in simulProblem.objective.keys()
]
decoder = DecoderProtKnockouts(ids)
evalFunc = build_evaluation_function("targetFlux", ["r_2056"])
eaConfig = EAConfigurations()
optimProbConf = OptimProblemConfiguration(
simulProblem,
type=optimType.PROTEIN_KO,
decoder=decoder,
evaluationFunc=evalFunc,