marsi.cobra.strain_design package

Submodules

marsi.cobra.strain_design.design_methods module

class marsi.cobra.strain_design.design_methods.GenericMARSIDesignMethod(model=None, nearest_neighbors_model=None, min_tanimoto=0.75, currency_metabolites=None, essential_metabolites=None)[source]

Bases: object

Generic wrapper for Metabolite Analog design method.

This one just runs a optimization method

Example

>>> from marsi.cobra import strain_design
>>> from cameo import models
>>> from cameo.strain_design import OptGene
>>> designer = strain_design.GenericMARSIDesignMethod(model=models.bigg.iJO1366)
>>> designer.optimize_with_reaction("succ_e", max_interventions=5, substrate="EX_glc__D_e",
>>> biomass="BIOMASS_Ec_iJO1366_core_53p95M", max_results=25, design_method=OptGene, manipulation_type="reactions")

Methods

essential_metabolites_reactions()
optimize_with_metabolites(target[, …])
optimize_with_reaction(target[, …])
optimize_with_reaction(target, max_interventions=1, substrate=None, biomass=None, design_method=<class 'cameo.strain_design.deterministic.linear_programming.OptKnock'>, max_results=100, non_essential_metabolites=False, max_evaluations=20000, **design_kwargs)[source]
optimize_with_metabolites(target, max_interventions=1, substrate=None, biomass=None, max_results=100, non_essential_metabolites=False, max_evaluations=20000, **design_kwargs)[source]
essential_metabolites_reactions()[source]
class marsi.cobra.strain_design.design_methods.RandomMutagenesisDesign(model=None, nearest_neighbors_model=None, min_tanimoto=0.75, currency_metabolites=None, essential_metabolites=None)[source]

Bases: marsi.cobra.strain_design.design_methods.GenericMARSIDesignMethod

Apply only knockout like designs where total loss of functions are expected.

Methods

essential_metabolites_reactions()
optimize_with_metabolites(target[, …])
optimize_with_reaction(target[, …])
optimize_with_reaction(target, max_interventions=1, substrate=None, biomass=None, design_method='optgene', max_results=100, non_essential_metabolites=False, max_evaluations=20000, **designer_kwargs)[source]
optimize_with_metabolites(target, max_interventions=1, substrate=None, biomass=None, max_results=100, max_evaluations=20000, **design_kwargs)[source]
class marsi.cobra.strain_design.design_methods.ALEDesign(model=None, nearest_neighbors_model=None, min_tanimoto=0.75, currency_metabolites=None, essential_metabolites=None)[source]

Bases: marsi.cobra.strain_design.design_methods.GenericMARSIDesignMethod

Apply both knockout and flux modulation. The strains will be selected via growth rate proxy.

Methods

essential_metabolites_reactions()
optimize_with_metabolites(target[, …])
optimize_with_reaction(target[, …])
optimize_with_reaction(target, max_interventions=1, substrate=None, biomass=None, design_method='differential_fva', max_results=100, non_essential_metabolites=False, **designer_kwargs)[source]
optimize_with_metabolites(target, max_interventions=1, substrate=None, biomass=None, max_results=100, max_evaluations=20000, **design_kwargs)[source]

marsi.cobra.strain_design.evolutionary module

class marsi.cobra.strain_design.evolutionary.OptMet(model, evolutionary_algorithm=<class 'inspyred.ec.ec.GA'>, essential_metabolites=None, plot=True, *args, **kwargs)[source]

Bases: cameo.core.strain_design.StrainDesignMethod

Attributes

manipulation_type
plot

Methods

__call__(*args, **kwargs)
run([target, biomass, substrate, …])
param target:The design target
manipulation_type
plot
run(target=None, biomass=None, substrate=None, max_knockouts=5, variable_size=True, simulation_method=<function fba>, growth_coupled=False, max_evaluations=20000, population_size=200, max_results=50, seed=None, **kwargs)[source]
Parameters:
  • target (str, Metabolite or Reaction) – The design target
  • biomass (str, Metabolite or Reaction) – The biomass definition in the model
  • substrate (str, Metabolite or Reaction) – The main carbon source
  • max_knockouts (int) – Max number of knockouts allowed
  • variable_size (bool) – If true, all candidates have the same size. Otherwise the candidate size can be from 1 to max_knockouts.
  • simulation_method (function) – Any method from cameo.flux_analysis.simulation or equivalent
  • growth_coupled (bool) – If true will use the minimum flux rate to compute the fitness
  • max_evaluations (int) – Number of evaluations before stop
  • population_size (int) – Number of individuals in each generation
  • max_results (int) – Max number of different designs to return if found.
  • kwargs (dict) – Arguments for the simulation method.
  • seed (int) – A seed for random.
Returns:
Return type:

OptMetResult

marsi.cobra.strain_design.metaheuristic module

class marsi.cobra.strain_design.metaheuristic.MetaboliteKnockoutOptimization(metabolites=None, essential_metabolites=None, n_carbons=2, compartments='c', skip_essential_metabolites=False, *args, **kwargs)[source]

Bases: cameo.strain_design.heuristic.evolutionary.optimization.TargetOptimization

Knockout optimization using metabolites.

model

SolverBasedModel – A constraint-based model.

heuristic_method

inspyred.ec.EvolutionaryComputation – An evolutionary algorithm.

objective_function

objective function or list(objective function) – The objectives for the algorithm to maximize.

seed

int – A seed for random. It is auto-generated if None is given.

termination

inspyred.ec.terminators – A termination criteria for the algorithm. The default is inspyred.ec.terminators.evaluation_termination.

simulation_method

flux_analysis.simulation – The method used to simulate the model.

wt_reference

dict – A reference initial state for the optimization. It is required for flux_analysis.simulation.lmoma and flux_analysis.simulation.room. If not given, it will be computed using flux_analysis.simulation.pfba

metabolites

list – A list of valid metabolites to knockout. If None, then all metabolites in the model will be knockout candidates except the ones defined in essential_metabolites

essential_metabolites

list – A list of metabolites that cannot be knocked out. If None, then all essential genes will be removed from the valid genes set.

run(view=config.default_view, maximize=True, **kwargs)

See also

inspyred, cameo.config.default_view

Example

>>> from cameo import models
>>> model = models.bigg.iJO1366
>>> from cameo.strain_design.heuristic.evolutionary.objective_functions import biomass_product_coupled_yield
>>> bpcy = biomass_product_coupled_yield(model.reactions.Ec_biomass_iJO1366_core_53p95,
>>>                                      model.reactions.EX_succ_e,
>>>                                      model.reactions.EX_glc__D_e)
>>> knockout_optimization = MetaboliteKnockoutOptimization(model=model, objective_function=bpcy)
>>> knockout_optimization.run(max_evaluations=50000)

Attributes

archiver
heuristic_method
metadata
objective_function
random
simulation_kwargs
simulation_method

Methods

run([max_size, variable_size, diversify, view]) Runs the evolutionary algorithm.
simulation_method
simulation_kwargs

marsi.cobra.strain_design.post_processing module

marsi.cobra.strain_design.post_processing.find_anti_metabolite_knockouts(reaction, ref_flux=0, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True)[source]

Generates a dictionary {species_id -> MetaboliteKnockoutTarget}.

Parameters:
  • reaction (cobra.Reaction) – A COBRA reaction
  • ref_flux (float) – The flux from the reference state (0 if unknown)
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (currency metabolites, etc.)
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
Returns:
Return type:

dict
marsi.cobra.strain_design.post_processing.find_anti_metabolite_modulation(reaction, fold_change, essential_metabolites, ref_flux=0, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True)[source]

Generates a dictionary {species_id -> AntiMetaboliteManipulationTarget}.

If the fold change > 0:
  1. Search for metabolites that are essential.
  2. Calculate fraction using a link function.
If fold change < 0:
  1. Search for metabolites that are not essential
  2. Calculated fraction is 1 - fold change
Parameters:
  • reaction (cobra.Reaction) – A COBRA reaction.
  • fold_change (float) – The fold change of the reaction flux.
  • essential_metabolites (list) – A list of essential metabolites.
  • ref_flux (float) – The flux from the reference state (0 if unknown)
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc.).
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
Returns:

{MetaboliteID -> AntiMetaboliteManipulationTarget}
Return type:

dict
marsi.cobra.strain_design.post_processing.convert_strain_design_results(model, results, objective_function, simulation_method, simulation_kwargs=None, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True, nullspace_matrix=None, essential_metabolites=None, max_loss=0.2)[source]

Converts a StrainDesignMethodResult into a DataFrame of possible substitutions.

Parameters:
  • model (cobra.Model) – A COBRA model.
  • results (cameo.core.strain_design.StrainDesignMethodResult) – The results of a strain design method.
  • objective_function (cameo.strain_design.heuristic.evolutionary.objective_functions.ObjectiveFunction) – The cellular objective to evaluate.
  • simulation_method (cameo.flux_analysis.simulation.fba or equivalent) – The method to compute a flux distribution using a COBRA model.
  • simulation_kwargs (dict) – The arguments for the simulation_method.
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc).
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • nullspace_matrix (numpy.ndarray) – The nullspace of the model.
  • essential_metabolites (list) – A list of essential metabolites.
  • max_loss (float) – A number between 0 and 1 for how much the fitness is allowed to drop with the metabolite target.
Returns:

A data frame with the possible replacements.
Return type:

pandas.DataFrame
marsi.cobra.strain_design.post_processing.convert_design(model, strain_design, fitness, objective_function, simulation_method, simulation_kwargs=None, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True, nullspace_matrix=None, essential_metabolites=None, max_loss=0.2)[source]

Converts a StrainDesign into a DataFrame of possible substitutions.

Parameters:
  • model (cobra.Model) – A COBRA model.
  • strain_design (cameo.core.strain_design.StrainDesign) – The results of a strain design method.
  • objective_function (cameo.strain_design.heuristic.evolutionary.objective_functions.ObjectiveFunction) – The cellular objective to evaluate.
  • simulation_method (cameo.flux_analysis.simulation.fba or equivalent) – The method to compute a flux distribution using a COBRA model.
  • simulation_kwargs (dict) – The arguments for the simulation_method.
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc).
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • nullspace_matrix (numpy.ndarray) – The nullspace of the model.
  • essential_metabolites (list) – A list of essential metabolites.
  • max_loss (float) – A number between 0 and 1 for how much the fitness is allowed to drop with the metabolite target.
Returns:

A data frame with the possible replacements.
Return type:

pandas.DataFrame
marsi.cobra.strain_design.post_processing.test_target_substitutions(model, all_targets, target, replacement_targets, objective_function, fitness, base_fitness, simulation_method, simulation_kwargs, reference, loss_validation, results)[source]
marsi.cobra.strain_design.post_processing.replace_strain_design_results(model, results, objective_function, simulation_method, simulation_kwargs=None, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True, essential_metabolites=None, max_loss=0.2)[source]

Converts a StrainDesignMethodResult into a DataFrame of possible substitutions.

Parameters:
  • model (cobra.Model) – A COBRA model.
  • results (cameo.core.strain_design.StrainDesignMethodResult) – The results of a strain design method.
  • objective_function (cameo.strain_design.heuristic.evolutionary.objective_functions.ObjectiveFunction) – The cellular objective to evaluate.
  • simulation_method (cameo.flux_analysis.simulation.fba or equivalent) – The method to compute a flux distribution using a COBRA model.
  • simulation_kwargs (dict) – The arguments for the simulation_method.
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc).
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • essential_metabolites (list) – A list of essential metabolites.
  • max_loss (float) – A number between 0 and 1 for how much the fitness is allowed to drop with the metabolite target.
Returns:

A data frame with the possible replacements.
Return type:

pandas.DataFrame
marsi.cobra.strain_design.post_processing.replace_design(model, strain_design, fitness, objective_function, simulation_method, simulation_kwargs=None, ignore_metabolites=None, ignore_transport=True, allow_accumulation=True, essential_metabolites=None, max_loss=0.2, allow_modulation=True)[source]

Converts a StrainDesign into a DataFrame of possible substitutions.

Parameters:
  • model (cobra.Model) – A COBRA model.
  • strain_design (cameo.core.strain_design.StrainDesign) – The results of a strain design method.
  • objective_function (cameo.strain_design.heuristic.evolutionary.objective_functions.ObjectiveFunction) – The cellular objective to evaluate.
  • simulation_method (cameo.flux_analysis.simulation.fba or equivalent) – The method to compute a flux distribution using a COBRA model.
  • simulation_kwargs (dict) – The arguments for the simulation_method.
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc).
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • essential_metabolites (list) – A list of essential metabolites.
  • max_loss (float) – A number between 0 and 1 for how much the fitness is allowed to drop with the metabolite target.
  • allow_modulation (bool) – If False does not allow modulation targets (MILP and QP are not compatible)
Returns:

A data frame with the possible replacements.
Return type:

pandas.DataFrame
marsi.cobra.strain_design.post_processing.convert_target(model, target, essential_metabolites, ignore_transport=True, ignore_metabolites=None, allow_accumulation=True, reference=None, allow_modulation=True)[source]

Generates a dictionary {species_id -> (MetaboliteKnockoutTarget, MetaboliteModulationTarget)}.

Parameters:
  • model (cobra.Model) – A COBRA model
  • target (ReactionModulationTarget, ReactionKnockoutTarget) – The flux from the reference state (0 if unknown)
  • ignore_metabolites (list) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc.)
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • reference (dict) – A dictionary containing the flux values of a reference flux distribution.
  • essential_metabolites (list) – A list of essential metabolites
  • allow_modulation (bool) – If False does not allow modulation targets (MILP and QP are not compatible).
Returns:
Return type:

dict
marsi.cobra.strain_design.post_processing.convert_target_group(model, target_group, essential_metabolites, ignore_transport=True, ignore_metabolites=None, allow_accumulation=True, reference=None, allow_modulation=True)[source]

Generates a dictionary {species_id -> MetaboliteKnockoutTarget}.

Parameters:
  • model (cobra.Model) – A COBRA model
  • target_group (dict) – A dictionary of (target: relative coefficient)
  • ignore_metabolites (set) – A list of metabolites that should not be targeted (essential metabolites, currency metabolites, etc.)
  • ignore_transport (bool) – If False, also knockout the transport reactions.
  • allow_accumulation (bool) – If True, create an exchange reaction (unless already exists) to simulate accumulation of the metabolites.
  • reference (dict) – A dictionary containing the flux values of a reference flux distribution.
  • essential_metabolites (set) – A list of essential metabolites.
  • allow_modulation (bool) – If False does not allow modulation targets (MILP and QP are not compatible).
Returns:

A dictionary with the antimetabolites to test and another dictionary with the reaction where those metabolites belong.
Return type:

tuple

marsi.cobra.strain_design.target module

class marsi.cobra.strain_design.target.AntiMetaboliteManipulationTarget(species_id, fraction=0.5, ignore_transport=True, allow_accumulation=True, accession_id=None, accession_db=None)[source]

Bases: cameo.core.target.Target

Metabolite target used to manipulate fluxes. There are three modes of action: 1. Inhibition 2. Competition 3. Knockout (see Metabolite Knockout Target)

If the metabolite is essential, the fluxes around the metabolite will be increased for the cells to survive. If the metabolite is not essential, the fluxes will be decreased as a consequence of the metabolite presence.

Methods

apply(model[, reference])
get_model_target(model) Finds the metabolites in the model across compartments.
to_gnomic()
get_model_target(model)[source]

Finds the metabolites in the model across compartments.

Parameters:model (cameo.SolverBasedModel) – A constraint-based model.
Returns:A list of cobra.core.metabolite.Metabolite.
Return type:list
apply(model, reference=None)[source]
to_gnomic()[source]
class marsi.cobra.strain_design.target.MetaboliteKnockoutTarget(species_id, ignore_transport=True, allow_accumulation=True)[source]

Bases: marsi.cobra.strain_design.target.AntiMetaboliteManipulationTarget

Methods

apply(model[, reference])
get_model_target(model) Finds the metabolites in the model across compartments.
to_gnomic()
apply(model, reference=None)[source]
to_gnomic()[source]

Module contents

class marsi.cobra.strain_design.OptMet(model, evolutionary_algorithm=<class 'inspyred.ec.ec.GA'>, essential_metabolites=None, plot=True, *args, **kwargs)[source]

Bases: cameo.core.strain_design.StrainDesignMethod

Attributes

manipulation_type
plot

Methods

__call__(*args, **kwargs)
run([target, biomass, substrate, …])
param target:The design target
manipulation_type
plot
run(target=None, biomass=None, substrate=None, max_knockouts=5, variable_size=True, simulation_method=<function fba>, growth_coupled=False, max_evaluations=20000, population_size=200, max_results=50, seed=None, **kwargs)[source]
Parameters:
  • target (str, Metabolite or Reaction) – The design target
  • biomass (str, Metabolite or Reaction) – The biomass definition in the model
  • substrate (str, Metabolite or Reaction) – The main carbon source
  • max_knockouts (int) – Max number of knockouts allowed
  • variable_size (bool) – If true, all candidates have the same size. Otherwise the candidate size can be from 1 to max_knockouts.
  • simulation_method (function) – Any method from cameo.flux_analysis.simulation or equivalent
  • growth_coupled (bool) – If true will use the minimum flux rate to compute the fitness
  • max_evaluations (int) – Number of evaluations before stop
  • population_size (int) – Number of individuals in each generation
  • max_results (int) – Max number of different designs to return if found.
  • kwargs (dict) – Arguments for the simulation method.
  • seed (int) – A seed for random.
Returns:
Return type:

OptMetResult
class marsi.cobra.strain_design.RandomMutagenesisDesign(model=None, nearest_neighbors_model=None, min_tanimoto=0.75, currency_metabolites=None, essential_metabolites=None)[source]

Bases: marsi.cobra.strain_design.design_methods.GenericMARSIDesignMethod

Apply only knockout like designs where total loss of functions are expected.

Methods

essential_metabolites_reactions()
optimize_with_metabolites(target[, …])
optimize_with_reaction(target[, …])
optimize_with_reaction(target, max_interventions=1, substrate=None, biomass=None, design_method='optgene', max_results=100, non_essential_metabolites=False, max_evaluations=20000, **designer_kwargs)[source]
optimize_with_metabolites(target, max_interventions=1, substrate=None, biomass=None, max_results=100, max_evaluations=20000, **design_kwargs)[source]
class marsi.cobra.strain_design.ALEDesign(model=None, nearest_neighbors_model=None, min_tanimoto=0.75, currency_metabolites=None, essential_metabolites=None)[source]

Bases: marsi.cobra.strain_design.design_methods.GenericMARSIDesignMethod

Apply both knockout and flux modulation. The strains will be selected via growth rate proxy.

Methods

essential_metabolites_reactions()
optimize_with_metabolites(target[, …])
optimize_with_reaction(target[, …])
optimize_with_reaction(target, max_interventions=1, substrate=None, biomass=None, design_method='differential_fva', max_results=100, non_essential_metabolites=False, **designer_kwargs)[source]
optimize_with_metabolites(target, max_interventions=1, substrate=None, biomass=None, max_results=100, max_evaluations=20000, **design_kwargs)[source]