mfapy package

Submodules

mfapy.carbonsource module

carbonsource.py:CarbonSource class in mfapy

The module includes:

CarbonSource class

Todo

  • Subtraction of natural isotope

class mfapy.carbonsource.CarbonSource(carbon_sources, emu_dict)

Bases: object

Class for carbon source information.

Instance of this class is generated in the MetabolicModel instance

generate_carbonsource_MDV(carbonsource=[], correction='no')

Generator of MDVs of all EMUs of each carbon source.

This function is called in the mfapy.metabolicmodel.MetabolicModel

Parameters:

  • carbonsource (array) – List of carbon source metabolite (optional)

  • correction (str) – (yes/no) Correction of isotopomer distribution considering natural 13C occurence

Returns:

Dictionary of mdv data of carbon source.

Return type:

dict

Examples

>>> cs.generate_carbonsource_MDV()

History:

Correcton of IDV by natural 13C method was improved.

labeled by 2 13C and 3 13C was taken into consideration.

generate_dict()

Generator of a dictionary of MDVs of all EMUs of carbon sources

Parameters:

required. (Not) –

Returns:

Dictionary of MDVs of all EMUs of carbon sources

Return type:

dict

Examples

>>> mdvs = cs.generate_dict()
set_all_isotopomers(compound, list, correction='no')

Setter of IDV data by list of all mass isotopomer distribution

Parameters:

  • compound (str) – Name of carbon source.

  • list (array) – list of mass isotopomer distribution (from 000, 001, 010, to 111)

  • correction (str) – (yes/no) Correction of isotopomer distribution considering natural 13C occurence

Returns:

True/False

Return type:

Booleans

Examples

>>> cs.set_all_isotopomers('AcCoA', [0.3, 0.3, 0.3, 0.1])
set_carbonsources(filename, correction='no', format='text', output='normal')

Setter of isotopomer data of multiple carbon sourses from text file.

Parameters:

  • filename (str) –

    filename of MDV data with following format:

    Name    Isotopomer      Ratio
Asp     #0000   0.5
Asp     #1111   0.5
AcCoA   #00     0.5
AcCoA   #11     0.5

  • correction (str) – (yes/no) Correction of isotopomer distribution considering natural 13C occurence

  • format (str) – “text” (defalut) or “csv”

  • output (str) – “normal” (defalut) or “debug”

Returns:

True/False

Return type:

Boolean

Examples

>>> cs2.set_isotopomers_from_file('Example_1_carbonsource2.txt', correction = "yes")
set_each_isotopomer(compound, dict, correction='no')

Setter of IDV data of selected mass isotopomers

Parameters:

  • compound (str) – Name of carbon source

  • dict (dict) –

    Dictionary of mass isotopomer and its relative abundance:

    {'#111': 0.5, '#001': 0.5}

  • correction (str) – (yes/no) Correction of isotopomer distribution considering natural 13C occurence

Returns:

True/False

Return type:

Boolean

Examples

>>> cs.set_each_isotopomers('AcCoA', {'#11':0.5, '#10':0.25}, correction = 'yes')
set_labeled_compounds(compound, dict, correction='no')

Setter of IDV data by distribution of selected mass isotopomers

Following symbols are available:

"[13C]CO2": "#1",
"[12C]CO2": "#0",
"[13C]THF": "#1",
"[12C]THF": "#0",
"[1-13C]glucose": "#100000",
"[2-13C]glucose": "#010000",
"[1,2-13C]glucose": "#110000",
"[U-13C]glucose": "#111111",
"[1-13C]glutamine": "#10000",
"[2-13C]glutamine": "#01000",
"[5-13C]glutamine": "#00001",
"[U-13C]glutamine": "#11111",
Parameters:

  • compound (str) – Name of carbon source.

  • dict (dict) – Dictionary of mass isotopomer and its relative abundance {‘#111’: 0.5, ‘#001’: 0.5}

  • correction (str) – (yes/no) Correction of isotopomer distribution considering natural 13C occurence

Returns:

True/False

Return type:

Boolean

Examples

>>> cs.set_labeled_compounds('Glc',{'[1_13C]glucose': 0.5, '[U_13C]glucose':0.5}, correction = 'yes')
show()

Method to display contents of CarbonSource instance

Parameters:

required. (Not) –

Returns:

Nothing.

Examples

>>> cs.show()
Name: Asp
Carbon number: 4
#0000       1.0
Name: AcCoA
Carbon number: 2
#00 0.5
#10 0.25
#11 0.25
mfapy.carbonsource.main()

mfapy.mdv module

mdv.py:MdvData class in mfapy

The module includes:

MdvData class
MdvTimeCourseData class

Todo

  • I/O

mfapy.mdv.INV_correcting(formula)
class mfapy.mdv.MdvData(target_fragments)

Bases: object

Class of mass distribution vector (MDV) information. A instance of the MdvData class could be generate from an instance of MetabolicModel class using model.generate_mdv_templete()

add_gaussian_noise(stdev, iteration=1, method='absolute', normalize='on')

Addition of gausian noise to MDV data

Parameters:

  • stdev (float) – noise level (standard deviation of normal distribution)

  • iteration (int) – number of sampleing (experimental)

  • method (str) –

    method to add gausian noise

    • ”relative” intentsity = (randn() * stdev + 1) * original_intensity

    • ”absolute” (default) intentsity = randn() * stdev) + original_intensity

  • normalize (str) – on (default)/off sum of ratio is set to 1.0

Returns:

True/False

Return type:

boolean

Examples

>>> mdv.add_gaussian_noise(0.01)
add_natural_isotope()

Addition of natural isotope effect to fragments

Molecular formula information in “target fragment” of model file is used.

Parameters:

required. (Not) –

Returns:

Nothing.

Example

>>> mdv.add_natural_isotope()
check(output='debug')

Function to check missing values in the MDV data before MDV comparison

Parameters:

output (str) – Show details when “debug” mode

Returns:

True/False

Return type:

boolean

Examples

>>> mdv.check()
comparison_with_another_mdv(mdv)

Comparison of two MDV data

Parameters:

mdv – instanse of another MdvData class for comparison

Returns:

Nothing

Examples

>>> mdv.compare_mdv(mdv_fitted)
correct_natural_isotope(mode='normal')

Subtraction of natural isotope effect from fragments

Molecular formula information of “target fragment” in the model definition file is used.

Parameters:

mode (str) – “normal”(defalut) and “correction” (forced removal of negatives value and sum(ratio) is set at 1.0)

Returns:

Nothing

Examples

>>> mdv.correct_natural_isotope()
generate_observed_mdv()

Generator of MDV related inforamtion in list format.

This function is onely used in model.set_experiment()

Parameters:

required. (Nor) –

Returns:

  • id_array (array) List of ids

  • ratio_array (array) List of ratio data

  • std_array (array) List of std data

  • use_array (array) List of use data

  • observed_fragments (array) List of observed fragments

  • data (array) List of other information

Examples

>>> id_array, ratio_array, std_array, use_array, observed_fragments, data = mdv.generate_observed_mdv()
get_data(fragment, number)

Getter of [number] th isotopomer in the MDV of [fragment].

Parameters:

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

ratio (float) Relative abundance of mass spectra data

std (float) Standard deviation of measured MDV data

use (str) Data to be used for MDV comparison. ‘use’ or ‘no’

Examples

>>> ratio, std, use = mdv.get_data('Leu85', 2)
get_fragment_mdv(fragment)

Getter of MDV data of [fragment].

Parameters:

fragment (str) – name of target_fragment

Returns:

mdv_pattern (array) Array of MDV of fragment

Examples

>>> mdv_pattern = mdv.get_fragment_mdv('Leu85')
get_fragments_for_mdv_calculation()

Getter of fragment list whose MDVs are calculated in calmdv fucntion

Parameters:

required. (Not) –

Returns:

List of fragment names calculated in calmdv fucntion

Return type:

array

Examples

>>> list_of_fragment = mdv.get_fragments_for_mdv_calculation()
get_number_of_measurement()

Getter of number of measurement of the MDV data set.

Parameters:

required. (Nor) –

Returns:

Number of measurement

Return type:

int

Examples

>>> number_of_measurement = mdv.get_number_of_measurement()
has_data(fragment, number)

Checker whether the MdvData instance has data of [number] th isotopomer in the MDV of [fragment].

Parameters:

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

TrueFalse

Return type:

Boolean

Examples

>>> if mdv.has_data('Leu85', 2): brabra
load(filename, format='text', output='normal')

Method to load MDV data from text/csv file

Parameters:

  • filename (str) –

    filename of MDV data with the format:

    Name    Spectrum        Select  MDV     Std
Ala57   m0      1       0.566990778     0.000774686
Ala57   m1      1       0.148623963     0.000774686
Ala57   m2      1       0.039467636     0.000774686
Ala57   m3      1       0.244917622     0.000774686

  • format (str) –

    file format:

    • ’csv’ : CSV.

    • ’text’ (defalut) : tab-deliminated text.

  • output – “normal” (defalut) or “debug”

Returns:

True/False

Return type:

Boolean

Examples

>>> mdv.load('filename', format = 'text',output = "normal")
save(filename, format='text')

Method to save MDV data in text/csv file

Parameters:

  • filename (str) –

    filename of MDV data with the format:

    Name    Spectrum        Select  MDV     Std
Ala57   m0      1       0.566990778     0.000774686
Ala57   m1      1       0.148623963     0.000774686
Ala57   m2      1       0.039467636     0.000774686
Ala57   m3      1       0.244917622     0.000774686

  • format (str) –

    file format:

    • ’csv’ : CSV.

    • ’text’ (defalut) : tab-deliminated text.

Returns:

True/False

Return type:

Boolean

Examples

>>> mdv.save('filename', format = "csv")
set_data(fragment, number, ratio, std, use)

Setter of [number] th isotopomer in the MDV of [fragment].

Parameters:

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

  • ratio (float) – Relative abundance of mass spectra data

  • std (float) – Standard deviation of measured MDV data

  • use (str) – Data to be used for MDV comparison. ‘use’ or ‘no’

Returns:

TrueFalse

Return type:

Boolean

Examples

>>> mdv.set_data('Leu85', 2, 0.763, 0.0054, 'use')
set_mdv_for_comparison(fragment, number)

Setter of mass data to be used for MDV comparison

Parameters:

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

Nothing.

Examples

>>> mdv.set_mdv_for_comparison(fragment, number)
set_mdvs_for_comparison(threshold_ratio, threshold_std=1.0)

Setters of multiple data to be used for MDV comparison

Parameters:

  • threshold_ratios (float) – MDV data whose ratio (0.0-1.0) is above threshold_ratios are used for MDV comparison

  • threshold_std (float) – MDV data whose std is below threshold_std are used for MDV comparison

Returns:

True/False

Return type:

boolean

Examples

>>> mdv.set_mdvs_for_comparison(0.05, 0.005)
set_observed_fragments(fragments)

Setter of observed fragment.

When new MdvData instance is generated ‘observed_fragment’ == ‘target_fragment’.

Parameters:

fragments (Array) – List of fragment names calculated in calmdv fucntion

Returns:

Nothing.

Examples

>>> mdv.set_observed_fragments(['Phe_85', 'Ala_57'])
set_std(value, method='absolute')

Setter of standard deviation level

Parameters:

  • value (float) – level of standard deviation.

  • method (str) –

    Method to calculate stdev levels.

    • ’relative’: Levels are set by stdev = [value] * signal intensity.

    • ’absolute’ (detault) : Levels are set by stdev = [value].

Returns:

Nothing.

Examples

>>> mdv.set_std(0.05, method = 'absolute')
set_unused_mdv_for_comparison(fragment, number)

Setter of mass data to be ignored for MDV comparison

Parameters:

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

Nothing.

Examples

>>> mdv.set_unused_mdv_for_comparison(fragment, number)
class mfapy.mdv.MdvTimeCourseData

Bases: object

add_gaussian_noise(stdev, iteration, method='absolute', normalize='on')

Addition of gausian noise to MDV data

Parameters:

  • stdev (float) – noise level (standard deviation of normal distribution)

  • iteration (int) – number of sampleing (experimental)

  • method (str) –

    method to add gausian noise

    • ”relative” intentsity = (randn() * stdev + 1) * original_intensity

    • ”absolute” (default) intentsity = randn() * stdev) + original_intensity

  • normalize (str) – on (default)/off sum of ratio is set to 1.0

Returns:

True/False

Return type:

boolean

Examples

>>> mdv.add_gaussian_noise(0.01)
add_time_point(time, mdv_instance)

Addition of new time point

Parameters:

  • time (float) – time point

  • mdv_instance (MdvData) – MdvData instance

Returns:

nothing

Examples

>>> mdv.add_time_point(3.0, mdv3)
generate_observed_mdv()

Generator of MDV related inforamtion in list format.

This function is onely used in model.set_experiment()

Parameters:

required. (Nor) –

Returns:

  • id_array (array) Array of list of ids at each time point

  • ratio_array (array) Array of list of ratio data at each time point

  • std_array (array) Array of list of std data at each time point

  • use_array (array) Array of list of use data at each time point

  • observed_fragments (array) Array of list of observed fragments at each time point

  • data (array) Array of list of other information at each time point

Examples

>>> id_array, ratio_array, std_array, use_array, observed_fragments, data = mdv.generate_observed_mdv()
get_data(time, fragment, number)

Getter of [number] th isotopomer in the MDV of [fragment] at [time].

Parameters:

  • time (float) – time point

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

ratio (float) Relative abundance of mass spectra data

std (float) Standard deviation of measured MDV data

use (str) Data to be used for MDV comparison. ‘use’ or ‘no’

Examples

>>> ratio, std, use = mdv.get_data('Leu85', 2)
get_number_of_measurement()

Getter of number of measurement of the MDV data set.

Parameters:

required. (Nor) –

Returns:

Number of measurement

Return type:

int

Examples

>>> number_of_measurement = mdv.get_number_of_measurement()
History

Revised at 27/5/2018

get_timecourse_data(fragment, number)

Getter of time course MDV data

Examples

>>> tc = mdv.get_timecourse_data('Leu85', 2)
Parameters:

  • fragment (name of target_fragment) –

  • number (Number of isotope) –

  • Reterns –

  • ---------- –

  • tc (Array of time course of Mass isotopomer) –

get_timepoints()

Getter of Time points

Parameters:

  • Reterns –

  • ---------- –

  • timepoints (list of time points) –

Examples

>>> timepoints = mdv.get_timepoints()
>>> timepoints
>>> [0., 1., 5,]
has_data(time, fragment, number)

Checker whether the MdvData instance has data of [number] th isotopomer in the MDV of [fragment] at [time].

Parameters:

  • time (float) – time point

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

Returns:

TrueFalse

Return type:

Boolean

Examples

>>> if mdv.has_data(0, 'Leu85', 2): brabra
save(filename, format='text')

Method to save MDV data in text/csv file

Files of each timepoint are generated with names filename+”timepoint”.txt

Parameters:

  • filename (str) –

    filename of MDV data with the format:

    Name    Spectrum        Select  MDV     Std
Ala57   m0      1       0.566990778     0.000774686
Ala57   m1      1       0.148623963     0.000774686
Ala57   m2      1       0.039467636     0.000774686
Ala57   m3      1       0.244917622     0.000774686

  • format (str) –

    file format:

    • ’csv’ : CSV.

    • ’text’ (defalut) : tab-deliminated text.

Returns:

True/False

Return type:

Boolean

Examples

>>> mdv.save('filename', format = "csv")
set_data(time, fragment, number, ratio, std, use)

Setter of [number] th isotopomer in the MDV of [fragment] at [time].

Parameters:

  • time (float) – time point

  • fragment (str) – name of target_fragment

  • number (int) – [number] th isotopomer

  • ratio (float) – Relative abundance of mass spectra data

  • std (float) – Standard deviation of measured MDV data

  • use (str) – Data to be used for MDV comparison. ‘use’ or ‘no’

Returns:

TrueFalse

Return type:

Boolean

Examples

>>> mdv.set_data(0, 'Leu85', 2, 0.763, 0.0054, 'use')
set_mdvs_for_comparison(threshold_ratio, threshold_std=1.0)

Setters of multiple data to be used for MDV comparison

Parameters:

  • threshold_ratios (float) – MDV data whose ratio (0.0-1.0) is above threshold_ratios are used for MDV comparison

  • threshold_std (float) – MDV data whose std is below threshold_std are used for MDV comparison

Returns:

True/False

Return type:

boolean

Examples

>>> mdv.set_mdvs_for_comparison(0.05, 0.005)
set_observed_fragments(fragments)

Setter of observed fragment.

When new MdvData instance is generated ‘observed_fragment’ == ‘target_fragment’.

Parameters:

fragments (Array) – List of fragment names calculated in calmdv fucntion

Returns:

Nothing.

Examples

>>> mdv.set_observed_fragments(['Phe_85', 'Ala_57'])
set_std(value, method='absolute')

Set standard deviation levels from mass sepctral intensity

Parameters:

method (Method to calculate stdev levels:) – ‘relative’ (detault): Levels are set by stdev = [value] * signal intensity ‘abusolute’: Levels are set by stdev = [value]

Examples

>>> mdv.set_std(0.05, method = 'absolute')
mfapy.mdv.binomial(n, k)
mfapy.mdv.count_atom_number(formula)
mfapy.mdv.transition_matrix(formula)

mfapy.metabolicmodel module

metabolicmodel.py:MetabolicModel class in mfapy

This is a core module of mfapy.

The module includes:

MetabolicModel class

Todo

  • Rewriting model construction part.

  • Check callbacklevel.

class mfapy.metabolicmodel.MetabolicModel(reactions, reversible, metabolites, target_fragments, mode='normal')

Bases: object

Class of MetabolicModel

An instances of this class has fuctions to analyze mass spectrometry based 13C MFA and INST MFA data.

An metabolicModel instance generates MDVData and CarbonSouece classes

Returns:

instance of metabolic model

Return type:

instance

Examples

>>> reactions, reversible, metabolites, target_fragments = mfapy.mfapyio.load_metabolic_model("example_1_toymodel_model.txt", format = "text")
>>> model = mfapy.metabolicmodel.MetabolicModel(reactions, reversible, metabolites, target_fragments)

Attributes (incomplete):

self.reactions = reactions
self.metabolites = metabolites
self.reversible = reversible
self.target_fragments = target_fragments
self.symmetry = {}
self.carbon_source = {}
self.experiments = {}
add_perturbation(flux)

New state dict is generarate from a given state dict by an addition of small random perturbation to an independent flux

Parameters:

state_dict (dict) – Dictionary of metabolic state data.

Returns:

  • check (boolean)

  • perturbed_state (dict)Dictionary of metabolic state data.

  • flux_vector (array) List of metabolic flux, metabolite concentration data.

  • reaction_name_list (array) List of ids

Examples

>>> check, perturbed_state, flux_vector, reaction_name_list = model.add_perturbation(state_dict)
History

Newly created at 21/1/2021

calc_idv(flux, carbon_sources)

Calc IDV from given flux and carbon_sources.

Isotopomer distribution vector (IDV) was used to calc MDV at time = 0 in the INST mode.

Examples

>>> idv = model.calc_idv(flux, carbon_sources)
Parameters:

  • flux (dict) – Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • carbon_sources (instance) – Instance of carbon source object

Returns:

array of idv

Return type:

array

See also

set_experiment() generate_mdv()

calc_rss(fluxes, mode='flux')

Getter to calc residual sum of square (RSS) level between estimated MDVs of ‘flux’ and measured MDVs in ‘experiment(s)’

Parameters:

  • flux (dict) – Dictionary of metabolic flux distribution or independent flux vector

  • mode (str) –

    • “independent”:Calculation of RSS from independent flux vector

    • ”flux”:Calculation of RSS from state dictionary

Returns:

Residual sum of square (RSS)

Return type:

float

Examples

>>> rss = model.calc_rss(flux_opt)
>>> print rss
7564.123
History:

Call of calmdv funcion is removed.12/9/2014

check_independents(independents)

Checker method to test whether a given vector of independent fluxes is able to produce feasible status (flux) vector within lower and upper boundaries.

Parameters:

independents (array) – list of independent fluxes generated by self.get_independents

Returns:

  • check (Boolean) True means that a given vector of independent fluxes is able to produce feasible status (flux) vector within lower and upper boundaries.

  • flux_vector (array) List of metabolic flux, metabolite concentration data.

  • reaction_name (array) List of ids

Examples

>>> check, flux_vector, reaction_name = self.check_independents(independents)

History:

Newly created at 21/1/2021

clear_experiment()

Clear all ‘experiment set(s)’ in the metabolic model.

Parameters:

required (Not) –

Examples

>>> model.clear_experiment()

See also

model.set_experiment()

History:

Newly developed at 4/9/2014

fitting_flux(method='SLSQP', flux=[], output='result')

Method for fitting a metabolic model to multiple experiments

Used parameters in self.configuration().

  • ‘iteration_max’: Maximal number of interation in the optimizers. Example: model.set_configuration(iteration_max = 1000)

  • ‘number_of_repeat’: Number of repeated execution by ‘deep’ functions. Example: model.set_configuration(number_of_repeat = 2)

  • ‘ncpus’: Number of CPUs used in parallel processing. Example: model.set_configuration(ncpus = 16)

Parameters:

  • method (str) –

    Algorism used for optimization.

    • ’deep’: Repeated execution of SLSQP & LN_PRAXIS

    • ’SLSQP’: Sequential Least SQuares Programming algorithm implemented in scipy

    • ’COBYLA’: Constrained Optimization By Linear Approximation implemented by scipy

    • ”LN_BOBYQA”:The BOBYQA algorithm for bound constrained optimization without derivatives by nlopt

    • ”LN_SBPLX”: Sbplx (based on Subplex, (a variant of Nelder-Mead that uses Nelder-Mead on a sequence of subspaces) by nlopt

    • ”LN_NEWUOA”: The NEWUOA software for unconstrained optimization without derivatives by nlopt

    • ”LN_COBYLA”: Constrained Optimization By Linear Approximation implemented by nlopt

    • ”LN_PRAXIS”: Gradient-free local optimization via the “principal-axis method” implemented by nlopt

    • ”LN_NELDERMEAD”: the original Nelder-Mead simplex algorithm by nlopt

    • ”GN_ESCH”: Evolutionary Algorithm for global optimization by nlopt

    • ”GN_CRS2_LM”: Controlled random searchimplemented for global optimizatoin by nlopt

    • ”GN_DIRECT_L”: DIviding RECTangles algorithm for global optimization by nlopt

    • ”GN_IRES”: Improved Stochastic Ranking Evolution Strategy by nlopt

    • ”GN_AGS”: AGS NLP solver by nlopt

  • flux (dict) –

    Initial flux distribution generated by self.generate_initial_states():

    • When flux is a dict of one state, single fitting trial is executed.

    • When flux is a array of multiple dists of states, muptiple fitting trial is exected by using the parallel processing.

  • output (str) –

    Output method.

    • ’result’ (default): Fitting problem is solved inside of the function.

    • ’for_parallel’: Fitting problems is NOT solved inside of the function. In this mode, a tupple of parameters for fit_r_mdv_pyopt() functions are generated for parallel processinng.

Returns:

stop condition

  • [number_of_initial_states == 1 ] Str of stop condition.

  • [number_of_initial_states > 1 ] Array of str of stop condition.

  • flux_opt (array or dict): Initial flux distribution generated by self.generate_initial_states()

    • [number_of_initial_states == 1 ] dict of optimzed metabolic state.

    • [number_of_initial_states > 1 ] Array of dict of optimzed metabolic state. Sorted by ascending order of their RSS levels

  • RSS_bestfit (array or float) : Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

    • [number_of_initial_states == 1 ] Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

    • [number_of_initial_states > 1 ] Array of dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

Return type:

  • state (str or array)

Examples:

#
# Single fitting trial
#
>>> state, flux_initial = model.generate_initial_states(10, 1)
>>> state, RSS_bestfit, flux_opt = model.fitting_flux(method = 'deep', flux = flux_initial)
>>> results= [("deep", flux_opt)]
>>> model.show_results(results, pool_size = "off")
#
# Multipe fitting trials by using parallel python
#
>>> state, flux_initial = model.generate_initial_states(50, 4)
>>> state, RSS_bestfit, flux_opt_slsqp = model.fitting_flux(method = "SLSQP", flux = flux_initial)
>>> results= [("deep", flux_opt[0])]
>>> model.show_results(results, pool_size = "off")
#
# Generating parameters for parallel processing
#
>>> parameters = model.fitting_flux(method = "SLSQP", flux = flux_initial, output = "for_parallel")

See also

generate_initial_states()

generate_calmdv(mode='normal')

Constructer of the python function to calculate MDV of target fragments from a metabolic flux distribution.

This function is performed in reconstruct()

Parameters:

mode (str) – optimize EMU network reduction (experimental)

Returns:

string, cython_string (str) Texts of python code describing calmdv and diffmdv functions for python 3 and Cython (Experimental)

Examples

>>> calmdv_text, ccalmdv_text = self.generate_calmdv(mode)
generate_carbon_source_template()

Generator of a templete of CarbonSourse instance. Labeling information of carbon sources will be added to the instance.

Carbon source compounds are derived from the metabolic model (//Metabolites)

Parameters:

required. (Not) –

Examples

>>> carbon_source_idv = model.generate_carbon_source_templete()

See also

CarbonSource.py

generate_carbon_source_templete()
generate_ci_template(targets='normal')

Generator of a template dictionary to keep confidence interval search results.

The templatedisctionary generated by this method is used for:

  • Setting reactions, metabolites, and reversible reactions for CI searching in self.search_ci method.

  • Store all record of CI searching by self.search_ci method.

Parameters:

targets (str) –

  • ‘normal’ Free reversible and inreversible reactions

  • ’independent’ Independent reactions

  • ’all’ all reactions

  • ’without_reversible_reactions’ inreversible reactions

Returns:

template dictionary

Return type:

dict

Examples

>>> ci = model.generate_ci_templete(targets = [("reaction","r29_g6pdh"),("reaction","r27_pc"),("reaction","r28_mae")])
>>> ci = model.generate_ci_templete(targets = 'normal')

See also

search_ci.

generate_ci_templete(targets='normal')
generate_initial_states(iterations=100, initial_states=1, method='normal', template=[])

Generator of random initial metabolic states for model fitting

Initial metabolic states are randomly generated for “iterations” times from which better states with lower RSS (number_of_initial_states) were selected.

Used parameters in self.configuration().

  • ‘initial_search_iteration_max’: Maximal number of interation in the optimizers. Example: model.set_configuration(initial_search_iteration_max = 1000)

  • ‘ncpus’: Number of CPUs used in parallel processing. Example: model.set_configuration(ncpus = 16)

Parameters:

  • method (str) – computational method * ‘normal’ : Generation of flux distributions by single thread. * ‘parallel’ : Parallel generation of flux distributions using parallel processing.

  • iterations (int) – Number of trials to generate initial metabolic state. A trial sometimes failed due to a bad optimization state.

  • initial_states (int) – Number of initial metabolic state(s) with lower RSS to be generated.

  • template (array) – Array of templete metabolic state(s) dictionaries. When template is available, initial metabolic state(s) similiar to the templete metabolic state are generated. This function is used in the grid search fucntion.

Returns:

Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • [number_of_initial_states > 1 ] Array of dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • [number_of_initial_states == 1 ] Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • state (str or array)stop condition

    • [number_of_initial_states > 1 ] Str of stop condition.

    • [number_of_initial_states == 1 ] Array of Str of stop condition.

Return type:

  • flux (array or dict)

Examples

>>> state, dict = model.generate_initial_states(50,1)
>>> state, dict = model.generate_initial_states(50, 2, template = flux)

See also

Fitting

generate_mdv(flux, carbon_sources, timepoint=[], startidv=[])

Generator of a MdvData instance including MDV data generated from given flux and carbon sources.

Parameters:

  • flux (dict) – Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • carbon_sources (instance) – Instance of CarbonSource

  • timepoint (array) – For INST-13C MFA. An array of time points of measurement in MDV

  • startidv (array) – array of idv as starting isotope distribution for INST

Returns:

instance* MdvData instance

Examples

>>> mdv = model.generate_mdv(flux, mdv_carbon_sources)

See also

generate_carbonsource_MDV

History:

140912 calc_MDV_from_flux instead of calmdv is used.

generate_state(template=[])

Generator of a random metabolic flux distribution.

Used parameters in self.configuration().

  • ‘initial_search_iteration_max’: Maximal number of interation in the optimizers. Example: model.set_configuration(initial_search_iteration_max = 1000)

Parameters:

template (dict) – Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels. When template is available, metabolic state most similar to the template is generated. The function is used in the grid search.

Returns:

  • flux (dict) Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

  • independent_flux (array) List of independent flux

Examples

>>> flux, independent_flux = model.generate_intial_flux_distribution()
generate_state_dict(tmp_r)

Generator of a state dict from a given vector.

Parameters:

tmp_r (array) – tmp_r = numpy.dot(matrixinv, Rm_intial)

Returns:

Dictionary of metabolic state inclucing metabolic flux and metabolite pool size levels.

Return type:

dict

Examples

>>> state_dict = model.generate_state_dict(tmp_r)
get_constrain(group, id)

Getter of constrains of metabolic reaction.

Parameters:

  • group (str) – group of parameters (“reaction”, “reversible”, “metabolite”)

  • id (str) – id in model definition

Returns:

  • type (str) type of constrain “fitting”, “fixed”, “free”, “pseudo”

  • value (float) Level of metabolic flux

  • stdev (float) Standard deviation of metabolic flux level used for ‘fitting’ type reactions.

Examples

>>> type, value, stdev = model.get_constrain("reaction","pgi")

See also

set_constrain

get_degree_of_freedom()

Getter of degree of freedom of the model

Examples

>>> degree_of_freedom = model.get_degree_of_freedom()
Returns:

degree of freedom.

Return type:

int

get_independents(flux)

Getter of independent status (flux) vector, its lower and upper boundaries and a list of independent flux ids of given state (flux) dict.

Examples

>>> independent_vector, lb, ub, independent_list = model.get_independents(state_dict)
Parameters:

flux (dict) – Dictionary of metabolic state data.

Returns:

  • independent_vector (array) list of independent flux values

  • lb (array) list of independent flux values

  • ub (array) list of independent flux values

  • independent_list (array) list of independent fluxes, metabolites, and reversible reactions.

History:

Newly created at 21/1/2021

get_number_of_independent_measurements()

Getter of number of independent measurements of the model

Returns:

number of independent measurements.

Return type:

int

History:

29/9/2014 Modified to consider “fitting” flux information.

Example

>>> number_of_independent_measurements = model.get_number_of_independent_measurements()
get_thres_confidence_interval(flux, alpha=0.05, dist='F_dist')

Getter to obtain a threshold level of RSS for searching confidence interval

Parameters:

  • flux (dict) – Dictionary of best fitted flux distribution.

  • alpha (float) – confidence level. For 95% confience interval, please set alpha = 0.05

  • dist (str) – Probability distribution for test * ‘F-dist’ (default): Using F-distribution * ‘chai-dist’:Using Chai-square distribution

Returns:

  • thres (float) a threshold level of RSS for searching confidence interval

  • number_of_measurements (int) number of measurements

  • degree_of_freedom (int) degree of freedom

Examples

>>> thres, number_of_reactions, degree_of_freedom = model.get_thres_confidence_interval(RSS, alpha = 0.05, dist = 'F_dist')
goodness_of_fit(flux, alpha=0.05)

Getter to calculate goodness-of-fit of a given flux distribution

Parameters:

  • flux (dict) – Dictionary of metabolic flux distribution

  • alpha (float) – Confidence level. Default is 0.05.

Returns:

  • pvalue (float) p-value determined by chi-squere test.

  • rss_thres (float) Threshold RSS at the alpha levels.

Examples

>>> pvalue, rss_thres= model.goodness_of_fit(flux_opt_fitted, alpha = 0.05)
load_mdv_data(filename, format='text', output='normal')

Load MDV data from the text file. This function generate new instance of mfapy.mdv class from a instance of mfapy.model.

Parameters:

  • filename (str) –

    filename of MDV data with following format:

    Name    Spectrum        Select  MDV     Std
Ala57   m0      1       0.566990778     0.000774686
Ala57   m1      1       0.148623963     0.000774686
Ala57   m2      1       0.039467636     0.000774686
Ala57   m3      1       0.244917622     0.000774686

  • format (str) – “text” (defalut) or “csv”

  • output (str) – “normal” (defalut) or “debug”

Returns:

True/False

Return type:

Boolean

Examples

>>> mdv = load_mdv_data('filename')
load_states(filename, format='text')

Load a text/csv file with ‘reacton type’ information to generate new states dict.

Parameters:

  • filename (str) –

    filename of flux data with following format:

    type    Id      type    flux_calue      flux_std        lb      ub
reaction        v1      fixed   100     1       0.001   100
reaction        v2      free    100     1       0.001   100
reaction        v3      free    50      1       0.001   100
metabolites     AcCoA   fixed   1       1       0.001   100
metabolites     OAC     fixed   1       1       0.001   100
metabolites     OACs    fixed   1       1       0.001   100
metabolites     OACx    fixed   1       1       0.001   100
reversible      FUM     free    1       1       0.001   100

  • format – ‘csv’ CSV or ‘text’tab-deliminated text.

Returns:

Dictionary of states dict.

Return type:

dict

Examples

>>> states = model.load_states("test.csv", format = 'csv')
modelcheck()

Method to check medel definition data.

Parameters:

nothing –

Returns:

Nothing.

Examples

>>> model.datacheck()

See also

reconstruct

History:

Newly developed for version 059

posterior_distribution(state_dict, number_of_steps=1000000)

Generator of posterior distribution by the Metropolis-Hastings algorism

Parameters:

  • state_dict (dist) – Dictionary of initial metabolic state

  • number_of_steps (int) – Length of Markov-chain

Returns

array: 2d array. Markov-chain of metabolic state vector

Examples

>>> record = model.posterior_distribution(perturbed_state, number_of_steps = numberofsteps)

History:

Newly created at 21/1/2021

reconstruct(mode='no')

Method to reconstruct the metabolic model to generate new stoichiometry matrix and a function to calculate MDV data.

Parameters:

mode (str) – experimental

Returns:

calmdv_text (dict) The dictionary has two keys. “calmdv” and “diffmdv” are instance of functions for 13C-MFA and INST-13C-MFA to generate MDV from a given metabolic flux, respectively.

Examples

>>> model.reconstruct()

See also

update()

save_states(dict, filename, format='text')

Save state dict to a text/csv file with ‘type’ information.

Parameters:

  • filename (str) – filename of MDV data with the format.

  • dict (dict) – dictinary of metabolic state (flux)

  • format – ‘csv’ CSV or ‘text’tab-deliminated text.

Returns:

True/False

Return type:

Boolean

Examples

>>> model.save_states(flux_dict, "test.csv", format = 'csv')
search_ci(ci, flux, method='grid', alpha=0.05, dist='F_dist', outputthres=2.0)

Method to estimate confidence intervals using information in a ‘ci’ dictionary.

Used parameters in self.configuration().

  • ‘grid_search_iterations’ (default = 1): MNumber of trials for model fitting at each point in grid search. Example: model.set_configuration(grid_search_iterations = 1)

  • ‘number_of_repeat’(default = 50): Number of initial metabolic flux disributions generated for each trial in grid search.Best initial metabolic flux disribution with smallest RSS is employed.

  • ‘number_of_repeat’: Number of repeated execution by ‘deep’ functions. Example: model.set_configuration(number_of_repeat = 2)

  • ‘ncpus’: Number of CPUs used in parallel processing. Example: model.set_configuration(ncpus = 16)

Argss:

ci (dict): dictionary for storeing confidence interval results. ci is generated by self.generate_ci_templete()

flux (dict): dict of metabolic state (best fitted)

method (str): ‘grid’ (Grid search method) is available

alpha (float): (default 0.05) Confidence interval level. 0.05 means a 95% confidence interval level

dist (str):

  • ‘F-dist’ (default): Using F-distribution

  • ‘chai-dist’:Using Chai-square distribution

outputthres (float): The grid search results larger than outputthres * threshold rss were removed from output

Returns:

dictionary of confidence interval results.

Return type:

dict

Examples

>>> ci = model.generate_ci_templete(targets = 'independent')
>>> ci = model.search_confidence_interval_parallel(ci, flux, method = 'grid')
set_boundary(group, id, lb, ub)

Setter of a lower and upper boundaries of metablic states

Parameters:

  • group (str) – metabolite, reaction, reversible

  • id (str) – id of state parameter

  • lb (float) – lower boundary

  • ub (float) – upper boundary

Returns:

True/False.

Return type:

Booleans

Examples

>>> model.set_boundary("reaction", 'HEX1', 0.001, 100.0)

See also

set_constrain

set_configuration(*args, **kwargs)

Setter of a configuration of model parameters.

Parameters:

*args (list) – “parameter_name = number”

Returns:

Nothing.

Parameters:

'callbacklevel': default = 0,
    'callbacklevel' determines a frequency level of callbacks from metabolic model.
    # 0: No call back
    # 1: Error report
    # 2: Simple pregress report for grid search
    # 3: Detailed pregress report for grid search
    # 4: Simple pregress report for flux fitting
    # 5: Detailed pregress report for flux fitting
    # 7: Detailed report for model constrution
    # 8: Debug
'iteration_max': default = 1000,
    Maximal number of interations (steps) in each fitting task.
'default_reaction_lb': default = 0.001,
'default_reaction_ub': default = 5000.0,
'default_metabolite_lb': default = 0.001,
'default_metabolite_ub': default = 500.0,
'default_reversible_lb': default = -300,
'default_reversible_ub': default = 300.0,
    Default lower and upper boundaries used only when there is no data in the model definition file.
'grid_search_iterations': default = 1,
    Number of trials for model fitting at each point in grid search.
'initial_search_repeats_in_grid_search': default = 50,
    Number of initial metabolic flux disributions generated for each trial in grid search.
    Best initial metabolic flux disribution with smallest RSS is employed.
'initial_search_iteration_max': default = 1000,
    Maximal number of interations (steps) allowed in each task to find feasible initial metabolic flux distribution.
'add_naturalisotope_in_calmdv':default = "no"
    For INST-13C MFA. Natural isotope is added to MDVs of all intermediate at time = 0.
'number_of_repeat':default = 3,
    For "deep" function of self.fitting flux. "Global" => "Local" optimization methods are repeated for n times.
'ncpus':default = 3,
    Number of CPUs used in parallel processing
'odesolver': default = "scipy" # or "sundials"
    Only "scipy" is avaliable due to memory leak problem of sundials

Examples

>>> model.set_configuration(callbacklevel = 1)
>>> model.set_configuration(iteration_max = 1000)
set_constrain(group, id, type, value=0.1, stdev=1.0)

Setter of a metabolic constrains.

Please perform update() after model modification.

Parameters:

  • group (str) – group of parameters (“reaction”, “reversible”, “metabolite”)

  • id (str) – id in model definition

  • type (str) – type of constrain “fitting”, “fixed”, “free”, “pseudo”: * “fitting” : For the calclation of RSS by using “value” and stdev * “fixed” : Fixed to “value”. Not used for the calculation of RSS. * “free” : Free between lower and upper boundary. Not used for the calculation of RSS. * “pseudo” : Pseudo reaction is a special “free” reaction to consider G-index. The reaction is ignored in a stoichiometry matrix. However, the reaction is considered in the MDV calculation.

  • value (float) – Level of metabolic flux

  • stdev (float) – Standard deviation of metabolic flux level used for ‘fitting’ type reactions.

Returns:

True/False.

Return type:

Booleans

Examples

>>> model.set_fixed_reaction("reaction", 'pgi', "fixed", 100.0,1)

See also

get_constrain

set_constraints_from_state_dict(dict)

Reaction types, value, stdev, lb, and ub are set from the state dict data at once

Parameters:

dict (dict) – Dictionary of flux. Data in ‘value’, ‘stdev’, ‘lb’, ‘ub’ and ‘type’ fields are used.

Examples

>>> model.set_constraints_from_state_dict(flux_dict)
set_experiment(name, mdv, carbon_sources, startidv=[])

Setter of an ‘experiment’ to metabolic model.

Here an ‘experiment’ indicated a set of a carbon source labeling pattern and a measured MDV data.

Parallel labeling experiment can be performed by setting multiple sets of ‘experiment’.

Parameters:

  • name (str) – name of the experiment (unique)

  • mdv (instance) – MDVdata instance including measured MDV data of target fragments.

  • carbon_sources (instance) – Instance of carbon source object

  • startidv (array) – array of idv as starting isotope distribution for INST

Examples

>>> model.set_experiment('ex1', mdv1, cs1)
>>> model.set_experiment('ex2', mdv2, cs2, startidv)
show_flux_balance(results, metabolite, filename='', format='csv')

List of metabolic flux levels of reactions related to given metabolite.

Parameters:

  • input – Tuple of (“name”, dic of metabolic statet) [(‘name1’, state1),(‘name2’, state2),(‘name3’, state3),]

  • metabolite – ID of metabolite of interest

  • filename – Results are saved when file name is not “”.

  • format – “csv” or “text”

Examples

>>> model.show_metabolite_balance(results, "Fum")
>>> model.show_metabolite_balance(results, "Fum", filename = "metabolite_balance.csv", format = "csv")
show_results(input, flux='on', rss='on', mdv='on', pool_size='off', checkrss='off', filename='', format='csv')

Method to output metabolic state data.

Parameters:

  • input (tuple) – Tuple of (“name”, dic of metabolic statet) [(‘name1’, state1),(‘name2’, state2),(‘name3’, state3),]

  • flux (str) – show flux data “on”/”off”

  • rss (check) – show rss data “on”/”off”

  • mdv (str) – show mdv data “on”/”off”

  • pool_size (str) – show metabolite data “on”/”off”

  • rss – show each RSS value “on”/”off”

  • filename (str) – Results are saved when file name is not “”/

  • format (str) – “csv” or “text”

Examples

>>> model.show_results(results, flux = "on", rss = "on", mdv = "on", pool_size = "off",  checkrss = "on")# Output to console
>>> model.show_results(results, filename = "show_results.csv", format = "csv")# Output to .csv file
show_results_in_map(flux, mapfilename, outputfilename, formattext='.1f')

Method to project metabolic state data into metabolic file (.GML).

Parameters:

  • flux (dict) – dict of metabolic state

  • mapfilename (str) – name of blank GML file

  • outputfilename (str) – name of uutput GML file

  • format (str) – format string of “format” method of string

Examples

>>> model.show_results_in_map(flux_opt1, "Example_2_cancer_map_new.gml", "Example_2_cancer_map_mapped.gml")
>>> model.show_results_in_map(flux_opt1, "Example_2_cancer_map_new.gml", "Example_2_cancer_map_mapped.gml", formattext = ".2f")
update()

Method to generate stoichiometry matrix.

A metabolic model have to be updated when any types (fixed, free, fitting, pseudo) of reactions, metabolites, and reversible reactions are modified.

Parameters:

required. (Not) –

Examples

>>> model.update()

See also

reconstruct

mfapy.mfapyio module

mfapyio.py:I/O funtions for mfapy

This module includes functions to read model description files.

Following functions are available:

load_metabolic_model_reactions
load_metabolic_model_metabolites
load_metabolic_model_reversibles
load_metabolic_model_fragments
load_metabolic_model

Example

>>> reactions, reversible, metabolites, target_fragments = mfapy.mfapyio.load_metabolic_model("Explanation_1_13CMFA_toymodel_model.txt", format = "text")

Todo

  • Support other file formats

mfapy.mfapyio.load_metabolic_model(filename, format='text', mode='normal')[source]

Function to load metabolic model information from a text file

CAUTION: Now this function has no error checking.

Parameters:

  • filename (str) – metabolic model file

  • format (str) – “text” (defalut, tab deliminated) or “csv”

  • mode (str) – “normal” (defalut) or “debug” (to show loaded metabolic file data)

Returns:

  • reactions (dict), Dictionary describing metabolite reactions

  • reversible (dict), Dictionary for defining reversible reactions

  • metabolites (dict), Dictionary including metabolite information

  • target_fragments (dict), Dictionary of target_fragments

Examples

>>> reactions, reversible, metabolites, target_fragments = load_metabolic_model("filename.txt')
# The obtaind data (dictionaries) are directly used for generation of new Metabolic Model object
>>> model = MetabolicModel(reactions, reversible, metabolites, target_fragments)
mfapy.mfapyio.load_metabolic_model_fragments(filename, format='text', mode='normal')[source]

Function to load mass fragment information from a metabolic model file with following format.

Examples:

//Target_fragments
Glue    gcms    Glu_1:2:3:4:5   use     C5H10N2O3
Gluee   gcms    Glu_1:2:3+Glu_4:5       use     C5H10N2O3
Parameters:

  • filename (str) – metabolic model file

  • format (str) – “text” (defalut, tab deliminated) or “csv”

  • mode (str) – “normal” (defalut) or “debug” (to show loaded metabolic file data)

Returns:

Dictionary of target_fragments

Return type:

dict

Examples

>>> target_fragments = load_metabolic_model_fragments('filename.txt')
>>> print target_fragments
{'Thr302': {'atommap': 'Thr_12', 'use': 'no', 'type': 'gcms', 'order': 35, 'number': 3}, ...}
mfapy.mfapyio.load_metabolic_model_metabolites(filename, format='text', mode='normal')[source]

Function to load Metabolite information from a file with following format.

Examples:

//Metabolites
CO2ex   1       no      no      excreted        (kegg:C00011)   0.0     300
AcCoA   2       no      carbonsource    no      (kegg:C00024)   0.0     300
OAC     4       no      no      no      (kegg:C00036)   0.0     300
Cit     6       no      no      no      (kegg:C00158)   0.0     300
AKG     5       no      no      no      (kegg:C00026)   0.0     300
Suc     4       symmetry        no      no      (kegg:C00042)   0.0     300
Fum     4       symmetry        no      no      (kegg:C00122)   0.0     300
Glu     5       no      no      no      (kegg:C00025)   0.0     300
Asp     4       no      carbonsource    no      (kegg:C00049)   0.0     300
Parameters:

  • filename (str) – metabolic model file

  • format (str) – “text” (defalut, tab deliminated) or “csv”

  • mode (str) – “normal” (defalut) or “debug” (to show loaded metabolic file data)

Returns:

Dictionary including metabolite information

Return type:

dict

Examples

>>> metabolites = load_metabolic_model_metabolites('filename.txt')
>>> print metabolites
{'GLUEX': {'excreted': 'no', 'carbonsource': 'carbonsource', 'C_number': 5, 'symmetry': 'no', 'order': 4}, ...}
mfapy.mfapyio.load_metabolic_model_reactions(filename, format='text', mode='normal')[source]

Function to load metabolic reaction information from a text or CSV file with following format.

Examples:

//Reactions
v1      AcCoA + OAC --> Cit     AcCoA + OAC --> Cit     AB + CDEF --> FEDBAC    (kegg:R00351)   0.0     300
v2      Cit --> AKG + CO2ex     Cit --> AKG + CO2ex     ABCDEF --> ABCDE + F    (kegg:R00709)   0.0     300
v3      AKG --> Glu     AKG --> Glu     ABCDE --> ABCDE         (kegg:R00243)   0.0     300
v4      AKG --> Suc + CO2ex     AKG --> Suc + CO2ex     ABCDE --> BCDE + A      (kegg:R01197)   0.0     300
v5      Suc --> Fum     Suc --> Fum     ABCD --> ABCD   (kegg:R02164)   0.0     300
v6      Fum --> OAC     Fum --> OAC     ABCD --> ABCD   (kegg:R01082)   0.0     300
v7      OAC --> Fum     OAC --> Fum     ABCD --> ABCD   (kegg:R01082)   0.0     300
v8      Asp --> OAC     Asp --> OAC     ABCD --> ABCD   (kegg:R00355)   0.0     300
Parameters:

  • filename (str) – metabolic model file name

  • format (str) – “text” (defalut, tab deliminated) or “csv”

  • mode (str) – “normal” (defalut) or “debug” (to show loaded metabolic file data).

Returns:

Dictionary describing metabolite reactions

Return type:

dict

Examples

>>> reactions = load_metabolic_model_reaction('filename.txt')
>>> print reactions
{'r4': {'atommap': 'AB+CDEF-->FEDBAC', 'reaction': 'ACCOA+OAA-->ICI', 'use': 'use', 'lb': 0.1, 'flux_value': 0.0, 'reversible': 'no', 'flux_var': 1.0, 'type': 'free', 'order': 3, 'ub': 300.0},...}
mfapy.mfapyio.load_metabolic_model_reversibles(filename, format='text', mode='normal')[source]

Function to load definitions of reversible reactions from a metabolic model file with following format.

Examples:

//Reversible_reactions
FUM     v6      v7      (kegg:R01082)   0.0     300
Parameters:

  • filename (str) – metabolic model file

  • format (str) – “text” (defalut, tab deliminated) or “csv”

  • mode (str) – “normal” (defalut) or “debug” (to show loaded metabolic file data)

Returns:

Dictionary for defining reversible reactions

Return type:

dict

Examples

>>> reversible = load_metabolic_model_reversibles('filename.txt')
>>> print reversible
{'MDH': {'flux_value': 0.0, 'reverse': 'r27', 'flux_var': 1.0, 'forward': 'r26', 'type': 'free', 'order': 6}, ...}

mfapy.optimize module

optimize.py:low level optimizer functions used in mfapy.

These functions were separated from model instance for the parallel execution.

Todo

  • Cleaning-up and support other optimizers

mfapy.optimize.calc_MDV_from_flux(tmp_r, target_fragments, mdv_carbon_sources, func, timepoint=[], y0temp=[])

Low level function to calculate mdv vector and mdv hash from metabolic flux and carbon source MDV using calmdv.

This funcition is called from mfapy.metabolicmodel.show_results.

Parameters:

  • tmp_r (array) – list of metabolic state data (tmp_r = numpy.dot(matrixinv, Rm_temp)

  • target_fragments (array) – list of targed mdvs for MDV calclation, model.target_fragments.keys()

  • mdv_carbon_sources (dict) – dict of mdv_carbon_sources in model.experiments[ex_id][‘mdv_carbon_sources’]

  • func (dict) – Dict of functions for MDV calclation in model.func

  • timepoint (array) – For INST mode only. timepoints for MDV comparison in model.experiments[ex_id][‘timepoint’] When the length of timepoint array >= 1, INST mode is used.

  • y0temp (dict) – Start IDV state for INST mode

Returns:

  • mdv (array) list of MDV data

    • mdv_hash (dict) dict of MDV data

INST-MFA mode:

  • mdv (array) array of mdv at each time point

  • mdv_hash (array) array of mdv_hash at each time point

Return type:

13C-MFA mode

Example

>>> mdv_exp, mdv_hash = optimize.calc_MDV_from_flux(tmp_r, target_fragments_temp, mdv_carbon_sources_temp, self.func)

See also

mfapy.metabolicmodel.show_results

mfapy.optimize.calc_MDV_residue(x, *args, **kwargs)

Low level function for residual sum of square calculation from mfapy.metabolicmodel.MetaboliModel.calc_rss

Parameters:

  • x (array) – vector of independent flux.

  • *args (array) – list of parameters.

  • **kwargs (dict) – dic of parameters.

Returns:

RSS + Penalty score (When out side of the lower and upper boundaries)

Return type:

float

See also

fit_r_mdv_scipy

mfapy.optimize.calc_MDV_residue_nlopt(x, grad, kwargs)

Low level function for residual sum of square calculation for model fitting using nlopt.nlopt

Parameters:

  • x (array) – vector of independent flux.

  • *args (array) – list of parameters.

Returns:

RSS + Penalty score (When out side of the lower and upper boundaries)

Return type:

float

See also

fit_r_mdv_scipy

mfapy.optimize.calc_MDV_residue_scipy(x, *args)

Low level function for residual sum of square calculation for model fitting using scipy.optimize.minimize

Parameters:

  • x (array) – vector of independent flux.

  • *args (array) – list of parameters.

Returns:

RSS + Penalty score (When out side of the lower and upper boundaries)

Return type:

float

See also

fit_r_mdv_scipy

mfapy.optimize.calc_protrude_nlopt(independent_flux, grad, kwargs)

Objective function used in initializing_Rm_fitting (nlpot)

Calc penalty score of metabolic state out side of the feasible space.

Parameters:

  • independent_flux (array) – vector of independent flux

  • grad – not used

  • *args (array) – list of parameters.

Returns:

Penalty score

Return type:

float

See also

initializing_Rm_fitting

mfapy.optimize.calc_protrude_scipy(independent_flux, *args)

Objective function used in initializing_Rm_fitting (SLSQP)

This function calculates penalty score of metabolic state out side of the feasible space.

Parameters:

  • independent_flux (array) – vector of independent flux

  • *args (list) – list of parameters.

Returns:

Penalty score

Return type:

float

See also

initializing_Rm_fitting

mfapy.optimize.fit_r_mdv_deep(configure, experiments, numbers, vectors, matrixinv, func, flux)

Low level function for model fitting by iterative fittings.

  • 1st iteration: GN_CRS2_LM (global optimizer)

  • 2n th iterations: SLSQP (local)

  • 2n + 1 th iterations: LN_SBPLX (local)

This combination is empirically best

Parameters:

  • configures (dict) – “model.configures” including various configulatoins of the model.

  • experiments (dict) – “model.experiments” including experiments defined in the model.

  • numbers (dict) – “model.numbers” including various numbers of the model.

  • vectors (dict) – “model.vector” including various vectors of the model.

  • matrixinv (2d array) – “model.matrixinv” is a inversed matrix for the flux calculation.

  • func (dict) – Dict of functions for MDV calclation in model.func

  • flux (dict) – Dictionary of initial metabolic state.

Returns:

  • state (str) finishing condition

  • kai (float) Residual sum of square of fitted metabolic state

  • opt_flux (array) list of fitted metabolix state

  • Rm_ind_sol (array) list of fitted independent flux

Example

>>> state, kai, opt_flux, Rm_ind_sol = optimize.fit_r_deep(configure, self.experiments, numbers, vectors, self.matrixinv, self.func, flux)

See also

optimize.fit_r_nlopt optimize.fit_r_scipy

mfapy.optimize.fit_r_mdv_nlopt(configure, experiments, numbers, vectors, matrixinv, func, flux, method='LN_PRAXIS')

Low level function for model fitting using nlopt.opt

Parameters:

  • configures (dict) – “model.configures” including various configulatoins of the model.

  • experiments (dict) – “model.experiments” including experiments defined in the model.

  • numbers (dict) – “model.numbers” including various numbers of the model.

  • vectors (dict) – “model.vector” including various vectors of the model.

  • matrixinv (2d array) – “model.matrixinv” is a inversed matrix for the flux calculation.

  • func (dict) – Dict of functions for MDV calclation in model.func

  • flux (dict) – Dictionary of initial metabolic state.

  • method (str) – “LN_COBYLA”, “LN_BOBYQA”, “LN_NEWUOA”, “LN_PRAXIS”, “LN_SBPLX”, “LN_NELDERMEAD”, “GN_DIRECT_L”, “GN_CRS2_LM”,”GN_ESCH”

Returns:

  • state (str) finishing condition

  • kai (float) Residual sum of square of fitted metabolic state

  • opt_flux (array) list of fitted metabolix state

  • Rm_ind_sol (array) list of fitted independent flux

Example

>>> state, kai, opt_flux, Rm_ind_sol = optimize.fit_r_mdv_nlopt(configure, self.experiments, numbers, vectors, self.matrixinv, self.func, flux, method = "LN_PRAXIS")

See also

calc_MDV_residue_nlopt

mfapy.optimize.fit_r_mdv_scipy(configure, experiments, numbers, vectors, matrixinv, func, flux, method='SLSQP')

Low level function for model fitting using scipy.optimize.minimize

Parameters:

  • configures (dict) – “model.configures” including various configulatoins of the model.

  • experiments (dict) – “model.experiments” including experiments defined in the model.

  • numbers (dict) – “model.numbers” including various numbers of the model.

  • vectors (dict) – “model.vector” including various vectors of the model.

  • matrixinv (2d array) – “model.matrixinv” is a inversed matrix for the flux calculation.

  • func (dict) – Dict of functions for MDV calclation in model.func

  • flux (dict) – Dictionary of initial metabolic state.

  • method (str) – “SLSQP” and “COBYLA” are available

Returns:

  • state (str) finishing condition

  • kai (float) Residual sum of square of fitted metabolic state

  • opt_flux (array) list of fitted metabolix state

  • Rm_ind_sol (array) list of fitted independent flux

Example

>>> state, kai, opt_flux, Rm_ind_sol = optimize.fit_r_mdv_scipy(configure, self.experiments, numbers, vectors, self.matrixinv, self.func, flux, method = "SLSQP")

See also

calc_MDV_residue_scipy

mfapy.optimize.initializing_Rm_fitting(numbers, vectors, matrixinv, template, initial_search_iteration_max, method='fitting')

Funcition to generate randomized initial flux dixtribution using scipy.optimize.minimize SLSQP

Parameters:

  • numbers (dict) – “model.numbers” including various number related data of the model.

  • vectors (dict) – “model.vector” including various vector related data of the model.

  • matrixinv (numpy 2d array) – “model.matrixinv” is a inversed matrix of stoichiometry matrix for flux calculation.

  • template (dict) – Dictionary of metabolic state. When template is available, metabolic state most similar to the template is generated. The function is used in the grid search.

  • initial_search_iteration_max (int) – “configure[“initial_search_iteration_max”]”. Maximal number of interations (steps) allowed in each task to find feasible initial metabolic flux distribution.

  • method (str) – “fitting” is only available.

Returns:

tmp_r (list) list of metabolic state data (tmp_r = numpy.dot(matrixinv, Rm_temp)

Rm_temp (list) metabolic state vector

Rm_ind (list) independent flux vector

state (str) State of finishing condition “Failed”/”Determined”

Examples

>>> tmp_r, Rm_temp, Rm_ind, state = optimize.initializing_Rm_fitting(numbers, vectors, matrixinv, template ,initial_search_iteration_max)

See also

calc_protrude_scipy

Module contents