Symbolic Kinetic Models in Python (SKiMpy): Intuitive modeling of large-scale biological kinetic models

Weilandt, Daniel; Salvy, Pierre; Masid, María; Fengos, Georgios; Eriksson, Robin Denhardt; Hosseini, Zhaleh; Hatzimanikatis, Vassily

doi:10.1101/2022.01.17.476618

Cited by 5 publications

(9 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Structural Kinetic Modelling, the ORACLE framework [ 28 , 29 , 30 , 31 , 32 ], and similar methods [ 33 , 34 , 35 , 36 ] replace reaction elasticities with random numbers. Reaction rates increase with the substrate concentrations and decrease with the product concentrations.…”

Section: Introductionmentioning

confidence: 99%

Structural Thermokinetic Modelling

Liebermeister

2022

Metabolites

View full text Add to dashboard Cite

To translate metabolic networks into dynamic models, the Structural Kinetic Modelling framework (SKM) assumes a given reference state and replaces the reaction elasticities in this state by random numbers. A new variant, called Structural Thermokinetic Modelling (STM), accounts for reversible reactions and thermodynamics. STM relies on a dependence schema in which some basic variables are sampled, fitted to data, or optimised, while all other variables can be easily computed. Correlated elasticities follow from enzyme saturation values and thermodynamic forces, which are physically independent. Probability distributions in the dependence schema define a model ensemble, which allows for probabilistic predictions even if data are scarce. STM highlights the importance of variabilities, dependencies, and covariances of biological variables. By varying network structure, fluxes, thermodynamic forces, regulation, or types of rate laws, the effects of these model features can be assessed. By choosing the basic variables, metabolic networks can be converted into kinetic models with consistent reversible rate laws. Metabolic control coefficients obtained from these models can tell us about metabolic dynamics, including responses and optimal adaptations to perturbations, enzyme synergies and metabolite correlations, as well as metabolic fluctuations arising from chemical noise. To showcase STM, I study metabolic control, metabolic fluctuations, and enzyme synergies, and how they are shaped by thermodynamic forces. Considering thermodynamics can improve predictions of flux control, enzyme synergies, correlated flux and metabolite variations, and the emergence and propagation of metabolic noise.

show abstract

Section: Introductionmentioning

confidence: 99%

Structural Thermokinetic Modelling

Liebermeister

2022

Metabolites

View full text Add to dashboard Cite

show abstract

“…While REKINDLE can use parameter sets of any kinetic modeling framework for the training, we employed the ORACLE framework 20,25,31–35 implemented in the SKiMpy tool 36 to generate a population of 80000 kinetic parameter sets for each physiology. The goal was to generate kinetic models that satisfy the observed steady-state and have dynamic responses that are faster than 6-7 minutes (which corresponds to 1/3 of the E. coli doubling time) and slower than 1 μ s. The kinetic models satisfying these two conditions can reliably reproduce experimentally measured metabolic responses in E. coli (Methods).…”

Section: Resultsmentioning

confidence: 99%

“…Kinetic nonlinear models used in this study represent the central carbon metabolism of wild-type E. coli . They are based on the model published by Varma and Palsson 30 and studied extensively using the SKimPy toolbox 36 (Supplementary Fig. 1).…”

Section: Methodsmentioning

confidence: 99%

“…A TensorFlow implementation of the REKINDLE workflow is available at https://github.com/EPFL-LCSB/rekindle and https://gitlab.com/EPFL-LCSB/rekindle. The ORACLE framework is implemented in the SKimPy (Symbolic Kinetic models in Python) 36 toolbox, available at https://github.com/EPFL-LCSB/skimpy.…”

Section: Code and Data Availabilitymentioning

confidence: 99%

“…We enumerated these four cases as Physiology 1 -4 (Table 1). 20,25,[31][32][33][34][35] implemented in the SKiMpy tool 36 to generate a population of 80000 kinetic parameter sets for each physiology. The goal was to generate kinetic models that satisfy the observed steady-state and have dynamic responses that are faster than 6-7 minutes (which corresponds to 1/3 of the E. coli doubling time) and slower than 1 𝜇s.…”

Section: Rekindle Generates Large-scale Kinetic Models Of E Coli Meta...mentioning

confidence: 99%

See 2 more Smart Citations

Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks

Choudhury

Moret

Salvy

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Kinetic models of metabolic networks relate metabolic fluxes, metabolite concentrations, and enzyme levels through well-defined mechanistic relations rendering them an essential tool for systems biology studies aiming to capture and understand the behavior of living organisms. However, due to the lack of information about the kinetic properties of enzymes and the uncertainties associated with available experimental data, traditional kinetic modeling approaches often yield only a few or no kinetic models with desirable dynamical properties making the computational analysis unreliable and computationally inefficient. We present REKINDLE (REconstruction of KINetic models using Deep LEarning), a deep-learning-based framework for efficiently generating large-scale kinetic models with dynamic properties matching the ones observed in living organisms. We showcase REKINDLE's efficiency and capabilities through three studies where we: (i) generate large populations of kinetic models that allow reliable in silico testing of hypotheses and systems biology designs, (ii) navigate the phenotypic space by leveraging the transfer learning capability of generative adversarial networks, demonstrating that the generators trained for one physiology can be fine-tuned for another physiology using a low amount of data, and (iii) expand upon existing datasets, making them amenable to thorough computational biology and data-science analyses. The results show that data-driven neural networks assimilate implicit kinetic knowledge and structure of metabolic networks and generate novel kinetic models with tailored properties and statistical diversity. We anticipate that our framework will advance our understanding of metabolism and accelerate future research in health, biotechnology, and systems and synthetic biology. REKINDLE is available as an open-access tool.

show abstract