Limits...
Computational Modeling of Lipid Metabolism in Yeast

View Article: PubMed Central - PubMed

ABSTRACT

Lipid metabolism is essential for all major cell functions and has recently gained increasing attention in research and health studies. However, mathematical modeling by means of classical approaches such as stoichiometric networks and ordinary differential equation systems has not yet provided satisfactory insights, due to the complexity of lipid metabolism characterized by many different species with only slight differences and by promiscuous multifunctional enzymes. Here, we present an object-oriented stochastic model approach as a way to cope with the complex lipid metabolic network. While all lipid species are treated objects in the model, they can be modified by the respective converting reactions based on reaction rules, a hybrid method that integrates benefits of agent-based and classical stochastic simulation. This approach allows to follow the dynamics of all lipid species with different fatty acids, different degrees of saturation and different headgroups over time and to analyze the effect of parameter changes, potential mutations in the catalyzing enzymes or provision of different precursors. Applied to yeast metabolism during one cell cycle period, we could analyze the distribution of all lipids to the various membranes in time-dependent manner. The presented approach allows to efficiently treat the complexity of cellular lipid metabolism and to derive conclusions on the time- and location-dependent distributions of lipid species and their properties such as saturation. It is widely applicable, easily extendable and will provide further insights in healthy and diseased states of cell metabolism.

No MeSH data available.


Model workflow scheme. Schematic representation of the model workflow, exemplified for PS synthase. (A) In every time step the order of reaction events is randomized. (B) A random number is drawn Nmax times, if the number exceeds the threshold P the reaction is performed once. (C) A random CDP-DG object i is taken from the substrate list, here CDP-DG list. The reaction consumes a serine precursor molecule and releases a CMP from the CDP-DG headgroup (sn3 position). The new PS object is appended to the product list, here PS list.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5037213&req=5

Figure 1: Model workflow scheme. Schematic representation of the model workflow, exemplified for PS synthase. (A) In every time step the order of reaction events is randomized. (B) A random number is drawn Nmax times, if the number exceeds the threshold P the reaction is performed once. (C) A random CDP-DG object i is taken from the substrate list, here CDP-DG list. The reaction consumes a serine precursor molecule and releases a CMP from the CDP-DG headgroup (sn3 position). The new PS object is appended to the product list, here PS list.

Mentions: We implemented a simulation framework that combines aspects of agent-based modeling and τ-leaping in Python. We defined the participating biomolecules (c.f. Table 1) as objects, which can be equipped with a set of attributes. These attributes include bound elements (fatty acids with different saturation states or headgroups) and the localization to a specific membrane (c.f. Figure 1C). Reactions are implemented to modify those attributes, in dependency of the substrate availability and specific kinetic parameters. The reactions can thereby use different allowed substrates (i.e., fatty acids or headgroups) as defined by the reaction rules and append them to a specific lipid. We adopted the time discretization from τ-leaping: The model is simulated in 1 s time steps for 120 min, representing the transition through one cell cycle. Within this time interval, more than one reaction can be executed. For the stochastic simulation we use the random number generator from the Python package numpy, which implements a version of the Mersenne twister sequence. To save computation time, we scaled all compound numbers by a factor of 104.


Computational Modeling of Lipid Metabolism in Yeast
Model workflow scheme. Schematic representation of the model workflow, exemplified for PS synthase. (A) In every time step the order of reaction events is randomized. (B) A random number is drawn Nmax times, if the number exceeds the threshold P the reaction is performed once. (C) A random CDP-DG object i is taken from the substrate list, here CDP-DG list. The reaction consumes a serine precursor molecule and releases a CMP from the CDP-DG headgroup (sn3 position). The new PS object is appended to the product list, here PS list.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5037213&req=5

Figure 1: Model workflow scheme. Schematic representation of the model workflow, exemplified for PS synthase. (A) In every time step the order of reaction events is randomized. (B) A random number is drawn Nmax times, if the number exceeds the threshold P the reaction is performed once. (C) A random CDP-DG object i is taken from the substrate list, here CDP-DG list. The reaction consumes a serine precursor molecule and releases a CMP from the CDP-DG headgroup (sn3 position). The new PS object is appended to the product list, here PS list.
Mentions: We implemented a simulation framework that combines aspects of agent-based modeling and τ-leaping in Python. We defined the participating biomolecules (c.f. Table 1) as objects, which can be equipped with a set of attributes. These attributes include bound elements (fatty acids with different saturation states or headgroups) and the localization to a specific membrane (c.f. Figure 1C). Reactions are implemented to modify those attributes, in dependency of the substrate availability and specific kinetic parameters. The reactions can thereby use different allowed substrates (i.e., fatty acids or headgroups) as defined by the reaction rules and append them to a specific lipid. We adopted the time discretization from τ-leaping: The model is simulated in 1 s time steps for 120 min, representing the transition through one cell cycle. Within this time interval, more than one reaction can be executed. For the stochastic simulation we use the random number generator from the Python package numpy, which implements a version of the Mersenne twister sequence. To save computation time, we scaled all compound numbers by a factor of 104.

View Article: PubMed Central - PubMed

ABSTRACT

Lipid metabolism is essential for all major cell functions and has recently gained increasing attention in research and health studies. However, mathematical modeling by means of classical approaches such as stoichiometric networks and ordinary differential equation systems has not yet provided satisfactory insights, due to the complexity of lipid metabolism characterized by many different species with only slight differences and by promiscuous multifunctional enzymes. Here, we present an object-oriented stochastic model approach as a way to cope with the complex lipid metabolic network. While all lipid species are treated objects in the model, they can be modified by the respective converting reactions based on reaction rules, a hybrid method that integrates benefits of agent-based and classical stochastic simulation. This approach allows to follow the dynamics of all lipid species with different fatty acids, different degrees of saturation and different headgroups over time and to analyze the effect of parameter changes, potential mutations in the catalyzing enzymes or provision of different precursors. Applied to yeast metabolism during one cell cycle period, we could analyze the distribution of all lipids to the various membranes in time-dependent manner. The presented approach allows to efficiently treat the complexity of cellular lipid metabolism and to derive conclusions on the time- and location-dependent distributions of lipid species and their properties such as saturation. It is widely applicable, easily extendable and will provide further insights in healthy and diseased states of cell metabolism.

No MeSH data available.