Simulation of mitochondrial metabolism using multiagents system

1
Simulation of mitochondrial metabolism using
multi-agents system

• Charles LALES 1Nicolas PARISEY 2 Jean-Pierre
  MAZAT 2 Marie BEURTON-AIMAR 1

1 LaBRI CNRS UMR 5800, Univ. Bordeaux 1, France2
Mitochondrial Physiopathology Lab. INSERM U688,
Univ. Bordeaux 2, France
Bordeaux Universities, FRANCE
2
Contents

• Biological problematic mitochondrial
  metabolism.
• Which modeling paradigm?
• MAS model for 3D membrane
• Overview.
• Granularity of agents.
• Abstraction of molecules.
• Interactions as forces.
• Implementation to simulate phospholipids.
• Conclusion and perspectives.

3
Mitochondrion

• Mitochondrion is an organelle which converts
  organic materials into cell energy ATP.
• It is an endosymbiot
• With its own DNA,
• With its own metabolism.

4
Mitochondrion an endosymbiot
its own metabolism
Its own DNA
5
Membranes

• Why looking after membranes ?
• Metabolism container
• Metabolism actor
• Problems induiced
• Modeling reconcilable membranes/métabolisme
• Structures complexes
• Structures dynamiques

6
Membranes

• Why study mitochondrial membranes ?
• metabolism container,
• metabolism player.
• Problems
• heterogeneous morphologies,
• complex structure,
• dynamical structure.

7
Membranes
Rossignol and al., 2004
8
Comparatif EDO/SMA
9
MAS paradigm?

• Why use differential equations?
• Modeling the whole system macro scale available
  data,
• Inherent mathematical proves,
• Limitations of this paradigm?
• Do not take into acount individual variabilities,
• Take hardly care of a compartmentalized space,
• Why trying multi-agents paradigm?
• To solve the above limitations,
• It is easier to translate biological hypotheses
  as microscopic behaviors.

10
Which modeling paradigms?

• Ordinary Differential Equations (ODE).
• Partial Differential Equations (PDE).
• Petri Nets.
• Mutli-Agent Systems (MAS).

11
MAS model - overview

• Biological Objects
• reactive BioAgents.
• Time
• discontinuous (time steps) for TimeAgent,
• continuous simulated (events) in futur.
• 3D Space
• continuous coordinates of BioAgents,
• discretized coordinates of GridAgents
• optimisation (neighbourhood),
• diffusion, (temperature, pH,).

12
MAS model - granularity
-- molecule set
-- molecule
-- atom set
-- atom
13
MAS model - agent granularity
-- molecule set
-- molecule
-- atom set
-- atom
14
MAS model - molecule abstraction

• Goals
• 3D Conformation,
• Spatial orientation,
• Inner dynamic.
• BioAgent model
• 1 gravity center,
• n interacting points.

15
MAS model - interactions

• A set of forces reflects physical and chemical
  properties.
• Forces depend on
• Type of interacting points,
• Distance between them.
• Forces take into account
• Forces generate linear mouvements whereas induced
  torques generate rotational mouvements

16
MAS model - interactions

• 3D rigid body dynamics
• Euler methode to approximate Newtons law of
  motion
• Quaternions used to manage 3D rotations

Linear mouvement
Rotational mouvement
17
MAS design

• A UML Class diagram shows the classes of the
  system
• A rendering engine uses OpenGL

18
(No Transcript)
19
Implémentation
20
Work in progress

• Parameters calibration
• Quantitative sources (used in molecular dynamics)
  
• PDB files (RMN, crystallography),
• force fields (physics).
• Qualitative sources (phase diagrams)
• micelles,
• membranes (monolayer, bilayer).
• Simulation of enzymatic reactions.
• Mixing model of membranes and enzymes.

21
Work in progress

• Dynamics engine
• Use of Open Dynamics Engine (ODE).
• MAS multi-scaling
• Vertical/horizontal flow of information.
• Constraint management.
• Example of micelles

22
Work in progress

• Parameters calibration
• pm(m1)8n
• p parameters to describe interaction function,
• m type of interacting points,
• n molecules.
• Quantitatives sources (used in molecular
  dynamics)
• PDB files (RMN, cristallography),
• force fields (physics).
• Qualitatives sources (phase diagrams)
• micelles,
• membranes (monolayer, bilayer).

23
Work in progress

• Enzymatic reactions
• Typical (tripsine),
• In mitochondria.
• Mixing model of membranes and enzymes
• Dynamics engine
• ODE (Open Dynamics Engine)
• MAS multi-scaling
• Vertical/horizontal flow of information,
• Constraint management.

24
Conclusion and perspectives

• Features of the model
• 3D Conformation,
• Spatial orientation,
• Inner dynamic.
• Membrane simulation
• Micelle, monolayer
• Still work to do
• Calibration membranes (bilayers)
• Quantitatives sources (PDB files,...),
• Qualitatives sources (phase diagrams).
• Simulation of enzymatic reactions,
• Mixing simulations of the membranes and the
  enzymes.

25
Perspectives
26
MitoScoP

• Mitochondria in Silico Project (ACI IMPBio).
• Knowledge base on mitochondria,
• Multi-paradigm modeling
• ODE,
• MAS,
• Graphes (Petri Nets),
• Mitochondria simulations framework.
• Team
• Marie BEURTON-AIMAR aimar_at_labri.fr
• Charles LALES - lales_at_labri.fr
• Jean-Pierre MAZAT - JP.Mazat_at_phys-mito.u-bordeaux2
  .fr
• Christine NAZARET - nazaret_at_sm.u-bordeaux2.fr
• Nicolas PARISEY - nicolas.parisey_at_etud.u-bordeaux2
  .fr
• Sabine PERES - sabine.peres_at_etud.u-bordeaux2.fr
• Christine REDER - Christine.Reder_at_math.u-bordeaux1
  .fr
• Thanks for your attention )

27
MitoScoP

• Capitalisation de connaissances sur la
  mitochondrie,
• Modélisation multi-paradigme
• ODE,
• SMA,
• Graphe (Réseau de Pétri).
• Plateforme de simulation pour ces modèles.

28
Thanks for your attention )
MitoScoP Mitochondria in Silicon Project
29
Thanks for your attention )

• Team
• 1 Charles LALES - lales_at_labri.fr
• 2 Nicolas PARISEY - nicolas.parisey_at_etud.u-bordeau
  x2.fr
• 2 Jean-Pierre MAZAT - JP.Mazat_at_phys-mito.u-bordeau
  x2.fr
• 1 Marie BEURTON-AIMAR aimar_at_labri.fr
• Lab.
• 1 LaBRI CNRS UMR 5800, Univ. Bordeaux 1, France
  www.labri.fr
• 2 Mitochondrial Physiopathology Lab. INSERM U688,
  Univ. Bordeaux 2, France - www.phys-mito.u-bordeau
  x2.fr
• Special thanks to
• Guillaume Beslon, computer sciences professor
  BIM INSA,
• Jean-Michel Fayard, director BIM INSA.