Question icon
Your Current Search
Choose below to refine your search
Research Topic
Download abstract book

Download the NI2012 abstract book here. The page numbers in the index are clickable for easy browsing.


A cytoskeleton-based 3D morphological simulation method for detailed local structure of the developing neural cell

Filed under:

Naoto Yukinawa (Graduate School of Informatics, Kyoto University), Honda Naoki (Graduate School of Informatics, Kyoto University), Shin Ishii (Graduate School of Informatics, Kyoto University)

To simulate the detailed morphological diversities of a developing neural cell, it is required to construct a model considering the mutliscale and multiphysics aspects of various cellular processes that underlies the morphological diversity and complexity; and this accordingly requires large computational power and time.

For this purpose, we have developed a distributed simulation framework which captures three different physical layers including reaction-diffusion processes, membrane dynamics, and F-actin-based cytoskeletal kinetics [1]. However the simulation scheme has two main drawbacks in the aspects of the biological plausibility for achieving to simulate complicated morphological processes such as formations of filopodia and neurites in neural development and the computational efficiency for large scale simulations: 1) the model limitations which are capable of only 2-dimensional morphological simulation and 2) poor performance scalability in cluster computers with over hundreds of cores due to a communication bottleneck during the filament kinetics computation.

In this study, we extend our previous simulation model for incorporating the binding by an actin cross-linker Fascin and the elastic properties of actin filaments based on a 3-dimensional spring network model. We also propose a two-step approach for resolving the computational performance which incorporates a subspace distribution method for the filament kinetics computation and a three-layered hierarchical management of computational nodes which optimizes the communication load. We show that our simulation framework well reproduces the fine local structure in the filopodial formation in computationally effective way.

[1] Nonaka. S., Naoki, H., and Ishii, S.: A multiphysical model of cell migration integrating reaction-diffusion, membrane and cytoskeleton. Neural Networks, Volume 24, Issue 9, November 2011, Pages 979–989, DOI: 10.1016/j.neunet.2011.06.009.
Preferred presentation format: Poster
Topic: Large scale modeling

Filed under: