Software science is the key discipline in the cell simulation projects. This is a quite different situation than other traditional simulation sciences like computational physics/chemistry.
E-Cell's approach, that is to construct cell models in a fully object-oriented fasion and make themselves working as computation engines, is rather unique. Actually I've never known of other bio-simulation software with this sort of architecture.
E-Cell 3 project (codename: Hekkoro) has two aspects concerning its relation to the previous ecell1.
- It is a re-design and re-construction of the software architecture of the ecell1
ecell3 aims at providing cell simulation community with common, highly flexible and high-performance software environment that can work as a community standard platform.
ecell3 is highly extendable. It is being developed as a set of extension module to the Python language interpreter. This means that various peripheral software components including graphical frontends, analysis modules and cell model editors/processors can easily be developed and reused by the community.
ecell1 was monolithic simulation software with an embedded GUI.
- It is a continuation of the development of the ecell1's simulation core
As I described above, E-Cell constructs cell models in an object-based mode. From our modeling, simulation and educational experiences in E-Cell Project, we have found that scalability, modeling productivity and multi-algorithmic computations are quite important for the whole cell simulation. We are sure that the object-based cell modeling and simulation is one of the essential key technologies to make it possible.
There are various subsystems with diverse computational properties in any single precise cell model, which cannot be simulated by any single computation method.
Cell models of ecell3 are constructed with object-orientation as in ecell1. The object-model, however, is much more flexible, simple and efficient than the previous version. Furthermore, ecell3 allows many simulation algorithms and time scales to coexist in a single cell model.
(Kouichi Takahashi)
