Core Grid Technology

Genome science has caused a shift in the entire life science paradigm. This resulted from the recognition that all biological entities are structured in finite complex combinations with each other, and that individual living organisms form high-dimensional, complex networks at different levels proteome, cellome and physiome which, in turn, construct a 3-dimensional hierarchy. From this perspective, many simulations should be performed at each level in the hierarchy. This fact differentiates our biological simulations from nano-simulations.

The Computing Grid Group 'gridifies' our original simulation programs based on our own algorithms at the levels of electrons, proteins, cells and organs (see the accompanying list), connects the input/output information from/to the databases being developed by the Data Grid Group of BioGrid, and integrates and hybridizes the individual programs at different levels.

List of Stored Contents

♦Electron level simulation 1
AMOSS (Ab initio Molecular Orbital Simulation on Supercomputer)
The quantum chemical program for computing ab initio molecular orbitals developed by the NEC Fundamental Research Laboratories offers fast, high-accuracy computation under the condition of a vast array of systems. The CI (Configuration Interaction) computation method is also under development. (figure 1)

♦Electron level simulation 2
DFT (Density Function Theory)/Generalized DFT
The program for DFT and generalized DFT, developed by the group led by Prof. Kizashi YAMAGUCHI, Graduate school of Science, Osaka Univ., can compute the electron states of complicated systems involving spin frustration, spin degeneracy, and spin canting. Thus, simulations are available for biochemical reactions, in particular, of complicated reactions involving oxidization and reduction.

♦Simulation for Protein Complexes
prestoX-basic
The molecular dynamics program is being developed in cooperation with JBIRC, Institute for Protein Research, Osaka Univ., Yokohama City University, Hitachi Corp. and Fujitsu Corp. for achievement of in silico drug screening. (figure 2)

♦Protein Fold Simulation
Prediction of ab initio protein folds from amino acid sequence information
The group led by Prof. Shoji TAKADA, Faculty of Science, Kobe Univ., develops a sophisticated algorithm to predict protein folds, and their program has received high commendation from the international prediction contest "CASP."

♦Cell Simulation
Simulation to study the drug inhibitions to ion channels in human cardiac muscle cells.
The study is being conducted by Prof. Yoshihisa KURACHI, Graduate School of Medicine, Osaka Univ.

♦Organ and Tissue Simulation
Simulations of surface muscle electric potential and lung ventilation
These studies at the organ and tissue levels are performed by Prof. Kenzo AKAZAWA and Dr. Hiroko KITAOKA, Graduate School of Information Science and Technology. Osaka Univ. They will offer an appropriate diagnosis and medical treatment.

figure 1

figure 2

A new platform for performing coupled computations by integrating application programs at different levels on the grid, which consists of various types of heterogeneous computers. In particular, the platform focusing on biosimulation is referred to as BioPfuga.

♦Methods for implementing BioPfuga

Separating application programs into smaller parts.
Standardize the data format for transferring data between different application programs.
·UDS-XML (Universal Data Set-XML)
·Scheme design and improvement of tools.

♦Coupling between Computations of Quantum Mechanics (QM) and Molecular Mechanics (MM) on BioPfuga

The hybrid-QM/MM simulation based on BioPfuga can solve many difficult problems for biological molecular systems, which have fundamentally dynamic features.
The hybrid-QM/MM simulation on the BioPfuga is now available using AMOSS for QM calculations and prestoX-basic for MM calculations. (figure 3)


figure 3	movie

Accomplishment of high quality and rapid computations using various 'gridified' application programs.
Development of the integrated technology enabling the efficient BioPfuga-based hybrid computations of various application programs on the grid computers.

The biological simulation programs will be applied to these developments at the levels of electrons, biomolecules and cells/organs to demonstrate their quality following the paradigm of life science.