| The research into the interaction between various biomacromolecules is the foundation to understand the basic mechanisms of biochemical processes. Regulation of biological activity of the receptor for the target peptide or simulate peptide ligands drug research and development is one of the most advanced areas in biology and medicine. As an important tool, molecular dynamics can simulate the motion of biomacromolecule system, conformational change mechanism and the interaction between molecules and ligands, and the interaction among each component of the molecules. It can clarify the sequence information of biomacromolecules and the complex, dynamic biological processes in micro-scale. Based on traditional experimental data, molecular dynamics can solve scientific issues more fundamentally, in order to make up for the traditional experimental techniques. Currently, it has been extensively used in the structures and functions of biomacromolecules in micro-scale such as proteins and nucleic acids.Recognition/binding/catalytic process among biomacromolecules usually occurs in the microsecond time scale or even longer, so molecular dynamics need femtosecond time step to simulate the process of calculation in order to accurately depicts the molecular motion. Besides, the ordinary personal computers and workstations cannot provide the computing resources, it must utilize the large clusters of many processors in parallel computing to implement the simulation tasks of hundreds of millions steps of tens of thousands to millions of atoms virus system, to complete the analysis of the dynamical behavior of key proteins, to understand the combination process and mechanism of cell surface ligands and receptors, and the assembly dynamics process of the viral capsid protein system. However, the first issue that we faced in practical applications is to reasonably complete and optimize the deployment of tasks in parallel clusters, to make full use of the large-scale cluster computing resources to complete the research efficiently, and finally we will build an molecular dynamics simulation research platform.Based on the parallel operating system MOSIX2, we built a PC cluster with a computing power of one hundred billion times and made molecular dynamics simulation deployment test of a million atoms system in picosecond time scale using NAMD software. Subsequently, we completed the deployment test of a million atoms system in nanosecond time scale in the ten trillion times cluster "ST10000" in the High Performance Computing Center of Shandong University (Shandong Province). Finally, by using the one hundred trillion times cluster "sunway4000A" and quadrillion times supercomputer "sunway blue light" of National Supercomputing Center in Jinan, we completed the ultra-large-scale deployment tests of tens of thousands atoms system in microsecond scales and millions atoms system in nanoscale using NAMD and GROMACS software. Furthermore, we also accomplished the replica exchanges dynamics simulations test using over ten thousand processors simultaneously. From the entire deployment test results we confirm that ordinary small-scale clusters suited to do preparatory work of the molecular dynamics simulation tasks, such as model construction, parameter optimization and super large-scale system must be deployed to one hundred trillion times and petascale cluster to use large-scale computing resources to complete the simulation task in an acceptable time range.The data of Molecular dynamics simulation is the moving trajectories of every atom of molecule in three-dimensional space and a certain period of time, so it is the three-dimensional image that can show the information contained in the macromolecules effectively. However, previous professional visualization technology and equipment are extremely expensive, thus, we built a biological macromolecules3D display and analysis platform for the analysis of the dynamic mechanism of autoimmune disease related protein molecular recognition, which is based on a low-cost3D VISION graphics solution.Based on the existing experiments, a growth factor (PGRN) can bind to tumor necrosis factor receptor (TNFR) to inhibiting the TNF-a mediated proinflammatory reaction. This paper confirms the binding mode of PGRN and TNFR2by the dynamics simulation PGRN combination of the binding mode of TNFR2and the key amino acids which mediate the combination of PGRN and TNFR2. Though building the PGRN virtual mutant, computer simulation further confirmed that electrostatic interactions mediated PGRN/TNFR2binding mode and the P as well as its anti-inflammatory function depend on these amino acids. This study discusses the primary cause of PGRN/TNFR2combination, which deepens the comprehension of their recognation progress. At the same time, the study provides the reference and basis for the further study the molecular mechanism of PGRN/TNFR signaling pathway as well as the design of new target drug source from PGRN (like Atsttrin etc), which promotes the clinical application conversion of anti-inflammatory drugs using the TNFR2as the target and the PGRN as the template, enhances the domestic development capabilities of original recombinant protein drug. |
PDF Full Text Request