Replica Exchange Molecular Dynamics Service

Replica Exchange Molecular Dynamics Simulation (REMD)

In REMD, a series of non-interactive replica systems are reconstructed, covering a wide temperature range from low to high temperatures. Perform independent molecular dynamics simulations for each copy. According to the Metropolis standard, the configuration of each copy adjacent in temperature can be exchanged. Therefore, REMD simulation can make the low-temperature configuration space escape from the lowest point of local potential energy. Compared with traditional dynamics, REMD simulation can sample a larger configuration space. But the disadvantage of this method is that the amount of calculation is very large, the number of copies is proportional to, and N is the degree of freedom of the system.

Figure 1. Illustration of replica exchange molecular dynamics (REMD) method.

Overall solutions

• The REMD method combines the molecular dynamics trajectory with the temperature exchange Monte Carlo process to effectively sample the conformational space.
• A series of replicas run in parallel at a parallel temperature from the required temperature to the high temperature, at which the replicas can easily overcome the energy barrier.
• The configurations of adjacent replicas will be exchanged from time to time, and the exchange will be accepted or rejected according to Metropolis acceptance criteria to ensure a detailed balance.

Algorithm

Once the peptide and its surrounding environment are determined, it is necessary to determine the range of the temperature space to be sampled, the number of processors used, and the length of time for the simulation. The system, the number of copies, the range of temperature space and the temperature distribution determine the average exchange between copies Probability. These values should be the same for all replicas. For this reason, if it is assumed that there is no known bottleneck in the free energy space, the potential energy of the system is expected to increase with increasing temperature, so the temperature distribution of the replicas should obey the exponential distribution.

Process of umbrella sampling simulation

• Define the system, for example: peptide + solvent (implicit or explicit).
• According to the number of available processors and the temperature range to be sampled (in fact, they have a very strong correlation), select the temperature distribution. Use exponential distribution: the exponential form can ensure that the temperature interval increases with increasing temperature. This is necessary, Because the distribution of total energy increases with increasing temperature, the exchange probability also increases. Keep the exchange probability unchanged with temperature.
• After obtaining the temperature distribution, balance the system separately at N temperatures.
• Then, run a short-term REMD simulation to obtain an estimate of the exchange probability. If the obtained result deviates greatly from the required value, we will try to modify the temperature.

Software

• GROMACS The REMD simulation can be performed using popular MD simulation packages including GROMACS, AMBER, CHARMM, NAMD, etc.
• High Performance Computing (HPC) Cluster REMD simulation is highly parallel and requires high resources. We have GROMACS-4.5.3 HPC cluster and standard message passing interface (MPI) library. Generally, having two cores per copy can increase productivity in HPC clusters equipped with Intel Xeon X5650 CPUs or higher.
• Visual Molecular Dynamics (VMD) In this protocol, the versatile package VMD is used for molecular modeling and structure visualization. VMD supports computers running MacOS X, Unix, or Windows systems, and is distributed free of charge.

AI-based approaches

With the development of high-precision force fields and the development of powerful computers, molecular dynamics (MD) has become a universal tool for studying the complex structure, thermodynamics and dynamic processes of interdisciplinary real systems. Therefore, we can provide AI-based copy exchange dynamics simulation, repeat the iterative MD-AI-MD-AI process until the required sampling is obtained. These spaces encode all relevant slow dynamics in the system, while the remaining fast dynamics are effectively replaced. Then, each round of AI will put a given sample of the slow mode, and after that the information is used to initiate a new biased round. Switching along the slow mode will result in an increase in exploration, and any incremental increase in exploration can be used in another round of AI to more accurately estimate the slow mode. Then, the unbiased thermodynamic and kinetic information can be recovered from the AI-enhanced MD trajectory thus obtained using standard rated procedures.

Why choose us?

• Innovate research design and technology to improve research reliability, improve results and increase efficiency
• Achieve flexibility by freeing up your R&D resources to focus on other strategic plans.
• High efficiency, fast response speed, flexible research direction, low price, short cycle and high quality.
• Tailor-made production for customers to help customers reduce capital and time costs.