I am currently using protein structure prediction servers to study A-beta folding pathway. I will like to learn to use molecular simulation programs that can give me a better insight of these protein dynamics.
Having done such simulations for the last several years, I can tell you that you will need many long simulations to gather sufficient data (see the papers in my profile if you are interested). AB is a difficult peptide to simulate adequately. High-performance software like GROMACS and NAMD are far better than CHARMM and AMBER, which have many features but are lower throughput. GROMACS, in my experience, is by far the fastest code out there and is very easy to use. Enhanced sampling methods like REMD are likely to be of use.
There are several programs that can do this kind of simulations, such as, CHARMM, NAMD, GROMACS, AMBER, which are the most popular. Depending of what you want to do, there exist several methods and approaches within each program. You should have a look at the documentation.
A reference for using MD simulations in studying the folding pathway is given in the following link:
www.ipcbee.com/vol23/2-CCEA2011-A007.pdf
Having done such simulations for the last several years, I can tell you that you will need many long simulations to gather sufficient data (see the papers in my profile if you are interested). AB is a difficult peptide to simulate adequately. High-performance software like GROMACS and NAMD are far better than CHARMM and AMBER, which have many features but are lower throughput. GROMACS, in my experience, is by far the fastest code out there and is very easy to use. Enhanced sampling methods like REMD are likely to be of use.
Just a brief addition to Justin's point:
If you want long simulations, you want fast code. Amber on CPUs may not be a good choice, but if you have access to GPUs to run on, than AMBER12 (pmemd.cuda) is probably the fastest out there. Just take a look at the info and their benchmarks (my own experience confirms that the code is really fast):
http://ambermd.org/gpus/
I heard from my colleagues, who are doing MD simulations on amyloid beta folding, that the simulated results can depend quite strongly on the type of program, which one is using. Perhaps one point to keep in mind could be to cross-check the results obtained by using program X with simulations using program Y.
The result of molecular dynamics simulations are strongly dependent on the force field used. There is no program that gives you the best results, although GROMACS and NAMD are the most trusted and most used. Everything depends on the experimental conditions that you impose. We recently did a test of this kind, on peptides of known structure (if the structure of what you want to simulate is not known a priori, how do you say which force field gives correct simulations ?). If you are interested this is the article: R Raucci, et al. Peptide Folding Problem: A Molecular Dynamics Study on Polyalanines Using Different Force Fields (2013) Int. J. Pept. Res. and Therap. 1-7.
This paper (Justin's) is useful. However, be aware that the experimental results on Abeta can be difficult to interpret. The NMR results referenced in the paper are likely averaged by exchange between the monomer and low molecular species of Abeta and a high molecular wt, Beta sheet species that is reversibly formed above 100 uM.
- Badges
- Science topic