Is periodicity important in MD simulations ?? if yes , why it so ? can anyone plx explain .
thanks in advance
anjibabu
Thanks for you answer. I have one more doubt, Are PBC same for QM region and MM region in QM/MM method ?? how QM and MM regions follow the PBC ?
Lulu: Does a cell or an organism have an infinite size?
---
Anji: PBC is very popular (there is no need to push water molecules back to the box for they simply appear on the opposite side of the box) but it has its problems: (i) one has to apply cut-off for non-bonded interactions; (ii) the box contains too many "unnecessary" water molecules (making the system unnecessarily too large).
The alternative is a non-periodic system (water sphere/droplet), in which the surface boundary can be treated with SCAAS:
http://scitation.aip.org/content/aip/journal/jcp/91/6/10.1063/1.456845
as implemented in the Q molecular dynamics simulation package: http://xray.bmc.uu.se/~aqwww/q/
The non-periodic system is especially suitable for free energy calculations, in which one cannot afford to discard long-range non-bonded interactions.
---
I would also consider the empirical valence bond (EVB) method as an alternative to QM/MM. It can be done with Q.
http://pubs.acs.org/doi/abs/10.1021/ja00540a008
Lulu: I think that most proteins do not float freely in the cell, including cytoplasmic proteins. The cellular environment is very crowded. Proteins interact with each other and become (partially, transiently) immobilized. How realistic is it for a protein to be surrounded by water from all sides all the time (as assumed in MD simulations using PBC)?
I am sorry for replying so late. In my opinion, I would like to answer the question from three aspects:
First, your idea is good. Without a doubt, the simulation is more realistic when the target protein is put into the crowded environment.
Second, every technique has its own limitation, so do MD simulations. The time scale and special scale that MD can reach are also limited. It is impossible to simulation the realistic cell in atomic level even in coarse grain model. Even with the development of GPU and HPC in recent years, the largest systems which MD can simulate is the HIV capsid(http://www.ks.uiuc.edu/Research/HIV/, as I have known)
The last but not the least, approximation is commonly used in theoretical study. It is reasonable to find a balance between computational capability and precision.
- Badges
- Science topic