From: Kenno Vanommeslaeghe (kvanomme_at_rx.umaryland.edu)
Date: Mon Jul 28 2014 - 13:33:00 CDT
On 07/28/2014 01:46 PM, Bala subramanian wrote:
> 1) It was by mistake i wrote 2b7 instead of charmm27.
And what did you use for the protein?
> 2) The experimental work indicates no implication on the time scale of the
> conformational changes.
Then my first line of thought is that your simulation is correct and the
protein doesn't move that much in such a short timespan. Increasing the
temperature will speed things up, but the speedup associated with a 17K
increase will be all but negligible, so my answer would be: "it probably
won't fall apart, but there's nothing to gain and you'll lose some
confidence in the relevance of the results".
*If* there are experimental indications that the pore should indeed open
on such short time scales, *then* I would recommend redoing the same
simulation with CHARMM36, which essentially has both increased lipid
mobility and increased protein flexibility. Conversely, if the
experimental time scales are in the microsecond to millisecond range,
you'll need biasing and/or enhanced sampling techniques to see anything of
interest happen in a 60ns simulation. Using CHARMM36 is still
advantageous, though, as it's more accurate overall.
Oh yeah, if this is a voltage-gated channel, then it's also important to
have the right ions at the right concentrations in the system. And more
generally spoken, careful system preparation is required, especially when
running NVT and when studying membranes, which you're both doing.
This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:04 CST