[gmx-users] Umbrella sampling and PMF calculations

[Alex]

>
>
>>
> Your situation is completely different.  Look into the cylinder settings
> for dealing with a layered system, but I have no experience with them.  To
> be clear - you always need two groups to define the vector (reaction
> coordinate) along which the bias is applied.  So you need to define a
> second group; you can't just say "I'm biasing an ion and that's it."


Okay, I see. However, I have a pretty simple situation: an ion and the rest
of the system, i.e. a membrane with all the atoms it contains, water, and a
counterion to the ion of interest -- which could be the groups in question,
after I use make_ndx properly. The simulations that result in the
pullf/pullx data of interest are using "pseudo" pull, because your pulling
rate is set to zero, so in terms of moving things, nothing actually gets
pulled and this is merely a way of obtaining the pullf data in the
direction of the bias. It sounds like there is little difference between
our systems. My topology of the ion of interest has no restraint in the
Z-direction, i.e. indeed a cylinder type of constraint in-plane. So, when
you say that my situation is completely different, what do you mean? There
is really nothing "layered" about my system...


> 2. It is my understanding that only pullx and pullf files are required for
>> WHAM, along with the tpr's for each of the MD runs, at least that's what
>> follows from the tutorial. Anything about trajectories or energy files?
>> May
>> I just clean those up in my script, or will they be needed at any point?
>>
>>
>


> Might be useful for troubleshooting or inspecting the microscopic behavior
> that gives rise to the free energy profile.  There's always useful stuff
> there; rarely do you only care about just a single number coming out of
> WHAM.
>
>
Sure. My HDD space is limited this time, hence the question on what's
necessary and what's a useful extra.

Thank you,

Alex


>