From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Mon Nov 18 2019 - 12:40:30 CST
There is no reference yet. In short, the code computes the PMF as a
solution of a Poisson equation, which is nice because it is deterministic
unlike the abf_integrate algorithm. The end result is the PMF whose
numerical gradients are as close as possible (in the least squares sense)
to the ABF estimate of the gradients.
Jerome
On Mon, 18 Nov 2019 at 19:27, Souvik Sinha <souvik.sinha893_at_gmail.com>
wrote:
> Great!
>
> So is it with NAMD 2.13?
> And Is there any reference on how it differs from the abf_integrate
> utility or how the multidimensional pmf is estimated?
>
> Thanks again for such support.
>
> On Mon, 18 Nov 2019, 23:48 Jérôme Hénin, <jerome.henin_at_ibpc.fr> wrote:
>
>> Hi Souvik,
>>
>> Colvars now includes a facility to integrate PMFs in 2d and 3d. The user
>> guide needs to be updated, that's on my TODO list. If you don't do
>> anything, the PMFs will be calculated by default.
>>
>> Jerome
>>
>> On Fri, 15 Nov 2019 at 07:01, Souvik Sinha <souvik.sinha893_at_gmail.com>
>> wrote:
>>
>>> Sorry, if I'm missing something here. As far as I know, for 2D or
>>> greater, *abf_integate* is the only available utility to reconstruct
>>> the landscape. Is there any update on that? I checked the NAMD 2.13 user
>>> guide, but couldn't find anything else than *abf_integrate * to build a
>>> 2D PMF.
>>>
>>> Souvik
>>>
>>> On Wed, Nov 13, 2019 at 4:59 AM Jérôme Hénin <jerome.henin_at_ibpc.fr>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> The PMF is significant. It is generated by solving a Poisson equation.
>>>> It is optimal in the sense that its numerical gradient is as close as
>>>> possible to the ABF gradients, given the boundary conditions. It's better
>>>> than that generated by the abf_integrate tool because the algorithm is
>>>> deterministic, and its convergence is much better behaved.
>>>>
>>>> Jerome
>>>>
>>>> On Mon, 11 Nov 2019 at 21:20, Ercalary . <ercalary20_at_gmail.com> wrote:
>>>>
>>>>> Dear NAMD community, I have the following doubt about the pmf
>>>>> generation when I apply abf in two dimensions. Is the PMF that is generated
>>>>> meaningful or I should work only with the gradient file instead?
>>>>>
>>>>> Beforehand thank you very much!
>>>>>
>>>>
>>>
>>> --
>>> Souvik Sinha
>>> Research Fellow
>>> Division of Bioinformatics
>>> Bose Institute, Kolkata
>>>
>>> Contact: 033 25693275
>>>
>>
This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:12 CST