Re: NANMA free energy surface vacuum map problems

From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Tue Nov 02 2010 - 08:18:30 CDT

Just to explain where I am coming from: traditionally, I would not use
CMAP for NANMA because that correction explicitly targets protein
conformation, and is knowledge-based (PDB) rather than physics-based.
So I thought its relevance in the case of NANMA was not established.

But on second thought, looking back at the CMAP papers, it does seem
to make NANMA agree better with QM/MM sampling, so it does seem to be
a good idea to include it. My bad.

Jerome

On 2 November 2010 13:23, JC Gumbart <gumbart_at_ks.uiuc.edu> wrote:
> That's a very interesting point.  I hadn't even noticed the files being
> distributed with the ABF tutorial (and FEP for that matter) don't include
> CMAP.  Looking at Tobias and Brooks (1992), their PMF is similar to the one
> in the tutorial though and they obviously didn't have CMAP.  Interestingly,
> the one in your colvars paper is a bit closer to what I found (although the
> C7ax is still slightly lower in energy).  I'll look into this further when a
> I get a chance.
>
> JC
>
> -----Original Message-----
> From: heninj_at_gmail.com [mailto:heninj_at_gmail.com] On Behalf Of Jérôme Hénin
> Sent: Tuesday, November 02, 2010 5:40 AM
> To: JC Gumbart
> Cc: Carles Corbi Verge; namd-l_at_ks.uiuc.edu
> Subject: Re: namd-l: NANMA free energy surface vacuum map problems
>
> Hi guys,
>
> Could that possibly be a problem of CMAP vs. no CMAP?
>
> Jerome
>
> On 2 November 2010 02:33, JC Gumbart <gumbart_at_ks.uiuc.edu> wrote:
>> Indeed, as you'll see in the provided png files, I saw the same thing.
>  I'm
>> attributing it to a sampling issue (taking the easy way out!), but I
> wanted
>> to be sure the simulations as given could be completed in a relatively
> short
>> time.  Probably it would be a good for you to run them longer to see if
> the
>> PMF improves relative to Fig. 5A.
>> JC
>>
>> On Nov 1, 2010, at 6:16 AM, Carles Corbi Verge wrote:
>>
>> Hi Jérôme,
>> I'm not talking about of a new simulation, I'm using the input files of
> the
>> tutorial, actually I use the example-output files to build the map. I
> merged
>> the outputs I then I run abf_integration following the instructions of the
>> tutorial. In this map  the C7ax spot have around 3 kcal/mol more than the
>> alphaL area.
>> Best,
>> Carles.
>>
>>
>> El 29/10/2010, a las 16:33, Jérôme Hénin escribió:
>>
>> Hi Carles,
>>
>> This could be a convergence issue. Is the (phi,psi) map well-sampled?
>> Did you visualize the histogram (.count file)?
>>
>> Best,
>> Jerome
>>
>> On 29 October 2010 15:03, Carles Corbi Verge <ccorbi_at_ugr.es> wrote:
>>
>> Dear NAMD users,
>>
>> I downloaded the files of the ABF tutorial, I started to follow the case
> 3,
>>
>> isomerization of N-acetyl-N'-methylalanylamide   ( NANMA ). With the
> results
>>
>> files you find in the namd website I try to reproduce the energy surface
> map
>>
>> of ABF simulations in vacuum. I merged them, then I run the
> abf_integration
>>
>> script, and The map I get is pretty different to the figure 5.a in the
>>
>> tutorial . The second minima appears in the region [48,52] instead the
> C7ax
>>
>> region [70,-80] . Is it that possible? Any suggestions?
>>
>> Best Regards,
>>
>> Carles Corbi Verge
>>
>> Departamento de Química Física.
>>
>> Facultad de Ciencias.
>>
>> Universidad de Granada.
>>
>> 18071 Granada, SPAIN
>>
>> Tel: (+34) 958 240437
>>
>> Fax: (+34) 958 272879
>>
>>
>>
>>
>>
>>
>>
>> Carles Corbi Verge
>>
>> Departamento de Química Física.
>> Facultad de Ciencias.
>> Universidad de Granada.
>> 18071 Granada, SPAIN
>> Tel: (+34) 958 240437
>> Fax: (+34) 958 272879
>>
>>
>>
>>
>>
>>
>
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:19 CST