Re: GBIS CUDA not yet compatible with fixed atoms

From: Siva Dasetty (sdasett_at_g.clemson.edu)
Date: Fri Sep 05 2014 - 18:01:13 CDT

Thanks to you both for the reply.

As Aron pointed out graphene being not charged will not have any
electrostatic interactions with either protein or solvent and I dont see
why GBIS (all equations will be 0 when charge=0) has a role here. In my
case I am just interested in the interactions between graphene-protein,
protein-protein and protein-solvent. Having said that, I am also not
calculating forces between fixed atoms which is the reason I didnt turn on
the fixedAtomForces=on in the first case.

So, if fixedAtomForces =on can solve the problem and if I can run the
simulation using GPU's, I will be still doing too many unnecessary
calculations and which leads to more computational time which I thought I
can get rid,if I exclude energy calculations for some groups or part of the
whole system.

Its true that SASA calculations are important for my graphene - implicit
solvent and protein - implicit solvent, but however if I just want to turn
off for one particular group (lets say graphene-solvent), can it be done?

Can you please explain how this is done?

Thanks,
Siva

On Fri, Sep 5, 2014 at 5:50 PM, Aron Broom <broomsday_at_gmail.com> wrote:

> I don't think you can turn it off for one particular portion of the
> system.
>
> Keep in mind also, that while it may not be charged, and hence the GBIS
> portion may not be needed, if you do not have solvent, you may still want
> the SASA component for the correct hydrophobic effect (which, by contrast
> to GBIS, would probably be very important for your graphene surface).
> Though it's been pointed out to me that implicit solvent calculations, by
> virtue of not having water molecules creating interactions, actually
> already reproduce a significant portion of the hydrophobic effect without
> the need for the SASA correction (which I believe you can turn off
> separately from running GBIS).
>
>
> On Fri, Sep 5, 2014 at 4:24 PM, Siva Dasetty <sdasett_at_g.clemson.edu>
> wrote:
>
>> Thank you for the reply. I forgot to add, but this note actually made me
>> realize that I don't need to do GBIS/SASA calculations for my graphene
>> surface as 1. it is not charged and 2. it is unnecessary.
>> So Is there any way that I can exclude GBIS for a particular group in
>> NAMD? In this case, exclude GBIS calculation for graphene and include for
>> the protein.
>>
>> Thanks,
>> Siva
>>
>>
>> On Fri, Sep 5, 2014 at 4:16 PM, Jim Phillips <jim_at_ks.uiuc.edu> wrote:
>>
>>>
>>> This message only applies to the GBIS implicit solvent code. The NAMD
>>> CUDA code for explicit solvent does support fixed atoms.
>>>
>>> I don't understand why this is and error. At the very least GBIS should
>>> be compatible with fixed atoms if fixedAtomsForces is set. I've revised
>>> the error test to automatically enable fixedAtomsForces and this should be
>>> available in the next nightly build. Please test tomorrow.
>>>
>>> Jim
>>>
>>>
>>>
>>> On Fri, 5 Sep 2014, Siva Dasetty wrote:
>>>
>>> Dear All,
>>>>
>>>> I am running a test simulation to check the performance of GBIS with
>>>> CUDA.
>>>> My system involves graphene which is fixed and protein in implicit
>>>> solvent.
>>>> However when I try to run I get the following error " GBIS CUDA not yet
>>>> compatible with fixed atoms".
>>>>
>>>> Does that mean, NAMD cuda doesnt support fixed atoms yet or am I doing
>>>> something wrong?
>>>>
>>>> Thanks,
>>>> Siva
>>>>
>>>>
>>
>>
>> --
>> Siva Dasetty
>> PhD Student
>> Department of Chemical and Biomolecular Engineering
>> Clemson University
>> Clemson, SC 29634
>> USA
>>
>
>
>
> --
> Aron Broom M.Sc
> PhD Student
> Department of Chemistry
> University of Waterloo
>

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:12 CST