[gmx-users] Protein-DNA simulation LINKS error

[Justin Lemkul]

On 6/25/15 3:43 AM, Timofey Tyugashev wrote:
>
> Looking through the paper for the FF I see that the values specified are 1.0 nm
> and 1.05 nm for cutoffs. If I used longer cutoffs than specified, then why it
> should affect the systme in a deleterious way?

Force fields behave unexpectedly when you start toying with cutoffs.  Your 
problem is probably the 8-oxoG, but the incorrect nonbond setup was worth 
pointing out.

> Anyway, now I'll try simulations with values from the paper, both for oxoguanine
> and plain guanine complexes.

Good; this will tell you where the issue really is.  It's probably the 8-oxoG 
parameters.

> Is the dodecahedron box size of 1.3 sufficient for this cutoffs, or should I
> pick even bigger one just to be on the safe side?

Probably good as-is.

-Justin

>> Message: 5
>> Date: Wed, 24 Jun 2015 07:44:41 -0400
>> From: Justin Lemkul<jalemkul at vt.edu>
>> To:gmx-users at gromacs.org
>> Subject: Re: [gmx-users] Protein-DNA simulation LINKS error
>> Message-ID:<558A9829.4030803 at vt.edu>
>> Content-Type: text/plain; charset=utf-8; format=flowed
>>
>>
>>
>> On 6/24/15 7:41 AM, Timofey Tyugashev wrote:
>>> >24.06.2015 15:58,gromacs.org_gmx-users-request at maillist.sys.kth.se  ?????:
>>> >
>>>> >>------------------------------
>>>> >>
>>>> >>Message: 5
>>>> >>Date: Wed, 24 Jun 2015 05:58:47 -0400
>>>> >>From: Justin Lemkul<jalemkul at vt.edu>
>>>> >>To:gmx-users at gromacs.org
>>>> >>Subject: Re: [gmx-users] Protein-DNA simulation LINKS error
>>>> >>Message-ID:<558A7F57.7000606 at vt.edu>
>>>> >>Content-Type: text/plain; charset=windows-1252; format=flowed
>>>> >>
>>>> >>
>>>> >>
>>>> >>On 6/24/15 5:20 AM, Erik Marklund wrote:
>>>>> >>>Hi Timofey,
>>>>> >>>
>>>>> >>>If I recall correctly, there is no constrainttype defined for the pair MCH3
>>>>> >>>CT in ffbonded.itp, which is needed for making a 4out virtual site for the
>>>>> >>>methyl group in thymine. You need to work out the geometry that yields the
>>>>> >>>correct moment of inertia for the methyl group and define a constraint type
>>>>> >>>accordingly.
>>>>> >>>
>>>> >>There are virtual sites in this system?  I saw no mention of this; have I
>>>> missed
>>>> >>something?
>>>> >>
>>>>> >>>Kind regards,
>>>>> >>>Erik
>>>>> >>>
>>>>> >>>
>>>>>> >>>>On 24 Jun 2015, at 06:38, Timofey Tyugashev<tyugashev at niboch.nsc.ru>
>>>>>> wrote:
>>>>>> >>>>
>>>>>> >>>>
>>>>>> >>>>I have ordinary DNA and an oxoguanine site? What actions should I take to
>>>>>> >>>>correct these vsite parameters?
>>>> >>If you're using virtual sites, back up and don't do that.  Solve simple
>>>> issues
>>>> >>before doing harder things.  If you have a non-standard residue like 8-oxoG,
>>>> >>then perhaps its parameters are insufficient.  Try running with a "normal"
>>>> DNA
>>>> >>(e.g. G instead of 8-oxoG) and see if that works first, before doing
>>>> something
>>>> >>atypical.
>>>> >>
>>>> >>-Justin
>>> >I haven?t planned to use virtual sites for this simulations yet. A short
>>> >calculation with normal guanine worked fine, one with longer time is still in
>>> >queue for calculation, so I'll definitely check it later.
>>> >Looking through the literature I've encountered different cutoff values, mainly
>>> >from 1.0 to 1.4 for forcefields, so I simply picked the middle one of 1.2. How
>>> >bad it is?
>> Read the AMBER99sb-ILDN paper.  Use those values.  You can't pick some arbitrary
>> value in the ballpark of what different models use.  That doesn't make sense.
>> Cutoffs are part of the force field; treat them as such.
>>
>> Your problem almost certainly comes from an inaccurate 8-oxoG topology if normal
>> guanine works fine.
>>
>> -Justin
>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================