[gmx-users] Protein-DNA simulation LINKS error

[Justin Lemkul]

On 7/1/15 8:08 AM, Timofey Tyugashev wrote:
> Thanks for the advice, Justin and Mark.
>
> After rerunning the simulation with guanine in the DNA I've found no noticeable
> problems.
> Now I'm looking for another possible sources for mistakes.
> 1.The input PDB file is formatted using amber conventions, i.e. DG5 and DG3 for
> terminal guanosine  residues, CYX for S--S linked cystine and so on. Could it
> result in any erroneous interpretations by pdb2gmx?

pdb2gmx throws obvious errors if there is a problem.  AMBER uses unique 
nomenclature for termini, which is consistent with what you describe here.

> 2. I think I correctly created 8-oxoguanine entry in amber force field files by
> copying and modifying existing one for normal guanine, but I want to re-check
> myself just to be sure. Is there anything more in-depth available than rather
> laconic instruction from the GROMACS site? I've tried searching through the
> maillist, but looks like nothing similar came up (It's a pity that no forum
> exist, looking for past answers would be far simpler).

Generating force field parameters is well beyond the scope of what is or should 
be included in GROMACS documentation.  We provide a few resources in terms of 
the implementation, but given the vast scope of this topic, and the expert 
knowledge required, it is impossible to teach people to derive parameters 
suitably.  For that, you need to read and understand the literature regarding 
the parametrization theory and methods of your chosen force field.  At least 
you've arrived at your answer - you have sub-optimal parameters for 8-oxoG that 
cause a crash.  Now the task is to derive parameters of suitable quality.

> 3. Could an accidental simultaneous execution of two or three jobs on one
> workstation (scheduling script was set incorrectly once) cause LINKS errors?

No.  Independent mdrun process do not affect each other.

-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

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