[gmx-users] LINCS warnings and ligand topology

[Justin Lemkul]

On 5/11/15 2:24 PM, MPI wrote:
> Dear Users,
>
> I prepared  ligand topologies by two different methods, one with
> Acpype and the other with Prodrg.
>
> It is sad that independent simulations give the same (or similar)
> numerous Lincs warnings and the system blew up at the end.
>
> Errors look like
>
> LINCS WARNING relative constraint deviation after LINCS:
> rms 0.019243, max 0.893307 (between atoms 2073 and 2121)
> bonds that rotated more than 30 degrees:
>   atom 1 atom 2  angle  previous, current, constraint length
>     2073   2121   90.0    0.1758   0.1844      0.0974
> Wrote pdb files with previous and current coordinates
>
>
> I was able to trace down errors that came mainly from the "bad" atom
> indexed 2073, which is an Oxygen in a gro file like
>
> 131DRG     O3 2073   1.789   5.047   4.061
>
> 131DRG    H45 2121   1.701   5.076   4.032
>
>
> I suspect  atom 2121 is a polar H and it should be deprotonated or
> removed ??? or did I miss something ?  but the question is how to
> prepare this ligand topology without causing LINCS warnings.
>

Hacking off atoms because the simulation fails is a pretty bad strategy :)

> Please also see attached ligand coordinates. This ligand is called
> Cholesterol-Sulfate.
>

General strategy for diagnosing situations like this is found in 
http://www.gromacs.org/Documentation/Terminology/Blowing_Up#Diagnosing_an_Unstable_System

Simplify the system.  Work scientifically!

It could also be a problem with your protocol and/or .mdp files, but only by 
systematically investigating what's going on will you figure out what the 
problem is.

-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

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