[gmx-users] LINCS warnings for O-P-O-H bond

[Luciane Vieira de Mello]

Dear Anton,

Thanks for your quick reply. I have tried your suggestion: "change the
mass of the hydrogen from 1 to e.g. 4 to reduce
its 'jittery' behaviour"
But unfortunately it hasn't worked very well. I tried a short relax_md and
still got the LINCS warning including the same atoms.

Step 4062, time 8.124 (ps)  LINCS WARNING
relative constraint deviation after LINCS:
max 8.816363 (between atoms 603 and 604) rms 0.123279
bonds that rotated more than 30 degrees:
 atom 1 atom 2  angle  previous, current, constraint length
    600    601   90.7    0.1480   0.2709      0.1480
    600    603   49.1    0.1610   0.1580      0.1610
    603    604   90.2    0.1000   0.9816      0.1000

And even worse, I got a segmentation fault this time.

I am thinking of restrain the dihedral angle envolving these atoms. Do you
think it would be a good strategy?

thanks in advance.

Luciane

> > I've added a singly negatively charged phosphoserine to my GROMACS
> > dictionary.   All seemed well and the MD runs.  What worried me now are a
> > series of warnings e.g.
> [...]
> > This bond is my O-P-O-H bond in the phosphoserine.  I uncommented the
> > following line in the ffgmxbon.itp file
> >
> >     P   OA    1     0.000       1.046      3
> >
> > Is this right?  Or do we not need to worry about a dihedral for
> > A-B-C-hydrogen?
>
> In general, you *would* worry about this. In this case, it is not
> unsurprising: there is a very light, and charged atom (H) in a heavily
> charged environment (the O-P-O etc.). Because of this, it will feel
> very large forces at times, and have quite a bit of chance of starting
> to spin. Then LINCS feels it is time to warn the user.
>
> I wouldn't recommend removing the dihedral. Rather, if you need a fix,
> you could change the mass of the hydrogen from 1 to e.g. 4 to reduce
> its 'jittery' behaviour. If you're interested in longer scale (~ns)
> properties and/or averages, that should still be a-o-k.
>
> > Another strange aspect is that it is taking more than a overnight to run
> > 20ps, what usually takes a couple of hours.
>
> Writing text can be very slow. It could be the LINCS warning that slow
> you down there, perhaps in combination with writing to nfs?
>
>
> --
> Groetjes,
>
> Anton
>   _____________ _______________________________________________________
> |             |                                                       |
> |  _   _  ___,| K. Anton Feenstra                                     |
> | / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
> |(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
> | \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610           |
> |             | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/       |
> |             | "If You See Me Getting High, Knock Me Down"           |
> |             | (Red Hot Chili Peppers)                               |
> |_____________|_______________________________________________________|
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
>