[gmx-users] Lipid-protein force field problem.

[Mark Abraham]

> Hi to all ! :D
>
> I've been trying to work with membrane
> DPPC+peptide+waters simulation.
> Following the "example.top" from Tieleman's website,
> I've done my topology as:
>
> #include "ffgmx.itp"
> #include "lipid.itp"
> #include "dppc.itp"
> #include "protein.itp" ;(inside it has been written
> #include "ffG43a2.itp")
> #include "ions.itp"
>
> And I've found:
> "Fatal error:
> Found a second defaults directive, file "lipid.itp",
> line 9".

Read chapter five of the manual to find out what a defaults directive is
and why you can't have two of them. Then don't #include more than one file
that has a default directive. Further, don't mix force fields (such as
ffgmx and ffG43a2 above). See
http://wiki.gromacs.org/index.php/Force_Fields for more info.

> When I comment the line 9 and execute grompp again:

That might help you diagnose the problem, but it's vanishingly unlikely to
fix it...

> WARNING 1 [file "lipid.itp", line 13]:
>   Overriding atomtype LO
> .
> .
> .
> WARNING 12 [file "lipid.itp", line 24]:
>   Overriding atomtype LC2
>
> So I've decide to comment #include "lipid.itp" in the
> topology file:
>
> "Fatal error:
> Invalid order for directive defaults, file
> ""ffG43a2.itp"", line 4".

Hmmm, what do you know? The same error... See above.

> I've tried several things such as to change the order
> of includes, etc and nothing has solved the
> problem(s)!
>
> Therefore I've decided to make a force field with
> ffgmx's parameters (to describe the lipids
> interactions) +
> ffg43a2's parameters (to describe the peptide
> interactions).

Do you have a basis for knowing that these will cooperate appropriately?

> I've done the field following the
> rules:
>
> Interactions between:
> lipid - lipid: gmx
> lipid - water: gmx
> lipid - protein: gmx
> protein - water: 43a2
> protein - protein: 43a2
> lipid - ions: gmx
> protein - ions: 43a2
> ions - ions: gmx
> ions - water: gmx
> water - water: gmx
>
> In the topology file I've written:
> #include "ffG43a2mod.itp" ;(modified)
> #include "dppc.itp"
> #include "peptide.itp"
> #include "spc.itp"
> #include "ions.itp"
>
> And I've had to face this error:
> Generated 846 of the 2080 non-bonded parameter
> combinations
> WARNING 1 [file "dppc.itp", line 223]:
>   No default Ryckaert-Bell. types, using zeroes
> .
> .
> .
> WARNING 12 [file "dppc.itp", line 234]:
>   No default Ryckaert-Bell. types, using zeroes
> WARNING 13 [file "dppc.itp", line 239]:
>   No default Ryckaert-Bell. types, using zeroes
> .
> .
> .
> WARNING 24 [file "dppc.itp", line 250]:
>   No default Ryckaert-Bell. types, using zeroes
>
> I think that it happens because I didn't define the
> "bonded parameters" from ffgmxbon.itp to
> ffg43a2modbon.itp.
> In afirmative case, how do I do to convert and include
> them? Note it's not enough copy and paste! :P
>
> As you can see, I've tried to apply all tips described
> in the gmx-users list e other like to make a mixture
> gmx-43a2 force field using brute force.
>
> So please, what am I doing wrong?

Not understanding how force fields work in general, and how this is
implemented in GROMACS. You should read chapter five of the manual for the
latter, and a good textbook for the former :-)

Mark