Difference between Force Field and topology, and other related questions

Question

I am performing some Molecular Dynamics computer simulations and I am learning about the various Force Fields (FF) such as OPLS-AA, OPLS-UA, AMBER, etc. Many things are still unclear to me. I put these questions together because I think they are strictly related.

1) What is exactly the difference between "Force Field" and "topology", since they define similar parameters?

2) Which parameters can be defined as "belonging" to the FF, and which not? It seems to me that sometimes topology files recall atoms from the force field and edit some parameters such as charge, mass, bond lengths, elastic constants.

3) Why is water model usually considered separately from the FF, and why is it not included in it?

4) When a molecule does not belong to a known topology and we have to model it, how do we establish the compatibility between our model and the FF used for the other molecules in the same simulation?

Thank you.

Answer 1

1 - "Force Field" (FF) means the set of equations that describe the interactions between atoms and the set of numerical parameters that people using those particular functions have come up with to describe some molecules (typically proteins, Nucleic Acids, lipids, and a few others). Many FFs exist because of different choices in functional forms, in the procedures adopted to obtain the numerical parameters etc etc. A topology is just the description of a molecule in terms of parameters that can be replaced by the numerical ones contained in the FF of your choice, so that the equations of the FF can be worked out numerically.

2 - see 1.

3 - Water is described separately because there are specific models for water (TIP3P and so on) that work reasonably well, so people stick to these models indipendently of their choices for the rest of the system. What do you mean by "it is not included"? I can see water parameters in all MD programs I have used. Usually, FFs are always divided by "optimisation" let's say, meaning that a set of parameters have been optimised for a particular task/family of molecules, and those are given together.

4 - if you want to model a molecule for which you don't have the parameters, then you should fit them according to the procedures that have been used to obtain the other parameters belonging to the FF (good luck). If parameters exist for a similar molecule, you can try to change those to represent your molecule.

In general, it is not a good idea to use mixed parameters, meaning e.g. AMBER for the protein, CHARMM for a DNA in the same simulation. Choose one FF and if you need parameters for some new molecule, you have all my sympathy (you can find tools to obtain these "pseudo-automatically", though).