Tag Info

15

Before talking about the pros and cons of ML-potentials, there is a huge conceptual difference between empirical- and ML-potential that needs to be clarified. In empirical potentials, one uses data from experiments to find the parameters of a fixed functional form that would explain the experiment. This is an inverse problem which is mathematically ill-posed ...

15

(Expanding my comment into an answer) When ML-based forcefields are compared to classical forcefield directly, I think we miss the most important points. ML-based models have several advantages: They do not need an a priori "correct" description of the system, nor are they limited by the applicability of specific theories to your system. Classical force ...

15

ForceBalance ForceBalance is one approach which aims at making it easier to create force fields from a combination of theoretical and experimental data. I believe it's within the OpenMM framework. The paper on the work can be found here https://pubs.acs.org/doi/10.1021/jz500737m From the GitHub page it says "The purpose of ForceBalance is to create ...

15

What is a force field? The Wikipedia entry on this is a good resource, but I'll give my own description below. In the context of molecular dynamics (MD), a force field is one way of describing the interactions between atoms. In classical MD, the motion of atoms is determined by the instantaneous forces acting on the atoms (i.e., we need forces in order to ...

11

Since you seem to have at least a start on automating the process of modifying the file, I will just address the other two parts of your question: the format of the nonbond_params section of a GROMACS .itp file and the meaning of the symbols used in the Martini force field. GROMACS format This is given in the GROMACS manual. To summarize: Atom1 Atom2 Func ...

11

We attempted to solve a similar problem when studying the (also highly symmetrical) $\ce{CH4}$ and $\ce{CF4}$ homo- and hetero-dimers. I found it easiest to use internal coordinates and fixed molecular geometries to generate the hypersurface, but we also had large computational resources at our disposal and we might have benefitted from using even more, ...

11

We performed some timing benchmarks as part of our recent paper, albeit not on molecular dynamics: "Assessing conformer energies using electronic structure and machine learning methods" Int J Quantum Chem. 2020; 121:e26381 It was a bit controversial, since we compared single-core CPU times and not in batch mode. Once the ML method runs the model, ...

10

I have looked for a similar set of benchmarks and haven't found one, unfortunately. Hopefully the data referred to in the comments comes through! In the meantime, I have an MIT-licensed implementation of the Wolf-like alternative to Ewald that @PhilHasnip linked to. Here is the repository. Forces and stresses are available. You could try benchmarking against ...

10

I am not a GROMACS user, but have used it once or twice during my course. As far as I can remember and understand, it is common to use different force field in MD if one simulate something solvated in something, e.g. formaldehyde (OPLSAA force field) in water (SPC,TIP3P, etc.). I guess we can also do that in GROMACS, but I do not really know how to specify ...

9

ForceBalance Cody is right, ForceBalance is designed for the task of generating custom force fields. Force fields have a lot of variability in their functional forms, representation of parameters (as text, XML files or objects in code), the possible values or ranges of parameters, the software that runs simulations using the force field, the experimental ...

8

On the fitting, I would recommend Dr. Nike Dattani's answer. You can also try some other methods such as those using neural networks. I read your discussion and find that may be you have also some issues with the scanning using ab initio methods. The intermolecular interaction is often read $$\Delta V_{\rm int}=\Delta V_{\rm int}(R, \theta, \phi),$$ where $... 8 For 1D analytic potentials, I highly recommend to use my potential energy form or a slight simplification of it, which I can help you with. This is known as the Morse/Long-Range (MLR) potential and is now the "industry standard" in high-accuracy spectroscopy and some other areas. For analytic multi-dimensional potential energy surfaces, there's a ... 8 Running simulations with small molecules is one thing that I found quite difficult especially because most online guides/tutorials are focused on proteins. So, here's a rough idea of what to do: Getting forcefield parameters: Your question is about this, so I will give a detailed explanation. There are many forcefields to choose from, e.g. MMFF94, UFF, ... 8 Allow me to begin with a disclaimer: what you have asked is a very contentious question among people who work on water models. I think much more contentious than it should be. Not necessarily because there's an obvious answer, but because there are many people who work on water models who really believe in their own model. The problem is that water is ... 8 The Axilrod-Teller force is not a "pairwise" force but a "3-body" force whose potential has the formula: $$\tag{1} V_{ijk}= E_{0} \left( \frac{1 + 3 \cos\gamma_{i} \cos\gamma_{j} \cos\gamma_{k}}{\left( r_{ij} r_{jk} r_{ik} \right)^3} \right),$$ as opposed to a pairwise force like the Van der Waals force which is commonly used in ... 8 I believe I have figured it out. For those trying to martinize their all-atom files into coarse grain representations for martini 3, use the following package: https://github.com/marrink-lab/vermouth-martinize With respect to modifying the water for IDP simulations, I pulled out the W, SW and TW interactions from the martini v3.0.0 forcefield, like so: W ... 8 Generally, if you want to perform simulations with some force field, you will have to search the literature for published FFs tailored for your problem. The Stillinger-Weber (SW) model is very popular to describe bulk (diamond) silicon, for which it was designed in the first place. But several reparametrizations have also been published. If you happen to ... 8 Changing the force field parameters is not a good/recommended approach. This is due to the high number of parameters you have to know. Many of them, you cannot obtain from experimental data. Instead, you will need high precision methods like Density Functional Theory, Hartree-Fock, pos Hartree-Fock or semi-empirical methods to calculate them for two, three ... 7 Any parameters are better than no parameters. Part of using a FF is checking that results make sense, so if you can find something in this bare market for anions, try it. Proof will be in the pudding. LigParGen can only generate ions between -2 and +2 charge, so$\ce{PO4^3-}$is out. When I tried with$\ce{SO4^2-}$, it did not work.$\ce{H2SO4}$also did not ... 7 ASE has a FixSymmetry constraint that preserves spacegroup symmetries. It works with a variety of structure optimization algorithms. You could use LAMMPS as the engine or one of the many other calculators, including some DFT options. 7 Short answer I am a regular GROMACS user. I do not think you can just easily have the total functional form "printed out" from GROMACS. However, you can obtain it from the topology file (top and itps) and the mdp file. Let me elaborate a little below. Longer answer The functional forms you are referring to in the GROMACS top/itp file are the ones used in ... 7 The term's origin goes back to vector field which is a function that returns a vector for any given point in space (as in the image below on the left). We can almost "see" that such fields exist, by putting iron filings and a magnet on a sheet of paper (the magnetic field causes the iron filings to move accordingly): In classical molecular ... 6 I would recommend GAFF or OPLS/A Camps has outlined the OPLS FF, so I won't elaborate very much, but to say that when using LigParGen, a nice feature is that forcefields are made for many different MD programs such as openMM, Gromacs, CHARMM and others, so it is pretty easy. If you email Jorgensen, you can get a version of LigParGen that you can use on your ... 6 I think the effect of larger cutoffs will not affect your simulation. Shorter cutoffs will be more problematic and can see artifacts than the larger cutoffs. Even if you have larger cutoffs, the interaction strength will be very very weak as you go outward and eventually die. This will just increase your computational cost (neighbour list calculation) and ... 6 Molecular dynamics is based in Molecular Mechanics. Molecular Mechanics do many approximations in order to treat the atomic/molecular system. A good definition can be seen here: Molecular Mechanics is a computational method that computes the potential energy surface for a particular arrangement of atoms using potential functions that are derived using ... 6 I'd use OpenMM's openmmforcefields package, which is available on GitHub. I haven't used it extensively, but I recommended to a student in a colleague's lab and they've found it useful. Small molecule force fields: GAFF 1.x and 2.x parameters OpenFF parameters Biomolecule force fields: Amber-type (albeit not Amber ff19SB yet) CHARMM-type And it will, if ... 5 This is actually a straightforward problem about unit conversions and careful manipulation based on the functional forms of the potential we want. To go from the Improper Torsion Parameters to the Improper Dihedral in GROMACS, simply do the following transformation on the amplitude:$\textrm{ampl}\times 4.184\times n/2~\$ where 4.184 is the unit conversion ...

5

Pairwise force-fields neglects 3 body and higher interactions (many body). As Nike pointed out, there are force-fields that can take into account higher order interactions and still be somewhat pairwise, but parameterizing them is hard(er). It is easier to parameterize a simple model, and often, more accurate. We have put alot of work into simple pairwise FF'...

5

Which force field would be the best candidate for my study case considering the physical system? The Optimized Potentials for Liquid Simulations (OPLS) force field can be a possible solution. It was developed with liquids in mind. The main developer is professor William Jorgensen, from Yale University. which one is more easy to use when it comes to infer ...

5

GULP GULP (General Utility Lattice Program) uses specific forcefields for different type of systems and it is designed to work with periodic conditions. Also, it has defined several type of potential models for two-body (Buckingham, Lennard-Jones, Morse, etc.), three-body (Three-body harmonic, Axilrod-Teller, Stillinger-Weber, etc.), four-body, six-body and ...

Only top voted, non community-wiki answers of a minimum length are eligible