Five Model Types, One Molecular System: A Practical Guide to Knowing What You’re Looking At

A molecular scene can look complete while still leaving an important question unanswered: what does each object actually do? A structure, its on-screen representation, the degrees of freedom used in a simulation, and the forces acting on it are related—but they are not the same thing. Mixing them up can make a document harder to understand and simulation setup harder to reason about.

A useful way to sort things out is to ask five different questions: What is there? How should it look? What can move? What forces act on it? What other properties should be recorded? In SAMSON, those questions correspond to five model categories.

Start with what exists: structural models

Structural models describe geometry and topology. In practical terms, they represent the molecular objects and how they are organized: atoms and bonds, as well as higher levels such as residues, chains, and molecules. If you are trying to identify the underlying molecular system, this is the category to think about first.

A useful check is to ask: if I changed how this scene was drawn, would the atoms, bonds, or organization change? If not, you are probably thinking about its visual representation rather than its structure.

Separate the scene from its appearance

Visual models provide graphical representations of the system. A protein cartoon, a surface, an electron-density isosurface, and a volumetric electrostatic-field display are examples of different ways to present information. They help you see and interpret a scene, but they are not the same as the underlying structural description.

This distinction is useful when a representation looks unfamiliar or when you want to change how a molecule is displayed without confusing that change with a change to the molecule itself. In SAMSON, visual models can be applied to selected nodes—or to the whole document when nothing is selected—through the selection context toolbar or Visualization > Add > Visual model. The shortcut is Ctrl/Cmd + Shift + V. Some visual models also expose parameters in the Inspector: select the visual model in the Document view and choose Inspect from its context menu.

Examples of default visual models in SAMSON

Depending on the visual model and selection filter, surfaces can also make it possible to highlight and select atoms, residues, or chains directly. That can be helpful when a visible surface is the easiest way to locate the part of the structure you want to work with.

For simulation, distinguish motion from forces

The next two categories are easy to conflate because they meet in simulation workflows, but they answer different questions.

Dynamical models indicate where the degrees of freedom are in structural nodes and describe the dynamical state. They store positions, momenta, and masses. For example, a particle-system dynamical model applied to a group of atoms assigns three translational degrees of freedom to each atom. When you are asking what can move, or how the system’s state is represented, think “dynamical model.”

Interaction models describe energies and forces. They provide forces corresponding to the degrees of freedom in a dynamical model. Examples include Lennard-Jones interactions, spring models, elastic network models, and universal force fields. When you are asking what energy or force acts on the system, think “interaction model.”

A simple way to remember the relationship is: the dynamical model describes the system’s available motion and state; the interaction model supplies the energy and forces associated with that motion. They are complementary, not interchangeable.

Keep other information in its own category

Property models represent properties that do not belong to the first four categories. Examples include a protein’s radius of gyration, a scalar field such as electron density, a vector field such as an electrostatic field, or a function. If a piece of information describes the system but is not its structure, visual display, dynamical state, or interaction energy and forces, property models are the category to consider.

One more piece: advancing the state

State updaters are not a sixth model category. They implement methods for advancing states during simulations. Minimization algorithms, Monte Carlo methods, and molecular dynamics algorithms are examples. Put another way, models describe aspects of the system; a state updater provides a method for taking it through successive states.

When a SAMSON document feels confusing, try identifying each item by its job rather than its appearance. Ask what the object represents, whether it changes the display, defines motion, supplies forces, stores another property, or advances a simulation state. That small classification step can make both the document hierarchy and the simulation setup easier to interpret.

For more detail, see the SAMSON models reference.

Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.

Comments are closed.