Determining the ideal shapes of molecules is a primary concern in computational chemistry. These shapes provide predictive information concerning the behavior of individual molecules as well as their behavior when interacting with other molecules.
When modeled alone (as in a vacuum), a molecule's shape gives some insight into its chemical stability and pharmacological properties. When molecules are modeled together, their shapes predict the ways they may interact and the potential strength of those interactions. When a small molecule (such as a pharmaceutical compound) is modeled alongside a very large molecule (such as a protein), their shapes predict whether the small molecule will stick, or 'bind', to the larger molecule, and how strong that binding might be.
A molecule's shape is governed by a complex system of forces which exist between each of the atoms' subatomic particles. That said, it's much easier to view these forces as a collection of springs and magnets. Chemical bonds (springs) can, within limits, be stretched, compressed, and bent. Partially charged atoms (magnets) attract and/or repel other partially charged atoms. The ideal shape is one where the cumulative stress of these stretch, bend, push, and pull forces has been reduced as much as possible; this process is normally referred to as energy minimization.
The video below illustrates the simple case of minimizing these forces in naphthalene to determine its ideal shape. For the sake of the example, all of its atoms were placed at the same location (an impossibly high-energy state), and then allowed settle into positions where the molecule's total energy has been reduced as much as possible. This minimization was performed via an algorithm developed internally at Collective Scientific; the movie was rendered using PyMOL.
From a computational standpoint, determining the ideal shape of naphthalene is relatively simple. However, minimization becomes increasingly complex when a molecule has many ideal or semi-ideal shapes, when the molecule is very large, or when multiple molecules are being simultaneously modeled.
The next installments in this energy minimization series will cover topics such as minimization of large molecules and molecular binding. Sign up for the Collective Scientific email list to be notified when they become available.