How Do Particles with Complex Interactions Self-Assemble ?
Authors : Lara Koehler, Pierre Ronceray and Martin Lenz
Summary :
Cells are largely powered by molecular machines made of protein components. In disease, proteins also assemble, albeit into deleterious fibrous structures. Predicting what structure will self-assemble from a collection of proteins is difficult, however, as their interactions result from a mixture of many competing physicochemical effects. Here, we use a model of this competition to show that, despite its complexity, it tends to favor a relatively small set of large-scale structures, from fibers to liquidlike droplets to crystals riddled with holes.
We introduce a minimal model of self-assembling particles on a lattice with 21 independent interaction parameters and show that it produces a range of equilibrium aggregate morphologies in numerical simulations. We then use machine learning to show that the resulting morphologies can be grouped into a small number of categories with the same aggregate dimensionality and orientational order, and to identify a simple predictor of the outcome of the assembly process. Particles with highly asymmetric interactions can result in morphologies reminiscent of those found in proteins.
This study helps identify features that determine the emergence of self-assembled structure from a given set of interactions, which could lead to a better understanding of proteins in disease and inspire designs for artificial self-assembling nanomachines.