Growth of equilibrium structures built from a large number of distinct component types

TL;DR

This paper investigates how equilibrium structures composed of many distinct components can grow with high fidelity despite undesigned interactions, using analytic and simulation methods to inform design strategies for self-assembly.

Contribution

It demonstrates that high-fidelity growth is achievable with substantial undesigned interactions if designed interactions are properly tuned and narrowly distributed, providing practical insights for DNA-based self-assembly.

Findings

High-fidelity growth occurs despite substantial undesigned interactions.

Narrow distribution of designed interactions enhances assembly fidelity.

Guidelines for selecting DNA sequences to optimize multicomponent self-assembly.

Abstract

We use simple analytic arguments and lattice-based computer simulations to study the growth of structures made from a large number of distinct component types. Components possess 'designed' interactions, chosen to stabilize an equilibrium target structure in which each component type has a defined spatial position, and 'undesigned' interactions that allow components to bind in a compositionally-disordered way. We find that high-fidelity growth of the equilibrium target structure can happen in the presence of substantial attractive undesigned interactions, as long as the energy scale of the set of designed interactions is chosen appropriately. This observation may help explain why equilibrium DNA 'brick' structures self-assemble even if undesigned interactions are not suppressed [Ke et al. Science 338, 1177 (2012)]. We also find that high-fidelity growth of the target structure is most probable when designed interactions are drawn from a distribution that is as narrow as possible. We use this result to suggest how to choose complementary DNA sequences in order to maximize the fidelity of multicomponent self-assembly mediated by complementary DNA interactions. We also comment on the prospect of growing macroscopic structures in this manner