Shape-anisotropic nanocrystals and patchy particles have been explored to construct complex superstructures, but most studies have focused on convex shapes. We report that nonconvex, dumbbell-shaped nanocrystals (nanodumbbells) exhibit globally interlocking self-assembly behaviors governed by curvature-guided depletion interactions. By tailoring the local curvature of nanodumbbells, we can precisely and flexibly adjust particle bonding directionality, a level of control rarely achievable with conventional convex building blocks. These nanodumbbells can undergo long-range ordered assembly into various intricate two-dimensional superlattices, including the chiral Kagome lattice. Theoretical calculations reveal that the Kagome lattice is a thermodynamically stable phase, with depletion interactions playing a crucial role in stabilizing these non–close-packed structures. The emergence of Kagome lattices and other unusual structures highlights the vast potential of nonconvex nanocrystals for creating sophisticated architectures.

Macromolecules contain molecular units as the coding information for their correlated structures in physical dimensions. The relationship between these two features is governed by the interaction energies of the involved molecular units and their encoded sequences. We present a neural network algorithm that treats molecular units themselves as neural networks, which has the flexibility to allow each unit to respond to its own environment and to influence others in the system. Through a deep neural network and a self-consistent procedure, molecular units in the network establish a strong correlation to produce the desirable features in the physical world. The proposed framework is applied to the HP model. Both the forward problem of predicting folded structures from given sequences and the inverse problem of predicting required sequences for a given structure are examined.