The selectivity of DNA recognition inspires an elegant protocol for designing versatile nanoparticle (NP) assemblies. We use molecular dynamics simulations to analyze dynamic aspects of the assembly process and identify ingredients that are key to a successful assembly of NP superlattices through DNA hybridization. A scale-accurate coarse-grained model faithfully captures the relevant contributions to the kinetics of the DNA hybridization process and is able to recover all experimentally reported to date binary superlattices (BCC, CsCl, AlB2, Cr3Si, and Cs6C60). We study the assembly mechanism in systems with up to 10(6) degrees of freedom and find that the crystallization process is accompanied with a slight decrease of enthalpy. Furthermore, we find that the optimal range of the DNA linker interaction strengths for a successful assembly is 12-16k(B)T, and the optimal mean lifetime of a DNA hybridization event is 10(-4)-10(-3) of the total time it takes to form a crystal. We also obtain the optimal percentage of hybridized DNA pairs for different binary systems. On the basis of these results, we propose suitable linker sequences for future nanomaterials design.