Recent advances in fluorescence in situ hybridization and three-dimensional microscopy have revealed a high degree of large-scale order in the nucleus, indicating that the position of each gene within the nucleus is not random. As with any other biological phenomenon, this large-scale organization must ultimately be specified by molecular interactions. Biochemical and molecular investigations have revealed a small set of local molecular-scale interactions that can be used together in a combinatorial fashion to establish a global large-scale nuclear architecture.