a, We started the designs in Autodesk Maya, importing or modelling our own 3D polygon mesh object. b, The triangulation step is not mandatory because the scaffold routeing and further processing is not limited to triangulated meshes, but it is used for all structures reported here to achieve extra rigidity by triangulation. Steps c–e are implemented as a series of scripts that process the mesh exported from the 3D design software. c, All odd-degree vertices are joined by helper edges using a minimum weight perfect matching algorithm (see Supplementary Note 1). d, The re-conditioned mesh is fed to a script implementing the A-trails routeing algorithm (see Supplementary Note 1). e, After scaffold routeing, the physical relaxation model reads the routed path. Up until now, the mesh has been treated as an abstract graph; in the relaxation step, however, an input is required to set the physical size of the desired DNA rendering, that is, the user sets a scaling value to fit the mesh to the scaffold available for the folding. The relaxation simulation and length-modification scheme (described in more detail in Supplementary Note 2) will rotate and shorten/lengthen some edges to find an overall best fit to the desired 3D shape while accounting for strain between nucleotides in the vertices. The output of the relaxation/length modification optimization is a file readable by vHelix, a plug-in for Autodesk Maya. f, As the file is imported into vHelix, the user has the option of automatically positioning staple-strand break-points by stating parameters for maximum staple length and the minimum length of edges with breakpoints. Alternatively, the staple-strand breakpoints can be edited manually in vHelix after importing. g, The DNA sequences of all staple strands given a scaffold input is calculated and exported to a spreadsheet by vHelix. h, The mixing of staple strands and scaffold is done by hand but a pipetting robot could conceivably also make this last step highly automated.