It was first suggested 1 more than 30 years ago that Watson-Crick base pairing might be used to rationally design nanoscale structures from nucleic acids. Since then, and especially since introduction of the origami technique 2 , DNA nanotechnology has seen astonishing developments and increasingly more complex structures are being produced [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] . But even though general approaches for creating DNA origami polygonal meshes and design software are available 14,16,17,[19][20][21] , constraints arising from DNA geometry and sense/antisense pairing still impose important restrictions and necessitate a fair amount of manual adjustment during the design process. Here we present a general method for folding arbitrary polygonal digital meshes in DNA that readily produces structures that would have been very difficult to realize with previous approaches. This is achieved with a high level of automation of the design process, which uses a routing algorithm based on graph theory and a relaxation simulation to trace scaffold strands through the target structures. Moreover, unlike conventional origami designs built from closed-packed helices, our structures have a more open conformation with one helix per edge and are thus stable in salt conditions commonly used in biological assays.The starting point of the present method is a 3D mesh representing the geometry one wishes to realize at the nanoscale. Focusing only on polyhedral meshes, i.e. meshes which enclose a volume inflatable to a ball, and in contrast to several previous approaches 14,17,19 (see Extended Data Fig. 1) we aim to replace the edges of the mesh by single DNA double helices such that the scaffold strand traverses each of these edges once. This problem is closely related to the Chinese Postman Tour problem 22 in graph theory, which we use to find solutions as doing so by hand would be practically impossible for most meshes. The main principles underpinning our design paradigm are that the technique should allow meshes to be triangulated to optimize structural rigidity; that each edge should be represented by one double helix to enable construction of large structures using as little DNA as possible (though some meshes require two helices to render certain edges as discussed below); and that vertices should be nonBenson et. al, -DNA Rendering of Polyhedral Meshes at the Nanoscale Confidential 2 crossing (i.e. the scaffold should not cross itself in the vertices to ensure non-knotted paths with fewer topological-and kinetic traps during folding, and planar vertex junctions that avoid mesh protrusions due to stacking of crossing helices at each vertex).The overall design scheme is split into four discrete steps: i) Drawing of a 3D polygon mesh in a 3D software, Fig. 1a. ii) Generating an appropriate routing of the long scaffold strand through all the edges of the mesh, Fig. 1b-e. iii) Determining the least strained DNA helix arrangement realizing the 3D mesh, Fig. 1f-i. And iv), Optional fine tuning of the des...