File Formats
1. caDNAno (.json) file
JSON is the input design file format used for CanDo analysis exported from caDNAno (http://caDNAno.org), which is an open source program that offers rapid prototyping of DNA origami connectivity maps.
2. caDNAno sequence (.csv) file
CSV is the input sequence file format exported from caDNAno.
3. BILD format
BILD is a text file that lists commands for lines, polygons, and geometric primitives used to generate a VRML model in UCSF Chimera. CanDo provides results to users for download including the 3D solution shape, thermal fluctuations of the solution shape, and lowest normal mode shapes using this format where DNA double helices are represented by cylinders. VRML models can be exported from UCSF Chimera for further uses in other visualization programs. More information about the format is available at http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/bild.html.
4. PDB format
PDB is a text file that lists for each atom its serial number, name, residue sequence number, residue name, chain identifier, and Cartesian coordinates (in Ångströms). CanDo provides a PDB file to users for download if an atomic model of the uploaded caDNAno design is included with the results. A PDB file can be the input of further full atomistic simulations such as molecular dynamics (MD) simulation. More information about the format is available at http://www.wwpdb.org/docs.html.
Tutorial
1. Preparing an input design file using caDNAno
In this tutorial, we will use the .json file for a 53 basepair long two-helix bundle design where three insertions and deletions exist in neighboring helix. The corresponding .json file can be downloaded here.
2. Preparing an input sequence file using caDNAno
Users may skip this step if they do not want an atomic model included with their results. Sequence information of the two-helix bundle is assigned using the “Add Sequence Tool” function in caDNAno. In this tutorial, we use the standard M13mp18 sequence. 
The sequence information of each staple strand is then exported in CSV format using the “export oligos as *.CSV” function in caDNAno. The corresponding .csv file can be downloaded here.
3. Filling out the submission form
a. Click the red box (Submit a caDNAno file for analysis…) to expand the submission form
b. Enter user information (Name, Affiliation, and E-mail address)
User information provided will be used only for internal evaluation of the resource by counting the number of unique users and the number of unique affiliations. No other information will be stored on the server. Provided E-mail address will be used to send CanDo notifications when the analysis is queued, started, and completed. If not provided, no notification will be sent.
c. DNA geometry
Default values for average B-form DNA geometry are pre-entered. Alternatively, users may enter their own values.
d. DNA mechanical properties
Default values for average B-form DNA mechanical properties are pre-entered. Users may enter their own values. Nicks are modeled by reducing backbone bending and torsional stiffness by a factor of 100 by default (corresponding to the default nick stiffness factor, 0.01) whereas stretching stiffness is retained at double-helix values. It is not recommended to use a nick stiffness factor less than the default value as it may result in much slower or no convergence of the analysis.
e. Model resolution
CanDo analysis is performed at the coarse model resolution by default. Users have an option to use the fine model resolution that computes the shape and flexibility at a single basepair resolution. However, the use of the coarse model is appropriate to obtain quick feedback for initial designs as it significantly reduces the computation time. Here we choose the fine resolution for purposes of this tutorial.
f. caDNAno (.json) file
Browse to the location of your caDNAno design file.
g. Lattice type
Users must choose the lattice type, either honeycomb or square.
h. Choose whether movies visualizing thermal fluctuations of the solution shape will be included with the results
If choosing "Yes" in this item, CanDo will compute thermal fluctuations of the DNA nanostructure at temperature 298K and visualize the result as movies.
i. Choose whether the atomic model of the solution shape will be included with the results
CanDo offers a new option for users to generate an atomic model of the solution shape of DNA nanostructure. Users can choose to generate an atomic model or compute the thermal fluctuations as well.
CanDo analysis must be performed at the fine model resolution to generate the atomic model. If the atomic model is to be included in the results, make sure that the fine resolution is selected in step e. If thermal fluctuations of the atomic model are to be included, users must choose "Yes" in item h.
j. (Optional) caDNAno sequence (.csv) file
Browse to the location of your caDNAno sequence file. This item does not appear if choosing "No" in step i.
4. Submitting the form
Pressing the submit button will automatically bring users to the result page.
5. Viewing the results
Once the CanDo analysis is completed, users will receive results in ZIP files attached in separate emails sent to the email address they filled in the submission form.
a. Solution shape of DNA structures
Results in the BILD file format can be viewed using the UCSF Chimera.
b. Thermal fluctuations of solution shape
If users choose “Yes” in item “Would you like a movie included with your results?” in the submission form, users will receive a second email with a ZIP file containing movies (fluctuations_view1.avi, fluctuations_view2.avi, and fluctuations_view3.avi) showing thermal fluctuation in three orthogonal views. The structure in the movies is colored according to root-mean-square-thermal fluctuations.
c. Atomic model
If users choose to generate an atomic model in item “Would you like the atomic model included with your results?” in the submission form, users will receive a third email with a ZIP file containing the atomic model in PDB format as well as its images in three orthogonal views. The color of each strand in the atomic model is the same as in the caDNAno design (.json) file uploaded in the submission form. If users choose to visualize thermal fluctuations, movies in three orthogonal views are also included in the ZIP file.
Modeling Guide
1. Crossovers between neighboring helices
CanDo assumes that all crossovers between neighboring helices in the design are located at their natural positions defined in the caDNAno at which torsional mismatch between neighboring helices is minimal. These natural crossover positions can be seen in the caDNAno design panel when a strand is clicked as shown in the figure below.
Currently, users must follow this crossover rule to design DNA origami structures for CanDo analysis. Below is an example two-helix bundle design that follows the crossover rule whose 3D solution shape is expected to be straight.
If a design does not follow the crossover rule, unexpected results may be obtained from the CanDo analysis. For example, in the modified two-helix design below, the right scaffold crossover is enforced at the position that is three basepairs left from its natural position as highlighted in the red box. Although there is torsional mismatch between basepairs connected by this crossover, expected to result in left-handed twist, CanDo predicts the straight solution shape because all crossovers are treated as natural ones.
To simulate the effect of this mismatch properly, the design should be modified using insertions or deletions with natural crossovers as below. Three deletions at each helix are added to the design while using natural crossovers only. CanDo predicts left-handed twist on the right half of the bundle correctly as expected.
Two more example design modifications are shown in the figure below where red dashed lines indicate natural crossover positions.
An algorithm to account for the mismatch due to unnatural crossovers is currently under development.
2. Crossovers between non-neighboring helices (Distant crossovers)
In wireframe structure design, crossovers between non-neighboring helices are used to connect substructures that are initially distant to each other but should be adjacent in the final relaxed configuration. CanDo defines these crossovers as distant crossovers that gradually shrink during the analysis until their length reduces to the distance between neighboring helices or the helix diameter. However, relative orientations of helices connected by distant crossovers remain unconstrained so that CanDo analysis is limited at present to on-lattice modeling of DNA origami structures. Thus, careful use of distance crossovers is advisable.
To illustrate this, let’s consider three test designs of four-helix bundle that have the same connectivity map. The first design consists of four neighboring helices with crossovers at the natural positions. Because all basepairs are aligned at crossovers, the deformed shape (right) is the same as the initial layout (left).
In the second design, the entire structure is divided into two sub-bundles (helices 0-1 and helices 2-3) separated vertically in the initial layout. Two sub-bundles are connected by distant crossovers between helix 1 and helix 2 at the same axial positions used in the first design so that basepairs connected by these distant crossovers are already torsionally aligned (highlighted with blue dashed lines). Hence the same deformed shape is obtained as in the first design.
In the last design, helices 2-3 are placed in an arbitrary position. Basepairs connected by distant crossovers between helix 1 and helix 2 are not torsionally aligned. In addition, all crossovers between helix 2 and helix 3 violate the natural crossover rule. In this case, the deformed shape predicted by CanDo will be different from that of the first and second designs.
We are currently developing a computational model to allow for more general crossovers, enabling lattice-free DNA origami design. Stay tuned!
3D Printing Guide
The 3D printing service we use is Shapeways (https://www.shapeways.com/) where users will need to make an account. Once having an account, users can upload a 3D object and decide on the material they want it printed with. For full color users will be limited to their Sandstone material, though there are many other plastics, ceramics and metals to choose from if users want something more robust and are not interested in full color. Once the material has been picked, it is straightforward to order, though only after certain tests have been run on the structure. A 3D printed model must have certain properties for it to be printed. e.g. any mesh must be water-tight. There are many tutorials on making sure the structure is ready (https://www.shapeways.com/tutorials).
Two programs are required to generate 3D printable models from CanDo:
1. UCSF Chimera (https://www.cgl.ucsf.edu/chimera/)
Chimera is used to read either the PDB files generated by CanDo, to prepare them visually (add a molecular surface, change the colour, etc) and to then write the scene to a 3D format file. We use the .wrl (VRML format) file type as it is the best full-colour data format for printing with the 3D printer service we use. Users can certainly use .obj or .stl if they are not interested in detailed color.
2. A CAD manipulation software
We recommend AutoDesk Maya which is free under an academic license (http://www.autodesk.com/education/free-software/maya). Users can also use the freeware Blender (http://www.blender.org/). The 3D printing service Shapeways also recommends Netfabb (http://www.netfabb.com/).
Maya/Blender is used to scale the size of your structure and to make alterations. Advance usage includes adding ambient occlusion shadings to users’ structures.
A 3D-printed model of DNA assembly is presented below:
