Simple and accessible analysis and iterative design of DNA and RNA nanostructure simulation

Analysing coarse-grained simulations of DNA and RNA nanostructures is now much easier. With the new oxView tool published by the Šulc lab, oxDNA simulations are now more accessible to experimental groups, allowing for iterative design and evaluation of structures.

The article has been published in the journal Nucleic Acid Research. Follow this link to read the full article or read Joakim’s post below for a shorter version.

A need for new tools

DNA and RNA nanotechnology is becoming an increasingly popular method of manufacturing structures on the nanometre scale, with possible applications ranging from targeted drug delivery to nanoelectronics. Especially successful is a method called DNA origami, where a long and single-stranded DNA scaffold can be made to fold into any desired structure by mixing it with a cleverly selected set of shorter staple strands that stick to different domains in the scaffold, holding them together.

200528_1
Here, a DNA origami icosahedron, from Douglas, et al 2009,  is folded using three different scaffolds (blue, green and yellow) and a large number of grey staple strands. By slowly lowering the temperature, complementary domains of the strands stick to each other to form the intended structure.

Together with improved experimental techniques, software models simulating the behaviour of DNA and RNA structures have also been developed. One particularly successful model is called oxDNA.

Instead of simulating and keeping track of every single atom, oxDNA uses a coarse-grained model that only has to keep track of individual nucleotides and the forces between them. This has a great advantage in that simulations can be much larger and run much faster, while still capturing the important dynamics. The disadvantage, however, is that you can no longer easily use established all-atom software to analyze the simulation results. Thus, since the launch of oxDNA, different research groups have been developing their own analysis tools, without any unifying standard or ease of access for the uninitiated.

However, with a new publication from the Šulc lab, this is about to change. The authors introduce oxView, an online oxDNA viewer that not only provides an easy way of looking at simulations, but that also lets you assemble and edit structures before simulating them. Moreover, the publication includes a set of generalized analysis scripts for calculating structure flexibility, averages and more.

 

An online visualizer and editor

The oxView web app, live at https://sulcgroup.github.io/oxdna-viewer/, allows users to drag and drop oxDNA files to inspect them in 3D, step through simulations and even export videos. The code is written in instanced typescript and can render over a million nucleotides while still being responsive. You can load an example DNA wireframe structure by clicking here.

In some cases, the structure you have needs some preparation before you can simulate it. You might have multiple components, designed separately, and want to combine them into a single structure. Now, this is easily solved, since you can drag each component into oxView to arrange and connect them as you wish.

Another similar problem might be that you have a structure drawn on a lattice, in software like caDNAno, where all helices need to be parallel. This restriction can cause a configuration very far from the intended one, which will take a long time to relax in oxDNA (or which might even be topologically incorrect). For example, a tensegrity kite structure as in the figure below can be easily drawn in caDNAno with both 6-helix bundles parallel to each other, but it will take considerable simulation time in oxDNA before the structure relaxes into the intended kite shape with the helix bundles at a 90-degree angle to each other. Thus, by rotating the upper helix in oxView before starting to simulate, you will have a significantly better starting configuration.

200528_2
A tensegrity kite structure, from Liedl, et al 2010. The initial configuration (top) has both helix bundles parallel to each other, while the bottom configuration has had one bundle rotated by 90 degrees in oxView to create a better starting point for simulation. The configuration can then be improved further by positioning the single-stranded sections along the sides.

As useful as these editing features are, manually rotating and translating components can become very tedious if you have enough of them. The below structure shows the three-scaffold icosahedron seen earlier. The leftmost image is the configuration freshly converted from caDNAno, with all helix bundles parallel to each other. While you could indeed orient and position each bundle manually and still save oxDNA simulation time, it would without a doubt take a long time. For this reason, oxView includes a rigid-body dynamics mode, where each component of your structure (identified manually or through automatic clustering) acts as a rigid body, experiencing spring forces across shared backbone connections and repulsion forces between their centres of mass. Thus, you can have the components orient themselves automatically into the intended icosahedron, which will rapidly relax in oxDNA (right-most image). If you want to try these dynamics yourself, you can have a look at the examples here.

200528_3
The DNA origami icosahedron, designed in caDNAno by Douglas, et al 2009, relaxed and simulated in oxDNA. From left to right: Initial configuration when imported, individual components automatically re-arranged in oxView using rigid-body dynamics, the resulting icosahedral shape, and the structure fully relaxed in oxDNA.

Since oxView can read, write and edit simulation files, you can also edit structures after they have been simulated, modifying the design depending on the simulation results. This facilitates iterative design, without the need to return to the original design software. When you are happy with your structure in oxDNA, you can export a list of the strands and sequences to try them experimentally.

Standardised analysis

The paper also introduces a set of generalised analysis tools, available at https://github.com/sulcgroup/oxdna_analysis_tools. These python scripts can provide insight into oxDNA simulation results such as calculating structural flexibility, average structures, or the angle between selected helices. Several of the tools can also output JSON files that you can load back into oxView and visualise analysis results as a heatmap or as arrows drawn in the scene. In the figure below, you can see an oxView visualisation of bond occupancy in an RNA tile, calculated using the analysis tools.

200528_4
Bond occupancy of an RNA tile during a simulation, calculated by the newly introduced analysis scripts and visualised in oxView overlayed on the mean structure.

With these new tools, we hope more researchers can benefit from the insights of nanostructure simulations.

Written by Joakim Bohlin, PhD student in Andrew Turberfield’s research group at Oxford University

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: