Elezgaray, J. | bioRxiv Science

bioRxiv2025

Optimizing DNA origami assembly through selection of scaffold sequences that minimise off-target interactions

DNA origami is a mainstay of DNA nanotechnology and several efforts have been devoted to understanding how various factors of the self-assembly reaction affect the final yield of the target origami structure. This study analyses how base sequence affects origami yield through the generation of off-target side reactions during selfassembly. Off-target bindings are an under-explored phenomenon and can potentially introduce unwanted assembly barriers and kinetic traps in the origami folding pathway. We developed a multi-objective computational approach that takes a given origami design and scores different scaffold sequences (and their complementary staples) for the prevalence of four different types of off-target binding events. Using our method on DNA origami, we can select bad regions of biological sequences (like lambda DNA phage) that, when used as origami scaffold sequences, have an excessive number of off-target side reactions for each shape. We show, using high-resolution atomic force microscopy (AFM), that these scaffold sequences largely fail to fold into the target triangle or rectangle structure in vitro, despite the scaffold sequence having a fully complementary staple set present. Conversely, using our method we can also select good regions of biological sequences. These sequences are deficient in off-target reactions and when used as origami scaffolds, fold more successfully into their target structures as characterised by AFM. These results have been validated in "blind" folding experiments at two different laboratories in which the experimenters did not know which scaffolds were good or bad folders. To further investigate assembly behaviour, optical tweezers experiments revealed distinct mechanical response profiles, correlating with scaffold-specific off-target interactions. While variants with higher GC content show a high mean unfolding force, variants with lower off-target binding demonstrated more uniform force-extension curves. Our analysis confirmed that high off-target binding leads to increased structural heterogeneity, as seen in the clustering behaviour of unfolding traces of OT experiments. Over-all, our work demonstrates how the off-target reactions implicit in base sequences can derail the origami self-assembly process if sufficiently prevalent, and we provide a software tool to select scaffold sequences that minimise off-target reactions for any DNA origami design.

bioengineering↗

bioRxiv2023

Reverse Engineering DNA Origami Nanostructure Designs from Raw Scaffold and Staple Sequence Lists

Designs for scaffolded DNA origami nanostructures are commonly and minimally published as the list of DNA staple and scaffold sequences required. In nearly all cases, high-level editable design files (e.g. caDNAno) which generated the low-level sequences are not made available. This de facto raw sequence exchange format allows published origami designs to be re-attempted in the laboratory by other groups, but effectively stops designs from being significantly modified or re-purposed for new future applications. To make the raw sequence exchange format more accessible to further design and engineering, in this work we propose the first algorithmic solution to the inverse problem of converting staple/scaffold sequences back to a guide schematic resembling the original origami schematic. The guide schematic can be used to aid the manual re-input of an origami into a CAD tool like caDNAno, hence recovering a high-level editable design file. Creation of a guide schematic can also be used to double check that a list of staple strand sequences does not have errors and indeed does assemble into a desired origami nanostructure prior to costly laboratory experimentation. We tested our reverse algorithm on 36 diverse origami designs from the literature and found that 29 origamis (81%) had a good quality guide schematic recovered from raw sequences. Our software is made available at https://revnano.readthedocs.io.

bioinformatics↗