Coding Agents as a Mechanism for Formalizing and Transferring Domain Knowledge in DNA Origami Design
Fu, D.; Ke, Y.
Show abstract
This work investigates the use of a coding agent to design and manipulate DNA origami structures through the caDNAno2 design interface. The design methodology produced by the agent is saved as transferable, repeatable text-file instructions interpretable by another coding agent, converting informal domain expertise into formalized, distributable protocols. We demonstrate this through a parametric pipeline for simple single and multi-layer designs, identifying several failure modes that each required explicit domain knowledge from the human designer. Once encoded, this knowledge enabled the agent to autonomously extend designs to arbitrary lattice dimensions, generate structures to physical specifications, execute targeted edits from natural language prompts, and run the full pipeline from design through molecular dynamics simulation.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Accelerated design of Escherichia coli genomes with reduced size using a whole-cell model and machine learning surrogate 92%
- FlowDesign: Improved Design of Antibody CDRs Through Flow Matching and Better Prior Distributions 91%
- Sequence-based prediction of protein-protein interactions: a structure-aware interpretable deep learning model 91%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- The Synthesis Success Calculator: Predicting the Rapid Synthesis of DNA Fragments with Machine Learning 93%
- Quantitative measurement of synthetic repression curves reveals design challenges for genetic circuit engineering under growth arrest 92%
- CloneCoordinate: Open-source software for collaborative DNA construction 92%
Similar papers in this journal
- RatInABox: A toolkit for modelling locomotion and neuronal activity in continuous environments 91%
- pyControl: Open source, Python based, hardware and software for controlling behavioural neuroscience experiments. 91%
- Artistoo: build, share, and explore simulations of cells and tissues in the web browser 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.