- This event has passed.
October 15 @ 2:00 pm - 4:00 pm MDT
Title: In-Silico Sequence Optimization for the Reproducible Generation of DNA Structures
Program: Doctor of Philosophy in Materials Science and Engineering
Advisor: Dr. William Hughes, Materials Science and Engineering
Committee Members: Dr. Bernard Yurke, Materials Science and Engineering, Dr. Jeunghoon Lee, Chemistry and Biochemistry, Dr. Igor Medintz, Materials Science and Engineering, and Alexander Liddle, Materials Science and Engineering
Biologically, deoxyribonucleic acid (DNA) molecules have been used for information storage for more than 1.4 million years. Today, modern synthesis tools have made it possible to use synthetic DNA molecules as a material for engineering nanoscale structures. These self-assembling structures are capable of both resolutions as fine as 4 angstroms and executing programed dynamic behavior. Numerous approaches for creating structures from DNA have been proposed and validated, however it remains commonplace for engineered systems to exhibit unexpected behaviors such as low formation yields, poor performance, or total failure. It is plausible that at least some of these behaviors arise due to the formation of non-target structures, but how to quantify and avoid these interfering structures remains a critical question.
To evaluate the impacts of non-target structures on system behavior, three co-dependent scientific developments were necessary. First, three new optimization criteria for quantifying system quality were proposed and studied. This led to the discovery that
relatively small intramolecular structures lead to surprisingly large deviations in system behavior such as reaction kinetics. Second, a new heuristic algorithm for robustly identifying systems with fit sequences was developed. This algorithm enabled the experimental characterization of newly generated systems, thus validating the design criteria and confirming the finding that almost all kinetic variation can be explained by non-target intramolecular structures. Finally, these studies necessitated the creation of two software tools; one for analyzing existing DNA systems (the “Device Profiler” software) and another for generating fit DNA systems (the “Sequence Evolver” software). In order to enable these tools to handle the size and complexity of state-of-the-art systems, it was necessary to invent efficient software implementations of the metrics and algorithm. The performance of the software was benchmarked against several alternative tools in use by the DNA nanotechnology community, with the results indicating a marked improvement. Ultimately, these developments cooperatively enabled an improved method for implementing DNA systems with kinetically uniform behaviors.