Predictive Design of Defect Annealing within Enzymatic Reaction Networks of ATP-Fueled Non-Equilibrium Multicomponent DNA Systems
Project areas:
Bioinformatics and Theoretical Biology
Preparatory and Physical Chemistry of Polymers
Project leaders:
Walther, Andreas, Prof. Dr.
Johannes Gutenberg University Mainz
Department of Chemistry
Duesbergweg 10–14, 55128 Mainz
+49 (0)6131 39 25883
andreas.walther[a]uni-mainz.de
Gerber, Susanne, Prof. Dr.
Universitätsmedizin der Johannes Gutenberg University Mainz
Institute of Human Genetics
Anselm-Franz-von-Bentzel-Weg 3, 55128 Mainz
+49 (0)6131 39 27331
sugerber[a]uni-mainz.de
Summary
In this project we aim to break new ground in the fundamental design and in conceptual approaches for
understanding and predicting ATP-driven DNA-based reaction networks using dynamic ligation/restriction networks orchestrated through T4Ligase and restriction enzymes. As a particular challenge, we will set out to integrate structural point defects into the sticky ends of our DNA tiles, which not only influence the kinetics
(e.g. of ligation), but which also gives rise to autonomous reconfiguration of the structures. Such autonomous reconfiguration of structures would expand principles of pathway complexity known from classical soft matter into fuel-driven non-equilibrium systems and would extend autonomous systems design to new levels. At the same time, this poses a distinct challenge for reaching comprehensive and predictive mathematical models, as most elaborate designs will require multicomponent systems that cannot be measured with reasonable efforts anymore in all of their kinetic parameters. To address this challenge, we forge a team between Walther (chemistry, DNA nanoscience, systems chemistry) with Gerber (biology/medicine, systems biology, bioinformatics and machine learning) to address both experimental challenges and mathematical modelling. We will follow a strongly interconnected workplan, whereby WP1 focuses on measuring kinetics and advancing system design, WP2 focuses on formulating mathematical models for the CRNs and on rapid extension of these models using machine learning approaches, while in WP3 we aim to conduct prospective model-based verifications of system behavior that arises from extra- and interpolated machine learning spaces.