Towards firm foundations for the rational design of RNA molecules

International audienceThe overall architecture of an RNA can be predicted with acceptable accuracy using established computational methods, such as those implemented in the MFold or the RNAFold software. This opens the door for a rational design of RNA molecules, to serve as building blocks in synthetic biology, or as novel therapeutic agents. To that purpose, one must solve the inverse folding problem, i.e. construct an RNA sequence which adopts a specific secondary structure according to established structure prediction methods. However, contrasting with the prediction task, the underlying computational structure of inverse folding is poorly understood, and existing tools either fail to return admissible sequences for some structures, or require extraordinary computational resources. Consequently, the field of RNA bioinformatics is currently craving for both practical and theoretically-sound approaches for the problem. In this talk, I will introduce the many flavors of the RNA design problem and illustrate its relevance to the objectives of modern (synthetic) biology. I will outline the strengths and shortcomings of selected approaches, and will briefly present recently contributed methods, developed with collaborators at Univ McGill (Canada), Wuhan Univ. (China) and Univ. Paris-Sud. If time allows, I will also mention results recently obtained with collaborators at Simon Fraser University, which characterize large subclasses of (un)designable secondary structures in simple energy models. These results are currently being extended to more realistic energy models, and could serve as a foundation for exact, computationally-efficient, solutions for RNA design