
RNA (Ribonucleic Acid) is an interesting molecule that is composed of a sequence of 4 nucleotides: G, C, A, U. Compared to Deoxyribonucleic Acid (DNA), RNA is single stranded but can fold back onto itself where certain base pairs (usually G-C, A-U, and G-U) will form hydrogen bonds. The 2D structure identified by the sequence and these base pairs is referred to as RNA secondary structure. While the molecule will fold into a 3D shape, secondary structure is of interest as it tends to be preserved by evolution and function can be inferred from it. Traditional wet lab procedures to determine RNA secondary structures include Nuclear Magnetic Resonance (NMR) and X-ray Crystallography, which tend to be time consuming and expensive. Alternatively computational methods to determine RNA secondary structure can be designed. In this seminar I will motivate an approach based on representing RNA secondary structures as permutations of potential helices under certain constraints. I will then discuss an approach based on an evolutionary algorithm to search this space and return a plausible structure close to a minimum energy state, not necessarily at the global minimum energy, but biologically closer to an actual native fold. I will compare this approach with dynamic programming approaches and focus on the challenges and open problems in this area. Instead of benchmarking a wide array of algorithms I will focus more on the design advantages and challenges of different approaches. I will also briefly discuss pseudoknots and RNA visualization. The seminar aims to be motivating and explorative.

RNA (Ribonucleic Acid) is an interesting molecule that is composed of a sequence of 4 nucleotides: G, C, A, U. Compared to Deoxyribonucleic Acid (DNA), RNA is single stranded but can fold back onto itself where certain base pairs (usually G-C, A-U, and G-U) will form hydrogen bonds. The 2D structure identified by the sequence and these base pairs is referred to as RNA secondary structure. While the molecule will fold into a 3D shape, secondary structure is of interest as it tends to be preserved by evolution and function can be inferred from it. Traditional wet lab procedures to determine RNA secondary structures include Nuclear Magnetic Resonance (NMR) and X-ray Crystallography, which tend to be time consuming and expensive. Alternatively computational methods to determine RNA secondary structure can be designed. In this seminar I will motivate an approach based on representing RNA secondary structures as permutations of potential helices under certain constraints. I will then discuss an approach based on an evolutionary algorithm to search this space and return a plausible structure close to a minimum energy state, not necessarily at the global minimum energy, but biologically closer to an actual native fold. I will compare this approach with dynamic programming approaches and focus on the challenges and open problems in this area. Instead of benchmarking a wide array of algorithms I will focus more on the design advantages and challenges of different approaches. I will also briefly discuss pseudoknots and RNA visualization. The seminar aims to be motivating and explorative.