Kathleen Steinhoefel King's College London Peek-a-boo with Protein Structure Candidates Research on protein structure prediction and protein folding has a long history, dating back to the seminal work by Pauling and Corey in 1953. Research on these topics has been back in the spotlight with the Nobel prize 2013 awarded to Karplus, Levitt and Warshel for their work on computer-aided analysis of protein molecules and for introducing simplified all-atom energy functions for protein folding simulation. From a computational point of view, all-atom protein structure prediction is a challenging task. Interestingly, Finkelstein and Badretdinov approximated the worst case folding time of a protein of length n as exp(n^{2/3})ns. In terms of algorithmic complexity, this implies a quasi-exponential structure prediction time. Consequently, protein structure prediction has been shown to be NP-hard for various lattice models and contact-based energy functions. A common approach to tackle NP-hard problems are local search-based heuristics, such as simulated annealing. However, the performance of these methods depends strongly on properties of the underlying energy landscape induced by the objective function and the associated neighbourhood relation. In 2008, we devised a simulated-annealing procedure with a time-dependent cooling schedule and a neighbourhood relation determined by the pull-move set. We obtained experimental evidence that the maximum depth D of local minima of the underlying energy landscape can be upper bounded by D