Dynamic programming algorithms for RNA secondary structure prediction with pseudoknots

Prediction of consensus RNA secondary structures including pseudoknots

“…Therefore the maximum matching needs to be post-processed. During post-processing several edges are deleted from the original input graph resulting in a modified weighted graph Γ (1) . The computation of the maximum matching and post-processing are iterated to convergence.…”

“…Thermodynamic structure prediction based on the standard energy model is NP-complete [34,1] in general, albeit restricted classes of pseudoknots can be dealt with by polynomial algorithms. Nevertheless, these approaches are expensive in terms of CPU and memory usage [39,38,19,1,10] and in addition suffer from uncertainties of the energy model for pseudoknots [16]. Figure 1.…”

compPknots: A Framework for Parallel Prediction and Comparison of RNA Secondary Structures with Pseudoknots

“…Akutsu [13] analyzed Uemura's method and found that the tree-adjoining grammar was not crucial but the parsing procedure was. Since the parsing procedure is intrinsically a dynamic programming procedure, Akutsu re-formulated this method as a dynamic programming procedure without the tree-adjoining grammar.…”

Problems on RNA Secondary Structure Prediction and Design

“…The thermodynamic model for pseudoknot free secondary structures has been extended to include contributions of pseudoknotted stems and loops. Several algorithms have been proposed for predicting the mfe secondary structure from a class of secondary structures that allows limited types of pseudoknots [1,15,19,28]. Other algorithms are heuristic in nature, such as the genetic algorithm of Gultyaev et al [12].…”

“…Another limitation of the algorithm is its high running time of Θ(n 6 ). An algorithm of Akutsu [1] runs in O(n 4 ) time and O(n 2 ) space, but there are natural pseudoknotted structures that cannot be handled by this algorithm.…”