A faster algorithm for RNA co-folding

Michal Ziv-Ukelson, Irit Gat-Viks, Ydo Wexler, Ron Shamir

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The current pairwise RNA (secondary) structural alignment algorithms are based on Sankoff's dynamic programming algorithm from 1985. Sankoff's algorithm requires O(N 6) time and O(N 4) space, where N denotes the length of the compared sequences, and thus its applicability is very limited. The current literature offers many heuristics for speeding up Sankoff's alignment process, some making restrictive assumptions on the length or the shape of the RNA substructures. We show how to speed up Sankoff's algorithm in practice via non-heuristic methods, without compromising optimality. Our analysis shows that the expected time complexity of the new algorithm is O(N 4 ζ(N)), where ζ(N) converges to O(N), assuming a standard polymer folding model which was supported by experimental analysis. Hence our algorithm speeds up Sankoff's algorithm by a linear factor on average. In simulations, our algorithm speeds up computation by a factor of 3-12 for sequences of length 25-250. Availability: Code and data sets are available, upon request.

Original languageEnglish
Title of host publicationAlgorithms in Bioinformatics - 8th International Workshop, WABI 2008, Proceedings
Pages174-185
Number of pages12
DOIs
StatePublished - 2008
Event8th International Workshop on Algorithms in Bioinformatics, WABI 2008 - Karlsruhe, Germany
Duration: 15 Sep 200819 Sep 2008

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5251 LNBI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference8th International Workshop on Algorithms in Bioinformatics, WABI 2008
Country/TerritoryGermany
CityKarlsruhe
Period15/09/0819/09/08

Fingerprint

Dive into the research topics of 'A faster algorithm for RNA co-folding'. Together they form a unique fingerprint.

Cite this