On the DCJ median problem

Mingfu Shao, Bernard M.E. Moret

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

As many whole genomes are sequenced, comparative genomics is moving from pairwise comparisons to multiway comparisons framed within a phylogenetic tree. A central problem in this process is the inference of data for internal nodes of the tree from data given at the leaves. When phrased as an optimization problem, this problem reduces to computing a median of three genomes under the operations (evolutionary changes) of interest. We focus on the universal rearrangement operation known as double-cut-and join (DCJ) and present three contributions to the DCJ median problem. First, we describe a new strategy to find so-called adequate subgraphs in the multiple breakpoint graph, using a seed genome. We show how to compute adequate subgraphs w.r.t. this seed genome using a network flow formulation. Second, we prove that the upper bound of the median distance computed from the triangle inequality is tight. Finally, we study the question of whether the median distance can reach its lower and upper bounds. We derive a necessary and sufficient condition for the median distance to reach its lower bound and a necessary condition for it to reach its upper bound and design algorithms to test for these conditions.

Original languageEnglish (US)
Title of host publicationCombinatorial Pattern Matching - 25th Annual Symposium, CPM 2014, Proceedings
PublisherSpringer Verlag
Pages273-282
Number of pages10
ISBN (Print)9783319075655
DOIs
StatePublished - Jan 1 2014
Event25th Annual Symposium on Combinatorial Pattern Matching, CPM 2014 - Moscow, Russian Federation
Duration: Jun 16 2014Jun 18 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8486 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th Annual Symposium on Combinatorial Pattern Matching, CPM 2014
CountryRussian Federation
CityMoscow
Period6/16/146/18/14

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Shao, M., & Moret, B. M. E. (2014). On the DCJ median problem. In Combinatorial Pattern Matching - 25th Annual Symposium, CPM 2014, Proceedings (pp. 273-282). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8486 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-07566-2_28