The complexity of deriving multi-labeled trees from bipartitions

J Comput Biol. 2008 Jul-Aug;15(6):639-51. doi: 10.1089/cmb.2008.0088.

Abstract

Recently, multi-labeled trees have been used to help unravel the evolutionary origins of polyploid species. A multi-labeled tree is the same as a phylogenetic tree except that more than one leaf may be labeled by a single species, so that the leaf set of a multi-labeled tree can be regarded as a multiset. In contrast to phylogenetic trees, which can be efficiently encoded in terms of certain bipartitions of their leaf sets, we show that it is NP-hard to decide whether a collection of bipartitions of a multiset can be represented by a multi-labeled tree. Even so, we also show that it is possible to generalize to multi-labeled trees a well-known condition that characterizes when a collection of bipartitions encodes a phylogenetic tree. Using this generalization, we obtain a fixed-parameter algorithm for the above decision problem in terms of a parameter associated to the given multiset.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Evolution, Molecular*
  • Models, Genetic*
  • Phylogeny*
  • Polyploidy*