科研之友
微信
新浪微博
Facebook
分享链接
摘要
A main open question related to character-based tree reconstruction is designing generalizations of the Perfect Phylogeny approach that couple efficient algorithmic solutions to the capability of explaining the input binary data, by allowing back mutations of some characters. Following this goal, the Persistent Phylogeny model and the related tree reconstruction problem (the PPP problem) have been recently introduced: this model allows only one back mutation for each character. The investigation of the combinatorial properties and the complexity of the model is still open: the most important such question is whether the PPP problem is NP-hard. Here we propose a graph-based approach to the PPP problem by showing that instances can be represented by colored graphs, while the solutions are obtained by operations on such graphs. Indeed, we give a graph based characterization of the solutions to the PPP problem by showing the relationship between certain sequences of graph operations on the instance graphs and traversals of a persistent phylogeny solving these instances. Based on this result and on some combinatorial properties of the instance graphs we are able to give a polynomial time algorithm for a restricted version of the PPP problem.
- 出版日期2017-1-7
