Who is Who in Phylogenetic Networks

Publication no #2182   Type :   Html | Bib | Both

Created: 2012-03-20 13:44:38     Modified: 2017-07-25 16:12:53

2182 Hyun Jung Park and Luay Nakhleh. MURPAR: A fast heuristic for inferring parsimonious phylogenetic networks from multiple gene trees. In ISBRA12, Vol. 7292:213-224 of LNCS, springer, 2012. Keywords: explicit network, from unrooted trees, heuristic, phylogenetic network, phylogeny, reconstruction, software. Note: https://www.researchgate.net/profile/Hyun_Jung_Park2/publication/262318595_MURPAR_A_Fast_Heuristic_for_Inferring_Parsimonious_Phylogenetic_Networks_from_Multiple_Gene_Trees/links/54b7e7b50cf269d8cbf58cc4.pdf.         Toggle abstract "Phylogenetic networks provide a graphical representation of evolutionary histories that involve non-treelike evolutionary events, such as horizontal gene transfer (HGT). One approach for inferring phylogenetic networks is based on reconciling gene trees, assuming all incongruence among the gene trees is due to HGT. Several mathematical results and algorithms, both exact and heuristic, have been introduced to construct and analyze phylogenetic networks. Here, we address the computational problem of inferring phylogenetic networks with minimum reticulations from a collection of gene trees. As this problem is known to be NP-hard even for a pair of gene trees, the problem at hand is very hard. In this paper, we present an efficient heuristic, MURPAR, for inferring a phylogenetic network from a collection of gene trees by using pairwise reconciliations of trees in the collection. Given the development of efficient and accurate methods for pairwise gene tree reconciliations, MURPAR inherits this efficiency and accuracy. Further, the method includes a formulation for combining pairwise reconciliations that is naturally amenable to an efficient integer linear programming (ILP) solution. We show that MURPAR produces more accurate results than other methods and is at least as fast, when run on synthetic and biological data. We believe that our method is especially important for rapidly obtaining estimates of genome-scale evolutionary histories that can be further refined by more detailed and compute-intensive methods. © 2012 Springer-Verlag."