A scalability study of phylogenetic network inference methods using empirical datasets and simulations involving a single reticulation

Table 1 Topological distances between inferred phylogenies in the empirical study

Average (SE) topological distance between inferred phylogenetic
networks
	MLE-length		MP		SNaQ
MLE-length	.11	(.02)	.42	(.06)	.44	(.04)
MP			.36	(.03)	.52	(.05)
SNaQ					.23	(.02)

Phylogenies were inferred using a representative method from each category of multi-locus methods: MLE-length (a full likelihood probabilistic method), MP (a parsimony-based method), and SNaQ (a pseudo-likelihood-based probabilistic method). The normalized tripartition distance between solutions that included gene flow (i.e., phylogenetic networks with one reticulation) is shown as an average (standard error) across replicates (n=20). When constrained to infer a phylogenetic tree rather than a phylogenetic network, all methods inferred an identical species tree across all replicates. Each replicate dataset consists of randomly selecting a sample from the following mouse species and subspecies: Mus musculus domesticus, M. musculus musculus, M. musculus castaneus, M. spretus, M. spicilegus, and M. macedonicus

ISSN: 1471-2105