Skip to content

Review: Cophylogeny on a Fine Scale: Geomydoecus Chewing Lice and Their Pocket Gopher Hosts, Pappogeomys bulleri

Citation

  • Demastes, J. W., Spradling, T. A., Hafner, M. S., Spies, G. R., Hafner, D. J. & Light, J. E. (2012). Cophylogeny on a fine scale: Geomydoecus chewing lice and their pocket gopher hosts, Pappogeomys bulleri. Journal of Parasitology, 98(2), 262–270.
  • DOI

Abstract

Pocket gophers and their chewing lice show broadly congruent phylogenies over deep timescales. This study tests cophylogeny at a much finer scale, comparing populations within a single host species, Pappogeomys bulleri, and three louse species of the Geomydoecus bulleri complex, using mitochondrial and nuclear DNA. Contrary to the expectation that recently diverged taxa show weaker congruence, a statistically significant pattern of parallel cladogenesis is recovered, along with a higher rate of molecular evolution in the louse lineages.

The earlier gopher-louse studies compared distantly related taxa. Demastes and colleagues push the comparison down to the population level within one host species, where cospeciation is expected to be hardest to detect: recent divergence leaves time for gene flow, and incomplete lineage sorting and host switching can make a gene tree a poor proxy for the history of the lineage. They sequence mitochondrial COI and nuclear EF-1α, fit substitution models with among-site rate heterogeneity (GTR+I+G), and reconcile host and parasite trees with distance-based and event-based tests. The pattern of parallel cladogenesis survives even here, and the louse lineages again evolve several times faster than their hosts.

The methodological lesson is about phylogenetic uncertainty. At fine scales the tree that best fits one gene need not match the species history or the tree from another gene, because lineage sorting and gene flow act on each gene independently. A single point estimate can therefore hide as much as it shows, and the honest object of study is the set of trees a gene supports, not one winner.

Relevance to Hifuku

This is the case Hifuku's design targets. When gene trees disagree with each other and with the species history, a survey that maps the likelihood landscape over many trees is more informative than a single maximum-likelihood tree: the map shows where the support is concentrated and where it is spread across competing topologies. Charting several genes on one shared frame, as Hifuku's survey command does, lets the reader see directly where genes agree and where lineage sorting or rate variation pulls their maps apart, with the metric separating topological disagreement (Robinson-Foulds) from branch-length disagreement (clade branch score), Section 5. See Hafner & Nadler 1988 and Hafner et al. 1994 for the deeper-scale studies of the same assemblage.