Systematics Association Symposium: Tangled Trees: Phylogeny, Cospeciation, and Coevolution 26 August 1999, University of Glasgow, Glasgow, Scotland (In association with The Systematics Association's Biennial Meeting) Organiser: Roderic D. M. Page

John Huelsenbeck¹, Bruce Rannala² and Bret Larget³

¹ Department of Biology, University of Rochester, Rochester, NY 14627; ² Department of Ecology and Evolution, State University of New York, Stony Brook, NY 11794; ³ Department of Mathematics and Computer Science, Duquesne University, Pittsburgh, PA 15282.

E-mail: 1 johnh@brahms.biology.rochester.edu; 2 rannala@allele.bio.sunysb.edu; 3 larget@mathcs.duq.edu

Information on the history of cospeciation and host-switching for a group of host and parasite species is contained in the DNA sequences sampled from each. Here, we develop a Bayesian framework for the analysis of cospeciation. We suggest a simple model of host-switching by a parasite on a host phylogeny in which host-switching events are assumed to occur at a constant rate over the entire evolutionary history of associated hosts and parasites. The posterior probability density of the parameters of the model of host switching are evaluated numerically using Markov chain Monte Carlo. In particular, the method generates the probability density of the number of host switches and of the host-switching rate. Moreover, the method provides information on the probability that an event of host-switching is associated with a particular pair of branches. A Bayesian approach has several advantages over other methods for the analysis of cospeciation. For example, it does not assume that the host or parasite phylogenies are known without error; many alternative phylogenies are sampled in proportion to their probability of being correct.

Jean-Pierre Hugot

Museum National d'Histoire Naturelle, Mammiferes & Oiseaux (FR CNRS 1541), 55, rue Buffon, 75231 Paris cedex 05

E-mail: hugot@cimrs1.mnhn.fr

The evolutionary relationships among the subgroups Phiomorpha and Caviomorpha within the Hytricognathi have generated considerable debate. Two hypothesis are conflicting: (i) Phiomorpha is the sister-group of Caviomorpha with both lineages originating from a common ancestor in Africa, or (ii) the Hystricognathi are polyphyletic with the Caviomorpha originating from a North-American ancestor. Some arguments in support of the monophyly of the Hystricognathi rely on the pinworm nematode genus Wellcomia Sambon, 1907. If this genus, which includes all the pinworms known from porcupines is monophyletic, its present-day geographic distribution requires that its ancestor must have invaded the Hystricognathi before the break-up of the southern land masses. The morphological study of these parasites reveals original derived characters which are used to propose a phylogeny of the pinworms. The relationships between the parasites assessed in the resulting cladogram imply : (i) That the different genera parasitic in Hystricognathi form a monophyletic group and this is consistent with Hystricognathi monophyly; (ii) That all the pinworms parasitic on porcupines are closely related; (iii) The distribution of the parasites of Caviomorpha and Phiomorpha also suggests that these groups could be paraphyletic.

Joanne Martin¹, Elisabeth Herniou², and Michael Tristem³

^1&3 Department of Biology, Imperial College, Silwood Park, Buckhurst Road, Ascot, Berkshire, SL5 7PY, United Kingdom; ² Department of Zoology, The Natural History Museum, London, SW7 5BD, United Kingdom.

E-mail: 1 j.martin01@ic.ac.uk; 2 e.herniou@nhm.ac.uk; 3 m.tristem@ic.ac.uk.

Members of four classes of vertebrates have recently been shown to harbour Murine Leukaemia Virus (MLV)-related retroviruses. This has provided the first opportunity to assess the roles of horizontal and vertical transmission in the evolution of this viral genus. Phylogenetic analyses including sequences amplified from across the MLV host range show that viruses isolated from a particular host class generally cluster together, and the relationships of individual isolates within each viral cluster are largely in keeping with the phylogenies of their hosts, at least to the order/family levels. Although dependent on the position of the root of the MLV phylogeny, the positioning of the clusters of viruses from each vertebrate class relative to each other largely reflects the vertebrate tree of life. Alternative root positions suggest some infectious transfer may have occurred early in MLV evolution. The degree of congruence between host and virus phylogenies suggest that MLVs integrated into early vertebrates and have been inherited vertically, tracking the phylogenies of their hosts, over very long periods of time. Phylogenies based on sequences from the polymerase gene of 44 novel MLV-related retroviruses will be presented, together with a brief summary of the methodology used to generate these data.

Kevin Johnson and Dale H. Clayton

Department of Biology, University of Utah, Salt Lake City, UT 84112-0840, USA

E-mail: johnson@biology.utah.edu, clayton@biology.utah.edu

We reconstructed phylogenies for pigeons and doves (Columbidae) and their associated feather louse (Phthiraptera: Ischnocera) parasites. Phylogenies of lice included two separate groups of feather lice: wing lice and body lice. Both parasite phylogenies were more similar to the host phylogeny than expected by chance. We also evaluated whether cospeciation events in the two parasite groups were correlated.

Adrian Paterson¹, Ricardo Palma², and Russell Gray³

¹ Ecology and Entomology Group, Lincoln University, PO Box 84, Lincoln, New Zealand; ² Te Papa Museum of New Zealand, PO 467, Wellington, New Zealand; ³. Department of Psychology, University of Auckland, Private Bag 92019, Auckland, New Zealand

E-mail:- 1 PatersoA@lincoln.ac.nz; 2 RicardoP@tepapa.govt.nz; 3 rd.gray@Auckland.ac.nz;

There are two reasons why parasite species to explain absence from potential host species: 1) they have never been present on the host species, or 2) they were present on a host species, or its ancestor, but were subsequently lost through sorting events. Sorting events occur when parasite species go extinct from a host species or when a founding host population, by chance, does not carry a parasite species found in the parent population ("missing the boat"). Studies of coevolution using component and Brook's parsimony analyses typically find evidence for many sorting events relative to other coevolutionary events (cospeciation, intrahost speciation and host-switching). We tested whether sorting events were likely to occur by examining the New Zealand bird fauna and their ectoparasitic lice and identifying host parent-daughter species pairs. Forty eight such pairs were identified of which 34 had lost louse species. Daughter bird species were found to have significantly fewer louse species (one-tailed binomial test, P = 0.003) with a founding population typically losing one louse species per host founding or speciation event. This result confirms, at least for seabirds and lice, that sorting events are real and are likely to be numerous within a coevolutionary analysis.

Bruce Rannala¹ and Yannis Michalakis²

¹ Department of Ecology and Evolution, State University of New York, Stony Brook, NY 11794-5245 USA
² Fonctionnement et Evolution des Systemes Ecologiques CNRS UMR 7625, Universite P. & M. Curie, Bat. A 7eme et. CC 2377 quai St. Bernard 75252 Paris CEDEX 05 France

Email: 1 rannala@life.bio.sunysb.edu 2 imichala@hall.snv.jussieu.fr

A theoretical analysis is presented considering the effects of population structure and parasite transmission processes on patterns of host-parasite cospeciation. Coalescent theory is used to predict the probability that host and parasite gene trees are identical as a function of the effective population sizes of hosts and parasites, parasite transmission rates, and other demographic influences. The distribution of the mean-square difference between the coalescence times in the host and parasite trees is also considered for different values of these parameters. Conditions are described under which population-level processes alone can lead to large differences between host and parasite gene trees despite a history of synchronous allopatric isolation events in the associated host and parasite populations and strict host fidelity among parasite species (i.e., no host-switching).

Fredrik Ronquist

Dept. Systematic Zoology, Uppsala University, Norbyvägen 18D, SE-752 36 Uppsala, SWEDEN

email: fredrik.ronquist@zoologi.uu.se

In recent years, there has been an important shift from a pattern-based to a process-oriented approach to parsimony analysis of coevolving species associations. Process-oriented parsimony methods explicitly model the phylogenesis of species associations in terms of states (associated species pairs) and transitions between these states (events such as cospeciations and host switches). Each event is assigned a cost and the most parsimonious reconstruction of the history of an association is the one with the lowest total cost. I give a brief overview of the event-based parsimony methods proposed to date, including methods based on two-event, three-event and four-event models. I discuss similarities and differences among event-based protocols and some problems that have not yet been solved. I also show that some traditional pattern-based methods, such as Brooks Parsimony Analysis, are close approximations of event-based protocols and that the event-based derivation reveals their analytical properties.

Jason Taylor and Andy Purvis

Department of Biology, Imperial College at Silwood Park, Ascot, Berks, SL5 7PY UK

E-mail: jason.taylor@ic.ac.uk a.purvis@ic.ac.uk

We have constructed a phylogeny of 233 species and subspecies of chewing lice and compared this to various mammal phylogenies in order to detect events of putative cospeciation, host switch, lineage sorting, and duplication. These comparisons at the family and order level suggest that the chewing lice originated on an ancestral artiodactyl, probably an ancestor of the bovids. This places the origin of the chewing lice between 20 and 40 Mya, which was then followed by many host switches between mammal families. Preliminary comparisons at the species level using TreeMap, suggest the amount of cophylogeny varies between clades. The canids, for instance, have not experienced a statistically significant number of cospeciation events while the bovids have. These results suggest that cophylogeny between mammals and their chewing lice is patchy rather than common, pocket gophers aside, it occurs in only a few clades. This raises two important questions. Firstly, how can the phylogeny upon which these results are based be improved and secondly, what are the factors that break down cospeciation between mammals and their lice?