Integrative Biology 200B University of California, Berkeley "Ecology and Evolution" Spring 2009 March 12th, 2009 NM HallinanTree Shape: Phylogenies & MacroevolutionToday we move on to a discussion of macroevolutionary forces. We will expand our concerns beyond the consequences of character change in populations to include evolutionary forces that rely on differential creation and survival of species. In particular we will discuss methods which allow us to compare the distribution of diversity among the different clades in the tree of life. For a long time this subject was the sole purview of paleontologists, but in recent years with the advent of large accurate phylogenies there has been an explosion of research in this field among neontologists as well.DiversificationDiversity is the number of species in a given taxa. Here, we are not interested in diversity, per se, but in the processes that produced that diversity. I will call all those processes diversification. There is a tendency to think of diversification as only those processes that increase the number of species, but here it will refer to both speciation and extinction. In practice it is very difficult to separate the affects of extinction from the affects of speciation using only extant taxa. Today I will lay out a series of methods are for detecting differences in diversification, not diversity. Several things can lead to a difference in diversity without requiring that there be a difference in diversification. For example, we expect older clades to be more diverse. It is also possible for stochastic processes to lead to differences in diversity that we would consider within the range of equivalent outcomes. For example within the monkeys, there are about 50 species of new world monkeys, 80 species of old world monkeys and 20 species of apes. These are large differences in diversity, but do not require different diversification processes to explain them.What could cause diversification to vary? To answer that question we must first explore the types of patterns produced by diversification.Tree Shape Tree shape is a catchall term used to describe the properties of a phylogeny other than the taxa at the tips and the character states either at the tips or reconstructed along the branches and nodes of a tree. The shape of a phylogeny consists of two properties, the topology and the branch lengths.Topology is the particular branching pattern for a tree. A labeled topology represents the relationships among the taxa at the tips. An unlabeled topology has no taxa at the tips and thus consists only of the abstract tree shape. In this way two phylogenies have the same unlabeled topology if the taxa can be rearranged on the tips in such a way that the taxa have the same relationships, but only have the same labeled topology if all the taxa have the same relationships without rearrangement. Today we will mostly concern ourselves with unlabeled topologies; we will only considered labeled topologies at the end, when we include character data.Imbalance refers to the distribution of taxa among the different clades of a phylogeny. If taxa are evenly distributed among the clades, then the topology is balanced. On the other hand, if some clades have many more taxa than other clades of equivalent rank, then a tree is considered imbalanced.Excessively balanced topologies could result from several possible causes. Competition among close relatives could create a situation in which there is more competition and thus less diversification in large clades. If there is a period of time after speciation during which a lineage could not speciate again you would get clades that are more balanced than you would expect under equal branching. There are undoubtedly other processes that could increase balance. Excessively imbalanced topologies are usually assumed to be caused by heritable characters that effect diversification, such that diversification will show a phylogenetic signal. There is an infinite list of characters that could effect diversification.Key innovations could be defined as organismal characters that expand the available niche space and thus lead to an increase in diversification.Other types of organismal characters that do not lead to an increase of niche space could also lead to a change in diversification. For example sexual selection has been proposed to lead to increases in speciation. On the other hand broadcast spawners might be expected to show low diversification.Phylogeography can also lead to large differences in diversification, with a great deal of phylogenetic signal. Island clades are often more diverse than their closest main land relatives, as they faced little competition upon colonizing the island. One area may also have more available energy or a more heterogeneous environment than another and thus taxa in that area would have higher rates of diversification.LowHighImbalanceEarlyLateDiversification TimingWhen analyzing or describing macroevolution, the branch lengths of a tree are usually proportional to time; thus these trees are ultrametric. Branch lengths are also of critical importance, when likelihood models are used to analyze a topology. The branch lengths are also interpreted independently of the topology, in order to identify temporal shifts in diversification. Branching times, the timings of speciation events, can be compared on a lineages through time plot. Temporal variation in diversification that effects an entire clade could be attributed to several factors. Adaptive radiations would be expected to lead to high rates of diversification (probably high rates of speciation) early, and declining diversification through time. On the other hand a mass extinction would clearly show high across the board extinction for a brief period of time. It is difficult to separate the signal of mass extinction from generally high extinction, when examining only extant taxa. Other types of geological events could also affect an entire clade at once, as could a major ecological shift.Null Models Before we start asking questions about imbalance and branching times, we must carefully consider what we expect to see if there are no macroevolutionary forces acting. Several different null distributions have been proposed, we will focus only on the simplest and most commonly used.Equiprobable trees (Maddison and Slatkin 1991) considers each labeled topology to have the same
View Full Document
Unlocking...