1656American Journal of Botany 91(10): 1656–1665. 2004.MOLECULAR EVIDENCE ON PLANT DIVERGENCE TIMES1MICHAELJ. SANDERSON,2,5JEFFREYL. THORNE,3NIKLASWIKSTRO¨M,4ANDKA˚REBREMER42Section of Evolution and Ecology, University of California, Davis, California 95616 USA;3Bioinformatics Research Center, NorthCarolina State University, Campus Box 7566, Raleigh, North Carolina 27695 USA; and4Department of Systematic Botany,Evolutionary Biology Centre, Uppsala University, Norbyva¨gen 18D, SE-752 36 Uppsala, SwedenEstimation of divergence times from sequence data has become increasingly feasible in recent years. Conflicts between fossilevidence and molecular dates have sparked the development of new methods for inferring divergence times, further encouraging theseefforts. In this paper, available methods for estimating divergence times are reviewed, especially those geared toward handling thewidespread variation in rates of molecular evolution observed among lineages. The assumptions, strengths, and weaknesses of localclock, Bayesian, and rate smoothing methods are described. The rapidly growing literature applying these methods to key divergencetimes in plant evolutionary history is also reviewed. These include the crown group ages of green plants, land plants, seed plants,angiosperms, and major subclades of angiosperms. Finally, attempts to infer divergence times are described in the context of two verydifferent temporal settings: recent adaptive radiations and much more ancient biogeographic patterns.Key words: adaptive radiation; biogeography; divergence time; molecular clock; phylogeny; rates.Molecular sequence data have been used to estimate diver-gence times in plants for thirty years—often with controversialresults. For example, early studies aimed at the ‘‘abominablemystery’’ of angiosperm origins, undertaken assuming a strictmolecular clock, reported ages two to four times older thanthe first Early Cretaceous angiosperm fossils (Ramshaw et al.,1972; Martin et al., 1989). However, the broader molecularevolution community did not turn its attention to resolvingconflicts between molecules and fossils until the publicationof similarly surprising findings about the age of metazoans,birds, and mammals (Doolittle et al., 1996; Hedges et al.,1996; Wray et al., 1996). Although Doolittle et al. (1996) ob-tained paleontologically reasonable dates for the metazoan ra-diation, their study was overshadowed by the findings of Wrayet al. (1996), which estimated that the crown group radiationof metazoans lay deep in the Precambrian, perhaps twice asancient as implied by Cambrian fossils. In another provocativepaper, Hedges et al. (1996) estimated the age of orders ofmammals and birds to be much older than that suggested bythe fossil evidence. These studies and numerous subsequentanalyses at similar taxonomic scales (Ayala et al., 1998; Lee,1999; Wang et al., 1999; Benton and Ayala, 2003) spurredgreat interest in estimating divergence times from sequencedata. Among the consequences has been an increased appre-ciation for the role of fossil evidence (Lee, 1999; Smith andPeterson, 2002), development of new methodologies that im-pose less restrictive assumptions on the molecular substitutionprocess (Takezaki et al., 1995; Sanderson, 1997, 2002; Ram-baut and Bromham, 1998; Thorne et al., 1998; Huelsenbecket al., 2000; Thorne and Kishino, 2002), and development ofnew techniques to bring large quantities of heterogeneous se-quence data to bear on a single problem (Nei et al., 2001).Several recent reviews cover these issues from these differentperspectives (Sanderson, 1998; Arbogast et al., 2002; Smithand Peterson, 2002; Bromham and Penny, 2003; Benton andAyala, 2003).However, renewed enthusiasm has been tempered somewhatby a realization that inference methods are still far from per-1Manuscript received 30 December 2003; revision accepted 24 June 2004.5E-mail: [email protected]. Methods clearly need improvement when they produceconflicting answers to the same question. Although dramaticconflicts between ages based on sequence divergences vs. fos-sils may have raised the most concerns about the use of mo-lecular data to estimate ages, other lines of evidence have alsoraised questions. These include conflicting estimates derivedfrom different data sets, such as different taxon samples (San-derson and Doyle, 2001), sequence samples (Heckman et al.,2001), and sequence partitions (Sanderson and Doyle, 2001;Yang and Yoder, 2003); conflicts between dates estimated us-ing different calibration points (Soltis et al., 2002); and con-flicts between different inference methods themselves appliedto precisely the same data. The rapidly increasing number ofcase studies in plants promises to provide useful insights tosort through these conflicts and help improve methodology. Inthis paper, we examine the interplay between methodology anddata and review how new methods have shed light on thetiming of important events in plant evolution.METHODOLOGICAL ISSUESWhen using molecular sequence data to estimate divergence times, one keyproblem is how to disentangle times from evolutionary rates. The expecteddivergence between homologous sequences is determined by the time sincecommon ancestry and the rate at which differences have been accumulating.Although they provide information about the products of rates and times,sequence data do not by themselves allow rates and times to be disentangled.This explains why all such studies supply one or the other of these pieces ofinformation as external assumptions: either calibration times (e.g., fossils orother events in Earth history) or canonical rates (e.g., the legendary vertebratemitochondrial rate of 2% sequence divergence per million years: Klicka andZink, 1997).Methods assuming a clock—Inferring divergence times with a known to-pology and a molecular clock is relatively straightforward. Calibration pointsbased on fossil evidence can dictate the time duration of a branch, and mo-lecular sequence data can provide information about its length. A branchlength divided by its time duration yields the average evolutionary rate onthe branch. Under the assumption of a clock, rates in one part of the tree,derived from solid paleontological or biogeographical evidence, permit datingof nodes in other parts of the tree that lack paleontological evidence.Because the number of interior nodes on a rooted tree is guaranteed to beOctober 2004] 1657SANDERSON ET AL.—PLANT
View Full Document
Unlocking...