Molecular Biology and Evolution 19:101-109 (2002)
© 2002 Society for Molecular Biology and Evolution
Estimating Absolute Rates of Molecular Evolution and Divergence Times: A Penalized Likelihood Approach
Section of Evolution and Ecology, University of California, Davis
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Rates of molecular evolution vary widely between lineages, but quantification of how rates change has proven difficult. Recently proposed estimation procedures have mainly adopted highly parametric approaches that model rate evolution explicitly. In this study, a semiparametric smoothing method is developed using penalized likelihood. A saturated model in which every lineage has a separate rate is combined with a roughness penalty that discourages rates from varying too much across a phylogeny. A data-driven cross-validation criterion is then used to determine an optimal level of smoothing. This criterion is based on an estimate of the average prediction error associated with pruning lineages from the tree. The methods are applied to three data sets of six genes across a sample of land plants. Optimally smoothed estimates of absolute rates entailed 2- to 10-fold variation across lineages.
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Estimates of rates of evolution of genes and other elements of the genome have revealed much about molecular evolution across a diversity of taxa (e.g., Li 1997
). Comparisons of rates have contributed to the development of ideas about modes of selection on different genomic elements, such as introns versus exons, and have permitted correlates of rate differences, such as generation time and metabolic rate, to be identified (Gillespie 1991
; Martin and Palumbi 1993
). Comparisons of relative rates between lineages, in particular, have provided abundant evidence for departures from constant rates of substitution (Li and Wu 1985
; Britten 1986
; Li 1997
; Muse 2000
), a finding that has not, however, dampened enthusiasm for estimating divergence times from molecular data (e.g., Wray, Levinton, and Shapiro 1996
; Kumar and Hedges 1998
; Korber et al. 2000
).
Characterization of the timescale over which molecular rates change and of the extent of their autocorrelation in time has lagged behind characterization of relative rates (Gillespie 1991
). The reasons for this are several. Such studies require good estimates of absolute substitution rates. Estimates of absolute rates are often made by pairwise comparisons that necessarily average the rate differences on intervening branches (Easteal and Herbert 1997
; Ayala, Rzhetsky, and Ayala 1998
), thus underestimating the variability in rate. More comprehensive methods, such as estimating the slope of the regression of pairwise distance against calibrated divergence time (Wray, Levinton, and Shapiro 1996
; Leitner and Albert 1999
), have difficulty accounting for phylogenetic nonindependence (Pagel 1997
; Ayala, Rzhetsky, and Ayala 1998
). Estimation of absolute rates also requires reliable external information about time, usually from fossils. The use of fossil information in molecular rate estimation problems hinges, in turn, on the correct assignment of fossils to particular nodes in a phylogenetic tree (Marshall 1990
; Smith and Littlewood 1994
; Springer 1995
; Lee 1999
).
Relative rate comparisons, though robust, provide neither an estimate of absolute rate, nor an indication of how absolute rates change through time. Reconstructing how absolute rates change through time requires estimates of rate differences between two or more sequential branches in a tree. Relative rate comparisons only indicate differences in rate between sister branches, which are not sequential, are not descended from each other, and therefore do not indicate the direction of change. This problem cannot be corrected merely by performing multiple relative rate comparisons, whether they are nested within one another or not. Truly independent comparisons share no timepoint in common, making absolute comparisons impossible, and nested comparisons suffer from great sensitivity to what is assumed about the timing of nested nodes (Sanderson and Donoghue 1996
).
In recognition of these obstacles, several general approaches for estimating absolute rate variability have been proposed, mainly in conjunction with estimating divergence times in the presence of rate variation. Because divergence times and absolute rates of evolution are inextricably linked, one cannot be estimated without the other. Some methods involve pruning outlier taxa that appear to depart from a tree-wide rate (Takezaki, Rhetsky, and Nei 1995
). Some are local molecular clock methods, in which subtrees of the phylogeny are assigned different rates, but the rate is constant within each subtree (Hasegawa, Kishino, and Yano 1989
; Uyenoyama 1995
; Cooper and Penny 1997
; Rambaut and Bromham 1998
; Bromham and Hendy 2000
; Yoder and Yang 2000
). One potential problem with these approaches is that subset selection may be arbitrary, and in large trees, the number of possible ways to assign different rates to subtrees is large (Sanderson 1998
). More general, but still parametric, methods have been described (Thorne, Kishino, and Painter 1998
; Huelsenbeck, Larget, and Swofford 2000
; Kishino, Thorne, and Bruno 2001
), which assume a specific model for rate variation from branch to branch. Thorne, Kishino, and Painter (1998)
assumed a lognormal distribution of rate changes whereas Huelsenbeck et al. assumed a compound Poisson process in which rates change in a stepwise fashion within lineages, and the amount of change is governed by a gamma distribution.
Given that little is known about the timescale for rate variation (Gillespie 1991
), estimation procedures that are less reliant on parametric assumptions might prove useful. An entirely nonparametric method, nonparametric rate smoothing (NPRS: Sanderson 1997
), estimated rates and times via a least-squares smoothing criterion that penalized rapid rate changes on a tree (Sanderson 1997
). However, parametric models, coupled with maximum likelihood estimation methods, have proven to be statistically powerful and highly explanatory descriptions of molecular evolutionary processes. In this paper, I develop a semiparametric approach for estimating rates of molecular evolution. This approach attempts to combine the power of parametric methods and the robustness of nonparametric methods by the use of penalized likelihood (Green and Silverman 1994
), part of a general class of semiparametric techniques used in smoothing and regression problems.
Penalized likelihood takes a parameter-rich model that would ordinarily overfit the data and constrains fluctuations in its parameters by a roughness penalty. For rate variation, one roughness penalty would penalize how quickly rates varied from branch to nearby branch, as in the NPRS method (Sanderson 1997
). By adding the parametric component back in, it is possible to examine a broad spectrum of solutions with different levels of rate smoothing, ranging from highly penalized, nearly rate-constant models, to nearly unconstrained rate variability. The key to this approach is to find an objective method for selecting an optimal level of smoothing. After developing such an approach, based on cross-validation, it will be illustrated with three molecular data sets on land plants.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Models and Maximum Likelihood Estimation
Consider a rooted phylogenetic tree with M taxa and S + 1 internal nodes, in which the root node is labeled by 0, the remaining internal nodes are labeled by integers from {1, ..., S} and the terminal nodes are labeled with integers {S + 1, ..., S + M}. Branches are labeled by the node they subtend. Let node k have an age tk (measured backward from the present), and its ancestral node, anc(k), have an age tanc(k). The branch defined by these two nodes has a duration in time given by tanc(k) - tk.
Nucleotide substitution models are commonly cast in a general framework of a four-state Markov process (Rodriguez et al. 1990
), in which transition probabilities between states are explicitly modeled. This paper extends model realism in a different direction to account for rate variation between lineages and, for simplicity, the standard Markov formulation will be replaced with a simpler substitution process, in which the estimated number of substitutions along a branch is regarded as an observation, xk, drawn from a Poisson process which has a rate rk. The more complex formulation can be reintroduced with considerable computational costs.
Two very different models lie at the opposite extremes of a spectrum of rate variation among lineages. At one extreme is a clock (CL) model, in which the rate parameters are the same for every branch, rk = r. At the other extreme is a saturated (SAT) model, in which each branch is permitted to have a unique rate, rk. The unknown parameters can be written as 
CL = {t0, ..., tS; r} for the CL model and SAT = {t0, ..., tS; r1, ..., rS + M} for the SAT model, corresponding to S + 2 or 2S + M + 1 free parameters, respectively.
Let P(x|
) = x exp(-)/x! be the usual probability of an observation x taken from a Poisson distribution with parameter . Then the log likelihood of for the SAT model is given by
|
|
, can be obtained from these expressions by numerical methods (Langley and Fitch 1974
; Sanderson 1997
; Cutler 2000
). However, the situation for the SAT model is more problematic because there are more free parameters (2S + M + 1) than observations (S + M). The model is not identifiable, meaning that several parameter values can produce the same likelihood, and therefore it is not possible to estimate a unique SAT, without imposing some constraints on rate variation.
Penalized Likelihood
Most previous attempts to impose constraints on rate variation have relied on explicit parametric modeling of the rate variation. A less parametric alternative is to impose a roughness penalty (Green and Silverman 1994
; Simonoff 1994
), which forces rates to change smoothly from branch to branch. Instead of finding the parameter set that maximizes the log likelihood, for example, we can maximize the penalized likelihood, given by
|
|
is a roughness penalty, which increases as rates vary more rapidly across the tree, and 
is a smoothing parameter that controls the tradeoff between smoothness and goodness-of-fit of the data to the SAT model. At one extreme, = 0, is the SAT model described above. At the other extreme, as 
, the parameter estimates are expected to converge to those of the CL model because no variation in rate is tolerated.
The roughness penalty, , should be designed to reflect changes in rate between neighboring branches of the tree. After Sanderson (1997)
, this is chosen to penalize squared differences in rates between ancestral and descendant branches and the variance in rate between the branches descended from the root node:
|
|
The semiparametric formulation described by equations (2) and (3) is ad hoc, and can really only be justified in terms of its performance in the present problem. For this, it is necessary to develop an objective measure of performance.
Cross-validation and the Choice of Smoothing Parameter
The value of the smoothing parameter, , can be seen as indexing an infinite number of semiparametric models. The choice of will affect the estimated rates and times, and it is therefore desirable to have an objective, data-driven method for choosing this parameter. A widely used method for model selection in general (Burnham and Anderson 1998
) and smoothing in particular, is cross-validation (Green and Silverman 1994
), which sequentially removes small subsets of the data, estimates parameters from the remaining data for a given choice of the smoothing parameter, and then uses the fitted model parameters to predict the data that were removed. Ideally, some choice of smoothing parameter will lead to a best prediction of the removed data, signally the optimal level of smoothing. In practice this is done by constructing a cross-validation criterion (CV) related to prediction error and then selecting the smoothing parameter that minimizes CV.
One method for cross-validation on trees is the pruning of terminal branches from the tree. Removal of a terminal branch, m, from the tree leaves its immediate ancestral node in place, along with all the other branches in the tree. Label the set of observations on numbers of substitutions on the remaining branches {x}(-m). The idea is to use this reduced set of observations to estimate parameters of the model using the penalized likelihood method described above. This generates a set of M estimates, SAT(-m), corresponding to each pruned taxon.
After estimation, the observed value, xm, can be compared to the predicted value, xm*. One way to make this prediction is to base it on the rate of branch m's immediate ancestor branch, which is 
anc(m)(-m), plus information based on the estimated age of the ancestor of node m, 
anc(m)(-m). Both of these estimates are part of SAT(-m). In a Poisson process, this expected number of substitutions is just the product of the relevant rates and the duration of branch m, so
 |
The quality of the prediction can be measured by the squared deviation of the prediction from the observation, weighted by the inverse of the variance (which is equal to the mean in a Poisson process). A CV criterion can be constructed by taking the average of these prediction errors over all ways to prune the terminal branches:
|
|
, we may be able to select by minimizing CV.
In the data analyzed below, was varied on a log scale between 0.1 and 10,000. For comparative purposes, the CV score for two other estimators was also obtained, although these do not depend on . These were the NPRS method described in Sanderson (1997)
and the maximum likelihood estimation using the CL model, as outlined by Langley and Fitch (1974)
. A priori, these might be considered logical extremes. Because the real data sets depart significantly from clocklike substitution rates, a simulated data set was constructed to test the performance of the cross-validation method under those conditions. A tree of 15 taxa was constructed according to a stochastic pure-birth process (Cox and Miller 1977
), which generates a topology and branch durations. Numbers of substitutions were then assigned to branches by generating random Poisson deviates with means equal to the rate of substitution times the duration of the particular branch. The rate of substitution was set to 50 substitutions per unit time, in which a unit of time consisted of the distance from root to tip of the tree. The simulation protocol has been described elsewhere in more detail (Bininda-Emonds et al. 2001
) and is implemented in the author's program, r8s (http://ginger.ucdavis.edu/r8s).
Optimization of the Penalized Likelihood
Optimization entails the search for solution(s), ,that maximize the objective function, 
, in equation (2)
for a given set of observations, {xk}, and choice of . The maximization of this objective function is a nonlinear optimization problem which can be solved numerically by standard numerical techniques such as Powell's gradient-free method and quasi-Newton gradient-based methods (Gill, Murray, and Wright 1981
; Press et al. 1992
). Termination criteria were based on both the convergence of the objective function and the gradient (to zero). Any cases in which convergence was a problem were reanalyzed with different optimization settings or a different algorithm. The local stability of solutions was checked by perturbing them and restarting the search, and all searches were started from several different initial random guesses at the parameters.
Penalized likelihood optimization is implemented in the author's program, r8s.
Molecular Data
The methods are illustrated by reference to three published molecular sequence data sets spanning land plants. The timescale for these taxa extends back 450 MYA to the origin of land plants or to the origin of vascular plants at about 420 MYA. None of the data sets are clocklike, based on likelihood ratio tests (data not shown). The first data set consists of 3,795 nt concatenated from two chloroplast photosystem (PS) genes, psaA and psbB,in 19 species of land plants (Sanderson et al. 2000
: tree from their Fig. 3
; sequence data available at http://ginger.ucdavis.edu/wwwdata). The second data set consists of 1,428 nt of the plastid rbcL gene sampled in 37 land plants (Sanderson and Doyle 2001; data at http://ginger.ucdavis.edu/wwwdata). Data for the three protein coding genes were partitioned into approximate substitution classes to help distinguish nonsynonymous from synonymous changes: one class for first and second positions and the other for third positions. The third data set consists of 4,744 nt of concatenated sequence from small subunit (SSU) ribosomal DNA sequences from nuclear, chloroplast, and mitochondrial genomes of 28 land plants (Nickrent et al. 2000
: sequence data at http://www.science.siu.edu/landplants/Alignments/Alignments.html; the tree used is the single most parsimonious tree obtained from that data set, excluding rbcL,which was included in the original data set on the website, and excluding two algal outgroups).
|
Zero-length branches were collapsed to hard polytomies. Estimated numbers of substitutions along each branch of these trees (the observations, xk) were obtained by maximum likelihood using PAUP* 4.0 (Swofford 1999
) with a Jukes-Cantor model of substitution. This is an extremely simple substitution model but should provide the closest fit to the Poisson assumption described earlier. Subsequently, I discuss extensions to more complex substitution models. |
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Algorithmic Issues
Successful estimation of parameters—meaning that the optimization procedure converged to a stable solution—depended on the data, the smoothing parameter, and the numerical algorithm used. Data sets with as many as 100 taxa were tested, and for reasonable levels of smoothing, could converge in under 60 s using quasi-Newton methods on a 500 MHz Pentium III running Linux 6.0. All algorithms had trouble with extremely low smoothing parameters, when the optimization problem is ill-conditioned, meaning that a small change in the data translates into a large change in the estimated parameters (Gill, Murray, and Wright 1981
). Overall, quasi-Newton methods that relied on an explicit calculation of the gradient of the objective function succeeded much better over a wider range of smoothing values. Derivative-free methods ran into trouble both at low smoothing and at very high levels of smoothing. On the other hand, standard implementations of quasi-Newton methods in Press et al. (1992)
failed when any terminal branches had zero observed substitutions, apparently because the directional derivative for the rate along those terminal branches cannot be zero even at the obtained solution.
Cross-validation Results
In all the molecular data sets, plots of the CV score versus the level of smoothing, , indicate a minimum for intermediate levels of smoothing (fig. 1
). These optimal points indicate the level of smoothing corresponding to the least prediction error, and therefore these can be viewed as the best semiparametric models for the data. A range of patterns is observed, however, among the various data sets. In the PS and SSU data sets, very high values of introduce more prediction error than very low values. One does better with an undersmoothed than with an oversmoothed model. For the rbcL third position data, the reverse is true, and for first and second position data the CV curve is nearly symmetrical. In the data from the simulation that assumed a clock, no minimum is present. Instead, there is a monotonic improvement as smoothing increases in the direction of clocklike evolution.
|
Comparisons with the CV scores of other methods indicate that the penalized likelihood approach always performs better than either a clock-based method such as Langley-Fitch (CL) or a wholly nonparametric method, such as NPRS (Sanderson 1997
) whenever the data depart from constant rates. NPRS never achieves the low level of prediction error obtained by optimal smoothing in penalized likelihood. NPRS tends to overfit the data, in that the estimates of local substitution rates obtained from it tend to have higher variance than the variance found at optimal levels of smoothing. Even when the data are clocklike, penalized likelihood would select a value for smoothing that would be essentially clocklike itself.
NPRS outperformed CL in the real data, except in the rbcL first and second position partition. There the optimal smoothing level is nearly clocklike, and lower values of smoothing do progressively worse than assuming a clock, a result anticipated when data are nonclocklike, but there is simply not much of it (Sanderson 1997
; see Discussion).
Estimated Rates and Times as a Function of Smoothing Level
Specific results are illustrated with examples primarily from the two larger molecular data sets, SSU and PS. Detailed analyses of these data will be presented elsewhere. In all the data sets, higher levels of smoothing led to estimated absolute substitution rates that varied less from branch to branch, as indicated by the coefficient of variation of rates across branches (fig. 2 ). The rate variation for individual branches is shown for two contrasting examples: the branch subtending the Gnetales and the branch subtending angiosperms (fig. 3
). See figure 4 for the phylogenetic placement of these clades. Rates of substitution for the branch subtending Gnetales varied by a factor of 2–3 over the ranges of smoothing values described here, depending on the gene, whereas the range of variation for the branch subtending angiosperms was considerably less than that.
|
|
Differences in levels of smoothed rate estimates can be visualized across an entire tree in rate-calibrated phylogenies (fig. 4 ), which are tree diagrams in which branch lengths are drawn proportional to the absolute rates of substitution. These clearly show how smoothing can decrease the estimated variation in rate across the tree. For example, for first and second positions in PS, the rates of branches become steadily more similar to each other in moving from less smoothing to more smoothing. The rate along the branches leading to Huperzia and to Angiopteris, two low-rate lineages, increases, whereas the rate along the branch leading to Welwitschia and Ephedra, two high-rate lineages, decreases (albeit only slightly for this gene partition).
Estimated divergence times also depend on the choice of smoothing parameter. This is illustrated by the estimated age of the two nodes corresponding to the age of Gnetales and angiosperms (fig. 5 ). The estimated age of Gnetales is very sensitive to the smoothing level, showing a nearly monotonic increase in age as the model is made more clocklike, especially for the SSU data. The angiosperm age is a bit less sensitive to smoothing, but again the SSU data show relatively more sensitivity than the PS data.
|
Optimal Estimated Rates and Times
Given the results of the cross-validation analysis, estimates of ages and rates can also be obtained for each data set at the optimal level of smoothing. These are indicated for the exemplar branches and nodes in figures 3 and 5 . The variation in rates across each tree is substantial for all data sets (table 1 ): about twofold variation in rbcL first and second position data, threefold variation in the PS first and second position data, fivefold variation in the PS third position data, eightfold variation in the SSU data, and 10-fold variation in rbcL third position data.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Optimal Levels of Smoothing
In four of the five data partitions among the three data sets, cross-validation indicated an optimal level of smoothed rate variation that led to performance significantly better than that permitted by the assumption of a clock. In the fifth, rbcL first and second position, the improvement was only marginal but the method did no worse. These patterns can be understood intuitively. Consider the case in which rates vary substantially across a tree. For large
, the predicted values will essentially be the prediction under a molecular clock, but the observed number of substitutions along many branches will deviate from that expectation because rates are variable, leading to poor prediction of the observations in pruned branches. When is small, on the other hand, the model is overfit, and small changes in the data (i.e., subsamples of the data constructed during pruning) will lead to large changes in the parameter estimates. This leads, in turn, to a poor ability to predict pruned branches. For these reasons, one would expect some intermediate value of to give the best (smallest) CV score.
If the data are truly clocklike, on the other hand, CV should monotonically decrease with increasing , as it does for the simulated data in figure 1
, mainly because the noise in a constant-rate Poisson process causes unstable estimates of rates and times at low values of smoothing. However, if data are sufficiently noisy, as in the rbcL first and second positions, this pattern may well occur even for nonclocklike data because the simultaneous inference of both divergence times and absolute rates simply cannot be any more effective than a clock-based method. Previous simulation studies of the NPRS method (Sanderson 1997
) indicated that it performed best when the average numbers of substitutions on branches were relatively high. With less data, the assumption of a clock could often provide results as good as those with more sophisticated methods.
The estimation of is itself subject to error, of course. For the present paper, this error has been ignored, but it clearly will contribute to the error variance of the estimates of rates and divergence times, just as model misspecification does in conventional parametric inference. Determination of the error on is likely to be computationally expensive, at least if resampling methods are used.
Comparisons to Other Methods
Penalized likelihood outperformed CL and NPRS in every data set that departed from a clock, as long as cross-validation was used to determine the optimal level of smoothing. Even under the simulated clocklike data, cross-validation would lead the investigator to choose a level of smoothing that was clocklike and thereby retain optimal levels of prediction error. Penalized likelihood always outperforms NPRS, which tends to overfit the data, allowing too much rate variation and thereby losing predictive power. This does not, however, imply that NPRS is worse than assuming a clock. Usually NPRS is better, except perhaps in the case of few substitutions along branches.
The CV criterion proposed here provides an empirical method for comparing other recently proposed rate and time estimation procedures (Thorne, Kishino, and Painter 1998
; Huelsenbeck, Larget, and Swofford 2000
), which have mainly been evaluated on an absolute scale by simulation studies (Sanderson 1997
; Rambaut and Bromham 1998
). Studies of relative performance of methods on the same data should illuminate the strengths and weaknesses of these approaches.
Extensions
Penalized likelihood can be applied to better models of the substitution process, such as the full four-state Markov model commonly used in maximum likelihood phylogenetic inference—albeit at considerable computational cost. Unfortunately, the cross-validation procedures are computationally expensive. A tree with M terminal taxa will have M independent estimation steps corresponding to each pruning. The running time of the numerical algorithms is also polynomial in M. Thus, implementation of a full Markov model version of cross-validation will add one degree to the exponent in the running-time scaling factor. For anything but the smallest trees, addition of the full Markov model will require good algorithm engineering.
The simplification used here may not be too bad in general. Suitably corrected branch length estimates may be nearly sufficient, in a statistical estimation sense, to estimate rates and divergence times. The main problem may well be accounting properly for rate variation across sites, which seems to exert a very substantial influence on length estimates (Yang 1996
). Rate variation can be incorporated directly by using a negative binomial distribution for branch lengths rather than a Poisson as used here. If sites have rates that are chosen randomly from a gamma distribution—the usual approach (Yang 1996
)—then the probability of a substitution at any site is a negative binomial (Uzzell and Corbin 1971
), and the distribution of branch lengths for a particular branch is then the sum of R negative binomial distributions (where R is the number of sites), which is also a negative binomial.
The increasing availability of multigene data sets for many taxa suggests extensions of the model in a different direction. Different genes may well have different patterns of rate variation, but the divergence times are held in common between them (unless coalescence times of the separate gene trees differ significantly). A comprehensive model should then include a single set of divergence time parameters but a separate set of rate parameters for each gene. This could be easily implemented if all genes were subject to the same smoothing parameter, but if each gene is optimally smoothed to a different extent, then the cross-validation procedure must simultaneously optimize multiple smoothing parameters, a fairly daunting prospect. Obviously, the problem would be much easier if the divergence times themselves could be fixed prior to analysis. In that case, the problem would reduce to one that is much closer to semiparametric regression problems that are well characterized. Given enough data, this may become possible for some clades.
Rates of Evolution of Plant Genes
Recent studies have suggested dramatic differences in rates of molecular evolution among land plant lineages based on differences in branch lengths on estimated trees (Chaw et al. 2000
; Nickrent et al. 2000
; Sanderson et al. 2000
). This study confirms these inferences but places bounds on the rate variation. Certain lineages, especially Gnetales and some ferns, show much higher than average rates of evolution in all data sets and in both codon partitions of the protein-coding genes. Other lineages have much lower rates than average. The variation in rate is highest for rbcL third positions and SSU data, largely owing to extremely long branches in the Gnetales. In the PS genes, codon position partitions do not differ much in the level of variation in rate, despite the third codon partition mainly reflecting synonymous changes. The implications of this and its generality remain to be explored, but it agrees with previous findings in plant genes, which suggest strong lineage effects for both synonymous and nonsynonymous sites (Muse 2000
).
The interplay between estimates of divergence times and rates is exceedingly complex. The sensitivity of an age estimate for a node to the level of smoothing is partly influenced by the sensitivity of estimated rates in the local region around that node. For example, in the SSU data, the rate for the Gnetales branch varies over a factor of two along the smoothing axis and so does the age estimate. However, the estimated rate for the PS third position data along that branch shows the same level of sensitivity, and yet its age estimates are much less sensitive to different levels of smoothing. Clearly, the discovery of striking differences in substitution rates across land plants does not automatically dim the prospects for using molecular data to reconstruct ancient divergence times.
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Thanks are due to J. A. Doyle, J. Kim, S. Magallón, and M. F. Wojciechowski for useful suggestions. This research was supported by NSF grant 9726856.
|
Footnotes |
|---|
Michael Hendy, Reviewing Editor
Abbreviations: CL, clock; CV, cross-validation (criterion); NPRS, nonparametric rate smoothing; PS, photosystem; SAT, saturated; SSU, small subunit. 
Keywords: penalized likelihood
molecular clock
evolutionary rates
Address for correspondence and reprints: Section of Evolution and Ecology, One Shields Avenue, University of California, Davis, California 95616. mjsanderson{at}ucdavis.edu
.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Ayala J. A., A. Rzhetsky, F. J. Ayala, 1998 Origin of the metazoan phyla: molecular clocks confirm paleontological estimates Proc. Natl. Acad. Sci. USA 95:606-611
Bininda-Emonds O., S. G. Brady, J. Kim, M. J. Sanderson, 2001 Scaling of accuracy in extremely large phylogenetic trees Proc. Pacific Symp. Biocomputing 6:547–558
Britten R. J., 1986 Rates of DNA sequence evolution differ between taxonomic groups Science 231:1393-1398
Bromham L. D., M. D. Hendy, 2000 Can fast early rates reconcile molecular dates with the Cambrian explosion? Proc. R. Soc. Lond. B 267:1041-1047[Medline]
Burnham K. P., D. R. Anderson, 1998 Model selection and inference Springer-Verlag, New York
Chaw S.-M., C. L. Parkinson, Y. Cheng, T. M. Vincent, J. D. Palmer, 2000 Seed plant phylogeny inferred from all three plant genomes, monophyly of extant gymnosperms and origin of Gnetales from conifers Proc. Natl. Acad. Sci. USA 97:4086-4091
Cooper A., D. Penny, 1997 Mass survival of birds across the Cretaceous-Tertiary boundary: molecular evidence Science 275:1109-1113
Cox D. R., H. D. Miller, 1977 The theory of stochastic processes Chapman and Hall, London
Cutler D. J., 2000 Estimating divergence times in the presence of an overdispersed molecular clock Mol. Biol. Evol 17:1647-1660
Easteal S., G. Herbert, 1997 Molecular evidence from the nuclear genome for the time frame of human evolution J. Mol. Evol 44: (Suppl. 1) S121-S132
Gill P. E., W. Murray, M. H. Wright, 1981 Practical optimization Academic Press, New York
Gillespie J. H., 1991 The causes of molecular evolution Oxford University Press, New York
Green P. J., B. W. Silverman, 1994 Nonparametric regression and generalized linear models Chapman and Hall, London
Hasegawa M., H. Kishino, T. Yano, 1989 Estimation of branching dates among primates by molecular clocks of nuclear DNA which slowed down in Hominoidea J. Hum. Evol 18:461-476
Huelsenbeck J. P., B. Larget, D. Swofford, 2000 A compound Poisson process for relaxing the molecular clock Genetics 154:1879-1892
Kishino H., J. L. Thorne, W. J. Bruno, 2001 Performance of a divergence time estimation method under a probabilistic model of rate evolution Mol. Biol. Evol 18:352-361
Korber B., M. Muldoon, J. Theiler, F. Gao, R. Gupta, A. Lapedes, B. H. Hahn, S. Wolinsky, T. Bhattacharya, 2000 Timing the ancestor of the HIV-1 pandemic strains Science 288:1789-1796
Kumar S., S. B. Hedges, 1998 A molecular timescale for vertebrate evolution Nature 392:917-920
Langley C. H., W. Fitch, 1974 An estimation of the constancy of the rate of molecular evolution J. Mol. Evol 3:161-177[Web of Science][Medline]
Leitner T., J. Albert, 1999 The molecular clock of HIV-1 unveiled through analysis of a known transmission history Proc. Natl. Acad. Sci. USA 96:10752-10757
Lee M. S. Y., 1999 Molecular clock calibrations and metazoan divergence dates J. Mol. Evol 49:385-391[Web of Science][Medline]
Li W.-H., 1997 Molecular evolution Sinauer, Sunderland, Mass
Li W.-H., C.-I. Wu, 1985 Rates of nucleotide substitution are evidently higher in rodents than in man Mol. Biol. Evol 4:74-77[Web of Science][Medline]
Marshall C. R., 1990 Confidence intervals on stratigraphic ranges Paleobiology 16:1-10[Abstract]
Martin A. P., S. R. Palumbi, 1993 Body size, metabolic rate, generation time, and the molecular clock Proc. Natl. Acad. Sci. USA 90:4087-4091
Muse S. V., 2000 Examining rates and patterns of nucleotide substitution in plants Plant Mol. Biol 42:25-43[Web of Science][Medline]
Nickrent D. L., C. L. Parkinson, J. D. Palmer, R. J. Duff, 2000 Multigene phylogeny of land plants with special reference to bryophytes and the earliest land plants Mol. Biol. Evol 17:1885-1895
Pagel M., 1997 Inferring evolutionary processes from phylogenies Zool. Scripta 26:331-348
Press W. H., B. P. Flannery, S. A. Teukolsky, W. T. Vetterling, 1992 Numerical recipes in C. 2nd edition Cambridge University Press, New York
Rambaut A., L. Bromham, 1998 Estimating divergence data from molecular sequences Mol. Biol. Evol 15:442-448[Abstract]
Rodriguez F., J. L. Oliver, A. Marin, J. R. Medina, 1990 The general stochastic model of nucleotide substitution J. Theor. Biol 142:485-501[Web of Science][Medline]
Sanderson M. J., 1997 A nonparametric approach to estimating divergence times in the absence of rate constancy Mol. Biol. Evol 14:1218-1231[Web of Science]
———. 1998 Estimating rate and time in molecular phylogenies: beyond the molecular clock? Pp. 242–264 in P. Soltis, D. Soltis, and J. Doyle, eds. Plant molecular systematics. 2nd edition. Chapman and Hall, London
Sanderson M. J., M. J. Donoghue, 1996 Reconstructing shifts in diversification rates on phylogenetic trees Trends Ecol. Evol 11:15-20
Sanderson M. J., J. A. Doyle, 2001 Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 18S rDNA data Amer. J. Bot 88:1499-1516
Sanderson M. J., M. F. Wojciechowski, J.-M. Hu, T. Sher Khan, S. G. Brady, 2000 Error, bias, and long branch attraction in data for two chloroplast photosystem genes in seed plants Mol. Biol. Evol 17:782-797
Simonoff J. S., 1994 Smoothing methods in statistics Springer, New York
Smith A. B., D. T. J. Littlewood, 1994 Paleontological data and molecular phylogenetic analysis Paleobiology 20:259-273[Abstract]
Springer M., 1995 Molecular clocks and the incompleteness of the fossil record J. Mol. Evol 41:531-538
Swofford D. S., 1999 PAUP* 4.0: phylogenetic analysis using parsimony (*and other methods). Version 4b2 Sinauer Associates, Sunderland, Mass
Takezaki N., A. Rzhetsky, M. Nei, 1995 Phylogenetic test of the molecular clock and linearized trees Mol. Biol. Evol 12:823-833[Abstract]
Thorne J. L., H. Kishino, I. S. Painter, 1998 Estimating the rate of evolution of the rate of evolution Mol. Biol. Evol 15:1647-1657[Abstract]
Uyenoyama M. K., 1995 A generalized least-squares estimate for the origin of sporophytic self-incompatibility Genetics 139:975-992[Abstract]
Uzzell T., K. W. Corbin, 1971 Fitting discrete probability distributions to evolutionary events Science 172:1089-1096
Wray G. A., J. S. Levinton, L. H. Shapiro, 1996 Molecular evidence for deep precambrian divergences among metazoan phyla Science 274:568-573
Yang Z., 1996 Among-site rate variation and its impact on phylogenetic analyses Trends Ecol. Evol 11:367-372
Yoder A., Z. Yang, 2000 Estimation of primate speciation dates using local molecular clocks Mol. Biol. Evol 17:1081-1090
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Frajman, F. Eggens, and B. Oxelman Hybrid Origins and Homoploid Reticulate Evolution within Heliosperma (Sileneae, Caryophyllaceae)--A Multigene Phylogenetic Approach with Relative Dating Syst Biol, July 3, 2009; (2009) syp030v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Y. W. Ho and M. J. Phillips Accounting for Calibration Uncertainty in Phylogenetic Estimation of Evolutionary Divergence Times Syst Biol, July 3, 2009; (2009) syp035v1. [Full Text] [PDF]
|
|
|
|
 |
|
 S. Y.W. Ho An examination of phylogenetic models of substitution rate variation among lineages Biol Lett, June 23, 2009; 5(3): 421 - 424. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Antonelli, J. A. A. Nylander, C. Persson, and I. Sanmartin Tracing the impact of the Andean uplift on Neotropical plant evolution PNAS, June 16, 2009; 106(24): 9749 - 9754. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Q. Spinks and H. B. Shaffer Conflicting Mitochondrial and Nuclear Phylogenies for the Widely Disjunct Emys (Testudines: Emydidae) Species Complex, and What They Tell Us about Biogeography and Hybridization Syst Biol, May 28, 2009; (2009) syp005v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Kornilios, N. Poulakakis, M. Mylonas, and K. Vardinoyannis The phylogeny and biogeography of the genus Zonites Montfort, 1810 (Gastropoda: Pulmonata): preliminary evidence from mitochondrial data J. Mollus. Stud., May 1, 2009; 75(2): 109 - 117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Fawcett, S. Maere, and Y. Van de Peer From the Cover: Plants with double genomes might have had a better chance to survive the Cretaceous-Tertiary extinction event PNAS, April 7, 2009; 106(14): 5737 - 5742. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. D. Schrire, M. Lavin, N. P. Barker, and F. Forest Phylogeny of the tribe Indigofereae (Leguminosae-Papilionoideae): Geographically structured more in succulent-rich and temperate settings than in grass-rich environments Am. J. Botany, April 1, 2009; 96(4): 816 - 852. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Nadachowska and W. Babik Divergence in the Face of Gene Flow: The Case of Two Newts (Amphibia: Salamandridae) Mol. Biol. Evol., April 1, 2009; 26(4): 829 - 841. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Wang, M. J. Moore, P. S. Soltis, C. D. Bell, S. F. Brockington, R. Alexandre, C. C. Davis, M. Latvis, S. R. Manchester, and D. E. Soltis Rosid radiation and the rapid rise of angiosperm-dominated forests PNAS, March 10, 2009; 106(10): 3853 - 3858. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Herron, J. D. Hackett, F. O. Aylward, and R. E. Michod Triassic origin and early radiation of multicellular volvocine algae PNAS, March 3, 2009; 106(9): 3254 - 3258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. G. Moyle, C. E. Filardi, C. E. Smith, and J. Diamond Explosive Pleistocene diversification and hemispheric expansion of a "great speciator" PNAS, February 10, 2009; 106(6): 1863 - 1868. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J Givnish, K. C Millam, A. R Mast, T. B Paterson, T. J Theim, A. L Hipp, J. M Henss, J. F Smith, K. R Wood, and K. J Sytsma Origin, adaptive radiation and diversification of the Hawaiian lobeliads (Asterales: Campanulaceae) Proc R Soc B, February 7, 2009; 276(1656): 407 - 416. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Martin-Bravo, P. Vargas, and M. Luceno Is Oligomeris (Resedaceae) indigenous to North America? Molecular evidence for a natural colonization from the Old World Am. J. Botany, February 1, 2009; 96(2): 507 - 518. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Gentile, A. Fabiani, C. Marquez, H. L. Snell, H. M. Snell, W. Tapia, and V. Sbordoni From the Cover: An overlooked pink species of land iguana in the Galapagos PNAS, January 13, 2009; 106(2): 507 - 511. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Sauquet, P. H. Weston, C. L. Anderson, N. P. Barker, D. J. Cantrill, A. R. Mast, and V. Savolainen Contrasted patterns of hyperdiversification in Mediterranean hotspots PNAS, January 6, 2009; 106(1): 221 - 225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Magallon and A. Castillo Angiosperm diversification through time Am. J. Botany, January 1, 2009; 96(1): 349 - 365. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. E. Pfeil and M. D. Crisp The age and biogeography of Citrus and the orange subfamily (Rutaceae: Aurantioideae) in Australasia and New Caledonia Am. J. Botany, December 1, 2008; 95(12): 1621 - 1631. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Soria-Carrasco and J. Castresana Estimation of Phylogenetic Inconsistencies in the Three Domains of Life Mol. Biol. Evol., November 1, 2008; 25(11): 2319 - 2329. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.S Sequeira, M Sijapati, A.A Lanteri, and L Roque Albelo Nuclear and mitochondrial sequences confirm complex colonization patterns and clear species boundaries for flightless weevils in the Galapagos archipelago Phil Trans R Soc B, October 27, 2008; 363(1508): 3439 - 3451. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Smith and M. J. Donoghue Rates of Molecular Evolution Are Linked to Life History in Flowering Plants Science, October 3, 2008; 322(5898): 86 - 89. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Clark, R. H. Ree, M. E. Alfaro, M. G. King, W. L. Wagner, and E. H. Roalson A Comparative Study in Ancestral Range Reconstruction Methods: Retracing the Uncertain Histories of Insular Lineages Syst Biol, October 1, 2008; 57(5): 693 - 707. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Renner, G. W. Grimm, G. M. Schneeweiss, T. F. Stuessy, and R. E. Ricklefs Rooting and Dating Maples (Acer) with an Uncorrelated-Rates Molecular Clock: Implications for North American/Asian Disjunctions Syst Biol, October 1, 2008; 57(5): 795 - 808. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. F. Wright, E. E. Schirtzinger, T. Matsumoto, J. R. Eberhard, G. R. Graves, J. J. Sanchez, S. Capelli, H. Muller, J. Scharpegge, G. K. Chambers, et al. A Multilocus Molecular Phylogeny of the Parrots (Psittaciformes): Support for a Gondwanan Origin during the Cretaceous Mol. Biol. Evol., October 1, 2008; 25(10): 2141 - 2156. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S Thorpe, Y. Surget-Groba, and H. Johansson The relative importance of ecology and geographic isolation for speciation in anoles Phil Trans R Soc B, September 27, 2008; 363(1506): 3071 - 3081. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Bocak, M. Bocakova, T. Hunt, and A. P Vogler Multiple ancient origins of neoteny in Lycidae (Coleoptera): consequences for ecology and macroevolution Proc R Soc B, September 7, 2008; 275(1646): 2015 - 2023. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Scherson, R. Vidal, and M. J. Sanderson Phylogeny, biogeography, and rates of diversification of New World Astragalus (Leguminosae) with an emphasis on South American radiations Am. J. Botany, August 1, 2008; 95(8): 1030 - 1039. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Lemmon and E. M. Lemmon A Likelihood Framework for Estimating Phylogeographic History on a Continuous Landscape Syst Biol, August 1, 2008; 57(4): 544 - 561. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J. Harmon, J. Melville, A. Larson, and J. B. Losos The Role of Geography and Ecological Opportunity in the Diversification of Day Geckos (Phelsuma) Syst Biol, August 1, 2008; 57(4): 562 - 573. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M Percy, A. M Garver, W. L Wagner, H. F James, C. W Cunningham, S. E Miller, and R. C Fleischer Progressive island colonization and ancient origin of Hawaiian Metrosideros (Myrtaceae) Proc R Soc B, July 7, 2008; 275(1642): 1479 - 1490. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Mast, C. L. Willis, E. H. Jones, K. M. Downs, and P. H. Weston A smaller Macadamia from a more vagile tribe: inference of phylogenetic relationships, divergence times, and diaspore evolution in Macadamia and relatives (tribe Macadamieae; Proteaceae) Am. J. Botany, July 1, 2008; 95(7): 843 - 870. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Chapman, J. H. Leebens-Mack, and J. M. Burke Positive Selection and Expression Divergence Following Gene Duplication in the Sunflower CYCLOIDEA Gene Family Mol. Biol. Evol., July 1, 2008; 25(7): 1260 - 1273. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Pena and N. Wahlberg Prehistorical climate change increased diversification of a group of butterflies Biol Lett, June 23, 2008; 4(3): 274 - 278. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. L. Schardl, K. D. Craven, S. Speakman, A. Stromberg, A. Lindstrom, and R. Yoshida A Novel Test for Host-Symbiont Codivergence Indicates Ancient Origin of Fungal Endophytes in Grasses Syst Biol, June 1, 2008; 57(3): 483 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Merckx and M. I Bidartondo Breakdown and delayed cospeciation in the arbuscular mycorrhizal mutualism Proc R Soc B, May 7, 2008; 275(1638): 1029 - 1035. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J Peterson, J. A Cotton, J. G Gehling, and D. Pisani The Ediacaran emergence of bilaterians: congruence between the genetic and the geological fossil records Phil Trans R Soc B, April 27, 2008; 363(1496): 1435 - 1443. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. R. Schultz and S. G. Brady From the Cover: Major evolutionary transitions in ant agriculture PNAS, April 8, 2008; 105(14): 5435 - 5440. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. A. Nylander, U. Olsson, P. Alstrom, and I. Sanmartin Accounting for Phylogenetic Uncertainty in Biogeography: A Bayesian Approach to Dispersal-Vicariance Analysis of the Thrushes (Aves: Turdus) Syst Biol, April 1, 2008; 57(2): 257 - 268. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Mansion, G. Rosenbaum, N. Schoenenberger, G. Bacchetta, J. A. Rossello, and E. Conti Phylogenetic Analysis Informed by Geological History Supports Multiple, Sequential Invasions of the Mediterranean Basin by the Angiosperm Family Araceae Syst Biol, April 1, 2008; 57(2): 269 - 285. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. T. Burbrink and R. A. Pyron The Taming of the Skew: Estimating Proper Confidence Intervals for Divergence Dates Syst Biol, April 1, 2008; 57(2): 317 - 328. [Full Text] [PDF]
|
|
|
|
|
|
A. D. Cutter Divergence Times in Caenorhabditis and Drosophila Inferred from Direct Estimates of the Neutral Mutation Rate Mol. Biol. Evol., April 1, 2008; 25(4): 778 - 786. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. I. Smith, O. Pellmyr, D. M Althoff, M. Balcazar-Lara, J. Leebens-Mack, and K. A Segraves Pattern and timing of diversification in Yucca (Agavaceae): specialized pollination does not escalate rates of diversification Proc R Soc B, February 7, 2008; 275(1632): 249 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Jian, P. S. Soltis, M. A. Gitzendanner, M. J. Moore, R. Li, T. A. Hendry, Y.-L. Qiu, A. Dhingra, C. D. Bell, and D. E. Soltis Resolving an Ancient, Rapid Radiation in Saxifragales Syst Biol, February 1, 2008; 57(1): 38 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hardman and L. M. Hardman The Relative Importance of Body Size and Paleoclimatic Change as Explanatory Variables Influencing Lineage Diversification Rate: An Evolutionary Analysis of Bullhead Catfishes (Siluriformes: Ictaluridae) Syst Biol, February 1, 2008; 57(1): 116 - 130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. R. Smith and T. E. Dowling Correlating hydrographic events and divergence times of speckled dace (Rhinichthys: Teleostei: Cyprinidae) in the Colorado River drainage Geological Society of America Special Papers, January 1, 2008; 439(0): 301 - 317. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L Rabosky, S. C Donnellan, A. L Talaba, and I. J Lovette Exceptional among-lineage variation in diversification rates during the radiation of Australia's most diverse vertebrate clade Proc R Soc B, December 7, 2007; 274(1628): 2915 - 2923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. H Kozak and J. J Wiens Climatic zonation drives latitudinal variation in speciation mechanisms Proc R Soc B, December 7, 2007; 274(1628): 2995 - 3003. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Moore, C. D. Bell, P. S. Soltis, and D. E. Soltis Using plastid genome-scale data to resolve enigmatic relationships among basal angiosperms PNAS, December 4, 2007; 104(49): 19363 - 19368. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Raxworthy, C. M. Ingram, N. Rabibisoa, and R. G. Pearson Applications of Ecological Niche Modeling for Species Delimitation: A Review and Empirical Evaluation Using Day Geckos (Phelsuma) from Madagascar Syst Biol, December 1, 2007; 56(6): 907 - 923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Lepage, D. Bryant, H. Philippe, and N. Lartillot A General Comparison of Relaxed Molecular Clock Models Mol. Biol. Evol., December 1, 2007; 24(12): 2669 - 2680. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Cartwright and A. Collins Fossils and phylogenies: integrating multiple lines of evidence to investigate the origin of early major metazoan lineages Integr. Comp. Biol., November 1, 2007; 47(5): 744 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. C Ribas, R. G Moyle, C. Y Miyaki, and J. Cracraft The assembly of montane biotas: linking Andean tectonics and climatic oscillations to independent regimes of diversification in Pionus parrots Proc R Soc B, October 7, 2007; 274(1624): 2399 - 2408. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Britton, C. L. Anderson, D. Jacquet, S. Lundqvist, and K. Bremer Estimating Divergence Times in Large Phylogenetic Trees Syst Biol, October 1, 2007; 56(5): 741 - 752. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Knapp, R. Mudaliar, D. Havell, S. J. Wagstaff, and P. J. Lockhart The Drowning of New Zealand and the Problem of Agathis Syst Biol, October 1, 2007; 56(5): 862 - 870. [Full Text] [PDF]
|
|
|
|
 |
|
 C.-Y. Qiao, J.-H. Ran, Y. Li, and X.-Q. Wang Phylogeny and Biogeography of Cedrus (Pinaceae) Inferred from Sequences of Seven Paternal Chloroplast and Maternal Mitochondrial DNA Regions Ann. Bot., September 1, 2007; 100(3): 573 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. G. Pringle, S. W. Baxter, C. L. Webster, A. Papanicolaou, S. F. Lee, and C. D. Jiggins Synteny and Chromosome Evolution in the Lepidoptera: Evidence From Mapping in Heliconius melpomene Genetics, September 1, 2007; 177(1): 417 - 426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. G. Barraclough, D. Fontaneto, C. Ricci, and E. A. Herniou Evidence for Inefficient Selection Against Deleterious Mutations in Cytochrome Oxidase I of Asexual Bdelloid Rotifers Mol. Biol. Evol., September 1, 2007; 24(9): 1952 - 1962. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. F. Hugall, R. Foster, and M. S. Y. Lee Calibration Choice, Rate Smoothing, and the Pattern of Tetrapod Diversification According to the Long Nuclear Gene RAG-1 Syst Biol, August 1, 2007; 56(4): 543 - 563. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. A. Cranston and B. Rannala Summarizing a Posterior Distribution of Trees Using Agreement Subtrees Syst Biol, August 1, 2007; 56(4): 578 - 590. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Rutschmann, T. Eriksson, K. A. Salim, and E. Conti Assessing Calibration Uncertainty in Molecular Dating: The Assignment of Fossils to Alternative Calibration Points Syst Biol, August 1, 2007; 56(4): 591 - 608. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Thorne, S. C. Choi, J. Yu, P. G. Higgs, and H. Kishino Population Genetics Without Intraspecific Data Mol. Biol. Evol., August 1, 2007; 24(8): 1667 - 1677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Yang PAML 4: Phylogenetic Analysis by Maximum Likelihood Mol. Biol. Evol., August 1, 2007; 24(8): 1586 - 1591. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Hug and A. J. Roger The Impact of Fossils and Taxon Sampling on Ancient Molecular Dating Analyses Mol. Biol. Evol., August 1, 2007; 24(8): 1889 - 1897. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Pirie, M. P. B. Vargas, M. Botermans, F. T. Bakker, and L. W. Chatrou Ancient paralogy in the cpDNA trnL-F region in Annonaceae: implications for plant molecular systematics Am. J. Botany, June 1, 2007; 94(6): 1003 - 1016. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Marjanovic and M. Laurin Fossils, Molecules, Divergence Times, and the Origin of Lissamphibians Syst Biol, June 1, 2007; 56(3): 369 - 388. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. S. Manos, P. S. Soltis, D. E. Soltis, S. R. Manchester, S.-H. Oh, C. D. Bell, D. L. Dilcher, and D. E. Stone Phylogeny of Extant and Fossil Juglandaceae Inferred from the Integration of Molecular and Morphological Data Sets Syst Biol, June 1, 2007; 56(3): 412 - 430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-L. Lee, R. K. Jansen, T. W. Chumley, and K.-J. Kim Gene Relocations within Chloroplast Genomes of Jasminum and Menodora (Oleaceae) Are Due to Multiple, Overlapping Inversions Mol. Biol. Evol., May 1, 2007; 24(5): 1161 - 1180. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J Wiens, G. Parra-Olea, M. Garcia-Paris, and D. B Wake Phylogenetic history underlies elevational biodiversity patterns in tropical salamanders Proc R Soc B, April 7, 2007; 274(1612): 919 - 928. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Waters, D. L. Rowe, S. Apte, T. M. King, G. P. Wallis, L. Anderson, R. J. Norris, D. Craw, and C. P. Burridge Geological Dates and Molecular Rates: Rapid Divergence of Rivers and Their Biotas Syst Biol, April 1, 2007; 56(2): 271 - 282. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. B. Smith Intrinsic versus extrinsic biases in the fossil record: contrasting the fossil record of echinoids in the Triassic and early Jurassic using sampling data, phylogenetic analysis, and molecular clocks Paleobiology, March 1, 2007; 33(2): 310 - 323. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. K. Gonder, H. M. Mortensen, F. A. Reed, A. de Sousa, and S. A. Tishkoff Whole-mtDNA Genome Sequence Analysis of Ancient African Lineages Mol. Biol. Evol., March 1, 2007; 24(3): 757 - 768. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-J. Wang, J.-Q. Liu, and G. Miehe Phylogenetic Origins of the Himalayan Endemic Dolomiaea, Diplazoptilon and Xanthopappus (Asteraceae: Cardueae) Based on Three DNA Regions Ann. Bot., February 1, 2007; 99(2): 311 - 322. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. D. Akhunov, A. R. Akhunova, and J. Dvorak Mechanisms and Rates of Birth and Death of Dispersed Duplicated Genes during the Evolution of a Multigene Family in Diploid and Tetraploid Wheats Mol. Biol. Evol., February 1, 2007; 24(2): 539 - 550. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Roelants, D. J. Gower, M. Wilkinson, S. P. Loader, S. D. Biju, K. Guillaume, L. Moriau, and F. Bossuyt Global patterns of diversification in the history of modern amphibians PNAS, January 16, 2007; 104(3): 887 - 892. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Willyard, J. Syring, D. S. Gernandt, A. Liston, and R. Cronn Fossil Calibration of Molecular Divergence Infers a Moderate Mutation Rate and Recent Radiations for Pinus Mol. Biol. Evol., January 1, 2007; 24(1): 90 - 101. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. G.P Ericson, C. L Anderson, T. Britton, A. Elzanowski, U. S Johansson, M. Kallersjo, J. I Ohlson, T. J Parsons, D. Zuccon, and G. Mayr Diversification of Neoaves: integration of molecular sequence data and fossils Biol Lett, December 22, 2006; 2(4): 543 - 547. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Marshall, C. Simon, and T. R. Buckley Accurate Branch Length Estimation in Partitioned Bayesian Analyses Requires Accommodation of Among-Partition Rate Variation and Attention to Branch Length Priors Syst Biol, December 1, 2006; 55(6): 993 - 1003. [Full Text] [PDF]
|
|
|
|
|
|
S. G. Brady, T. R. Schultz, B. L. Fisher, and P. S. Ward From the Cover: Evaluating alternative hypotheses for the early evolution and diversification of ants PNAS, November 28, 2006; 103(48): 18172 - 18177. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. H. Crozier Charting uncertainty about ant origins PNAS, November 28, 2006; 103(48): 18029 - 18030. [Full Text] [PDF]
|
|
|
|
|
|
T. J. Near, S. K. Parker, and H. W. Detrich III A Genomic Fossil Reveals Key Steps in Hemoglobin Loss by the Antarctic Icefishes Mol. Biol. Evol., November 1, 2006; 23(11): 2008 - 2016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. HERMSEN and J. R. HENDRICKS THE HIERARCHY OF TIME Palaios, October 1, 2006; 21(5): 403 - 405. [Full Text] [PDF]
|
|
|
|
|
|
N. Wahlberg That Awkward Age for Butterflies: Insights from the Age of the Butterfly Subfamily Nymphalinae (Lepidoptera: Nymphalidae) Syst Biol, October 1, 2006; 55(5): 703 - 714. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Yesson and A. Culham Phyloclimatic Modeling: Combining Phylogenetics and Bioclimatic Modeling Syst Biol, October 1, 2006; 55(5): 785 - 802. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Smith, D. Pisani, J. A. Mackenzie-Dodds, B. Stockley, B. L. Webster, and D. T. J. Littlewood Testing the Molecular Clock: Molecular and Paleontological Estimates of Divergence Times in the Echinoidea (Echinodermata) Mol. Biol. Evol., October 1, 2006; 23(10): 1832 - 1851. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Crayn, M. Rossetto, and D. J. Maynard Molecular phylogeny and dating reveals an Oligo-Miocene radiation of dry-adapted shrubs (former Tremandraceae) from rainforest tree progenitors (Elaeocarpaceae) in Australia Am. J. Botany, September 1, 2006; 93(9): 1328 - 1342. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z.-L. Nie, H. Sun, P. M. Beardsley, R. G. Olmstead, and J. Wen Evolution of biogeographic disjunction between eastern Asia and eastern North America in Phryma (Phrymaceae) Am. J. Botany, September 1, 2006; 93(9): 1343 - 1356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. L. Pereira and A. J. Baker A Mitogenomic Timescale for Birds Detects Variable Phylogenetic Rates of Molecular Evolution and Refutes the Standard Molecular Clock Mol. Biol. Evol., September 1, 2006; 23(9): 1731 - 1740. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Pons, T. G. Barraclough, J. Gomez-Zurita, A. Cardoso, D. P. Duran, S. Hazell, S. Kamoun, W. D. Sumlin, and A. P. Vogler Sequence-Based Species Delimitation for the DNA Taxonomy of Undescribed Insects Syst Biol, August 1, 2006; 55(4): 595 - 609. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Won and S. S. Renner Dating Dispersal and Radiation in the Gymnosperm Gnetum (Gnetales)--Clock Calibration When Outgroup Relationships Are Uncertain Syst Biol, August 1, 2006; 55(4): 610 - 622. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kullberg, M. A. Nilsson, U. Arnason, E. H. Harley, and A. Janke Housekeeping Genes for Phylogenetic Analysis of Eutherian Relationships Mol. Biol. Evol., August 1, 2006; 23(8): 1493 - 1503. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Jakob and F. R. Blattner A Chloroplast Genealogy of Hordeum (Poaceae): Long-Term Persisting Haplotypes, Incomplete Lineage Sorting, Regional Extinction, and the Consequences for Phylogenetic Inference Mol. Biol. Evol., August 1, 2006; 23(8): 1602 - 1612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G Brady, S. Sipes, A. Pearson, and B. N Danforth Recent and simultaneous origins of eusociality in halictid bees Proc R Soc B, July 7, 2006; 273(1594): 1643 - 1649. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. RYDIN, K. R. PEDERSEN, P. R. CRANE, and E. M. FRIIS Former Diversity of Ephedra (Gnetales): Evidence from Early Cretaceous Seeds from Portugal and North America Ann. Bot., July 1, 2006; 98(1): 123 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J Roger and L. A Hug The origin and diversification of eukaryotes: problems with molecular phylogenetics and molecular clock estimation Phil Trans R Soc B, June 29, 2006; 361(1470): 1039 - 1054. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||