"PRINCIPLES OF PHYLOGENETICS: ECOLOGY AND EVOLUTION" Integrative Biology 200B Spring 2011 The Fossil Record and Dating: “How fast, as a matter of fact, do animals evolve in nature? That is the fundamental observational problem of tempo in evolution. It is the first question the geneticist asks the paleontologist. Some attempt to answer it is a necessary preliminary for the whole consideration of tempo and mode.” (Simpson 1944, p. 3) This is an area of intense recent interest, yet in need of much further thought and research. There are two fundamental steps in the process of putting time onto a node or a branch of a tree: 1. Establishing the clock What is ultrametricity? Ultrametric matrix and its tree: A non-ultrametric tree: from: http://www.diku.dk/~pawel/comp-bio/ev_trees/intro/intro/ultrametric.html A. Determining whether your data fit a clock model i. relative rate tests comparing three taxa at a time, in rooted context: GRate&‐&relative&rate&tests&with&groups&support ing&all&models&of&nucleotide&substitution&known&by&PAUP*&http://www.softpedia.com/get/Science-CAD/GRate.shtml ii. likelihood ratio test Testing the Molecular Clock using a likelihood ratio test (courtesy of John Huelsenbeck) Under the null hypothesis, the phylogeny is ultrametric (i.e., rooted and the branch lengths are constrained such that all of the tips can be drawn at a single time plane). Under the alternative hypothesis, each branch is allowed to vary independently. The alternative hypothesis invokes s - 2 additional parameters, where s is the number of sequences. The likelihood ratio test statistic is -2logL = 2(logL0 - logL1), where L0 and L1 are the likelihoods under the null and alternative hypotheses, respectively. The significance of the likelihood ratio test statistic can be approximated using a χ2 distribution (with s - 2 degrees of freedom) or by parametric bootstrapping. The following example shows how to perform the likelihood ratio test of the molecular clock. The data are s = 5 albumin sequences from vertebrates (a fish, frog, bird, mouse, and human). We assume the Hasegawa, Kishino, and Yano (1985) model of DNA substitution with among site rate variation described using a gamma distribution. The maximum likelihood under the null hypothesis is logL0 = -7585.343. The best estimate of phylogeny supports the monophyly of the mammals and amniotes. The maximum likelihood under the alternative hypothesis is logL1 = -7569.052. The likelihood under the alternative hypothesis is higher than under the null hypothesis because there are more free parameters in the substitution model (i.e., no constraints on branch lengths). The maximum likelihood estimate of phylogeny is consistent with the monophyly of mammals and amniotes (though the tree is unrooted). The likelihood ratio test statistic is -2logL = 32.582, which is asymptotically χ2 distributed under the null hypothesis with 3 degrees of freedom. Comparing the observed value of -2logL to a χ2 with 3 df shows that the null hypothesis can be rejected at P < 0.001. So, we conclude the data are not clock-like. B. If your data don't fit a clock model (and they usually don't), try smoothing the data to get an (at least locally) approximate clock. Two common methods (implemented in r8s by Mike Sanderson: http://loco.biosci.arizona.edu/r8s/index.html), both attempt to smooth the magnitude of changes in rate between neighboring branches, to give you something intermediate between the rigid clock assumption and completely unconstrained branch lengths: 1. non-parametric rate smoothing. This uses a least squares smoothing approach that penalizes rates that change too quickly from branch to neighboring branch. 2. penalized likelihood. This is a "semi-parametric" approach that combines a ML approach with the above penalty function. The user can specify the relative weight of the penalty function and the ML component (in which parameters are being fitted as typical in ML). The parametric model has a different substitution rate for each branch. 2. Calibrating the clock Some folks simply import a "known" rate from the literature into their analysis -- don't do this! You need to come up with a calibration from your analysis. A. Three ways that have been used to estimate the age of a node i. a fossil (see below for details) -- gives a minimum age for a nodeii. availability of necessary habitat -- gives a maximum age for a node (maybe) iii. geographic vicariance event -- neither a maximum or minimum age for a node B. How to use a fossil to date a node? Some principles: i. you never find a taxon in the fossil record, or a lineage; you find remains of an organism displaying some characters. These characters need to be analyzed using the principles talked about earlier in class, in relation to other fossils and extant organisms in the group. ii. therefore, a fossil can never be compared to a strictly molecular phylogeny (unless it has preserved molecular data!); all relevant morphological characters need to have been analyzed and incorporated in the phylogenetic reconstruction. iii. when a fossil can be placed using synapomorphies as sister to some other lineage, that other lineage (and the node connecting them) must be at least as old as the fossil. Nodes deeper must also have been in existence by that time. This is the important principle of equal age of sister groups. 3. Getting into the fossil record when fossil is placed here, it dates all the nodes with circles on them, but it doesn't date the nodes with an asterisk * *SHRIMP = sensitive high-resolution ion microprobe (U-Pb) IDTIMS = isotope dilution analysis using thermal ion mass spectrometry (U-Pb) U-Pb = Uranium Lead Ar-Ar = 40Argon 39 Argon Rb-Sr = rubidium-strontium fission track At Cambrian boundary (542 Ma) IDTIMS dates providing dates at error levels of ± 0.3 Ma.3. Why do molecular clock dates often indicate earlier divergence dates? The Signor-Lipps
View Full Document
Unlocking...