|
|
|
Welcome
Research interests Publications Curriculum vitae Software and manuals Contact |
Research interests Application and comparison of molecular dating methods to answer selected biogeographic and evolutionary questions related to specific plant taxa
Based on this central idea, molecular dating has been used in inumberable studies as one mean to investigate mechanisms and processes of evolution. For example, the Early Cambrian origin of the main phyla of animals (Cambrian explosion; Wray et al. 1996; Smith and Peterson 2002; Aris-Brosou and Yang 2003), the replacement of dinosaurs by modern birds and mammals in late Tertiary (Madsen 2001), or the age of the last common ancestor of the main pandemic strain of human immunodeficiency virus (HIV; Korber et al. 2000) have been investigated using molecular dating. In plants, molecular dating methods have been used to infer the origin and age of the Angiosperms (Wikström et al., 2001; Magallon-Puebla et al. 1999; Magallon 2001; Sanderson and Doyle, 2001) and - on a lower taxonomic level - in many other studies. As sequences from multiple species began to accumulate during the 1970's, it became apparent that a clock is not a particularly good metaphor for the process of molecular evolution (Langley and Fitch 1974). Variation in rates of sequence substitution, both along a lineage and between diffent lineages, is known to be pervasive (Britten 1986; Gillespie 1986; Li 1997). During the last years, different dating algorithms have been developed which are able to count for rate variation along lineages: Non-parametric rate smoothing (NPRS; Sanderson 1997) and penalized likelihood (PL; Sanderson, 2002) both relax the assumption of rate constancy by smoothing changes of substition rates across the tree. NPRS is an entirely non-parametric method that estimates rates and times via a least-squares smoothing criterion, whereas PL is a semi-parametric technique that attempts to combine the statistical power of parametric methods with the robustness of non-parametric methods. Both related methods are implemented in the software r8s 1.7 (Sanderson 2003). Bayesian dating methods (Thorne et al. 1998) use probabilistic models to describe the change in evolutionary rate over time and use the Markov chain Monte Carlo (MCMC) procedure to derive the posterior distribution of rates and times. They allow simultaneous use of different substitution models for multiple data partitions as well as multiple calibration windows, and provide direct credibility intervals for estimated divergence times and substitution rates. Did Crypteroniaceae disperse out-of-India?
|
Linus Pauling
