Estimation of divergence times when evolutionary rate varies.

     Inference of divergence times among macro molecules is one of the major subjects in molecular phylogenetic analyses to elucidate evolutionary histories of genes, genomes and species. Inference of divergence times requires either an assumption of constant rate throughout the tree (a molecular clock) or a statistical distribution to model the variation of evolutionary rates among lineages. Widely used Bayesian methods use a probability distribution of evolutionary rates in the tree (e.g., lognormal distribution) and whether the rates among lineages are correlated or independent. In contrast, our newly developed approach (RelTime) does not require such a probability distribution but estimates relative rates throughout the tree to generate relative node ages that can transformed into absolute dates by using temporal constrains for one or more nodes. RelTime has performed well in timetree analysis of many large empirical datasets, and it shows high accuracy in computer simulations. In this seminar, I would like introduce current state of the molecular clock applications as well as our RelTime approach.