The molecular clock hypothesis proposes that evolution at the sequence level occurs at a more-or-less constant rate. This hypothesis is demonstrably false: for example, the rate of substitution in anthozoan (corals etc) mitochondrial DNA is a hundred-fold slower than other animals. Yet such glaring counter-examples are swept aside with disturbing alacrity when the clock is used to draw inferences about the action of selection, or evolutionary history. Ignoring this variation in the tick of the molecular clock has been possible because, apart from such extreme examples, it is not easy to detect and analyse. One major difficulty is that the time over which genetic change has accumulated is imprecisely known. In addition to problems with interpreting the fossil record, different genes from the same species can have ancestors who lived millions of years apart1. For example only 60% of human genes have their closest relative in chimps. Most of the remainder share a more recent common ancestor with gorillas (X in the Figure) despite the extra 2-million years of evolution to the human-gorilla ancestor. Hence the human and chimp lineages must have remained separate for that entire period (T3).

The most recent Bayesian methods for phylogeny reconstruction attempt to allow for the unknown rate variation, but since we know little of the underlying pattern of rate variation, we cannot evaluate their basic design², let alone their accuracy or precision. The genomic data that have recently accumulated provide an exciting and unexploited opportunity to investigate the fundamental evolutionary processes affecting this variation, because sequences from multiple loci can now be obtained from many species. For example, the decreased anthozoan mtDNA rate has been attributed to a change in mitochondrial mutation, because nuclear genes show no such decrease. Similarly, the variance between rates of nuclear genes provided information on the effective population size of the human-chimp ancestor, which determined the rate at which lineages coalesced.

Previous work has used a few species (e.g. the human chimp gorilla comparison) or a few genes (e.g. the anthozoan comparison). This project exploits the explosion in the number of sequenced genomes to allow an order of magnitude increase in number of species and genes in the comparisons. Qualitatively different results can be obtained from this type of data. By comparing the same species-pair at many loci we have the power to identify lineage-specific effects, and comparing the same genes over many different species pairs, locus-specific effects. There are at least three sets of sequenced genomes that are sufficiently closely related to shed light on the fundamental population genetic parameters (variation in the mean and variance of mutation rate, and effective population size) using thousands of orthologous genes across entire genomes. These include 12 genomes of Drosophila spp., 28 genomes of vertebrates, and 19 genomes of hemiascomycete yeasts. One explanation for the incongruence between trees at different loci is the effect shown in the figure; our estimates of ancestral population sizes will allow us to predict its effects and so construct better models of the other possible sources of incongruence³.

The project will be supervised jointly by Dr James Cotton and Prof. Richard Nichols. The student will be equipped with computational skills and the theoretical understanding to contribute to this essential field of post-genomic science. This will include training in the use of bioinformatics software and tools, and in the programming necessary to automate analyses at the genomic scale, together with exposure to the population genetic methods required to draw inferences from the rates of substitution and their variation. It will suit a student interested in evolutionary biology, and, while programming skills are not a prerequisite, a reasonably sophisticated familiarity with computers and a willingness to learn basic programming skills are essential. The results of this project will overcome fundamental obstacles to the use of molecular clock techniques in studies of speciation, biogeography and many other areas of macroevolution, and more importantly, provide novel insights into the basic processes of molecular evolution.

References (click on pdf icon to download papers)

1. Nichols, R. A. (2001) Gene trees and species trees are not the same. Trends in Ecology and Evolution 16:358-364. 

2. Pulquério, M. J. F. and Nichols, R. A. (2007) Dates from the molecular clock: how wrong can we be? Trends in Ecology and Evolution 22:180-184. 

3. Cotton, J. A. and Page, R. D. M. (2004) Tangled trees from molecular markers in Phylogenetic Supertrees. Kluwer, Dordrecht.