The idea of being able to estimate dates, even if only roughly, from DNA or aminoacid sequences is a very appealing one. Ever since Zuckerkandl and Pauling (1965) suggested that mutations in proteins and their corresponding genes accumulate in a clock-like fashion, i.e., that they accumulate over time in a regular manner, biologists have been trying to date events using molecular data. There are many insights in modern biology which are based on these molecular-derived time estimates such as reconstruction of the history of epidemics from extant viral samples (Korber et al. 2000), the calculation of the dates of origin of selected groups of organisms (Cooper and Penny 1997, Bromham et al. 1999), the appraisal of rates of diversification within selected clades (Baldwin and Sanderson 1998, Sato et al. 2001), reconstruction of sources of colonisation and changes in population size (Griswold and Baker 1997), the effects of glacial cycles on speciation (Klicka and Zink 1997) and many more. Most biologists accept molecular clocks at least at a local level, i.e., the idea that some genes behave in a clock-like fashion across a narrow spectrum of taxa (Li 1993). All molecular clocks have to be calibrated with reference dates derived either from the fossil record or from geological events, but clock calibrations remain few and sparse as fossils are often lacking and few geological events are clear-cut enough to be suited for the task. Hence many studies still rely heavily on universal clocks. The best example is the ubiquitous assumption that animal mitochondrial DNA (mtDNA) evolves with a rate of approximately 2% sequence divergence per million years (myr) (Brown et al. 1979). In birds, this rate was originally estimated from RFLP data in geese (Shields and Wilson 1987). Some DNA-sequence studies found a similar rate in geese and other birds, although several more recent studies have come up with rather different estimates (Fig. 1). Nowadays the 2% rate is commonly applied to avian studies, but how truly universal is this clock among birds? Only a handful of avian mitochondrial clock calibrations have been attempted (reviewed by Lovette 2004) because the fossil record is relatively poor for most avian lineages. Nevertheless, although most calibrations cluster around the 2% value, there is considerable variance around that value (Fig. 1). It can be seen from Fig. 1 that calibrations based on events that occurred within the last million years (young calibrations) are faster than those based on events or fossils older than one million years (old calibrations; MannWhitney U test P /0.0082). How can these differences be explained, since a correlation between evolutionary rate and calibration date is not expected a priori? With one exception, young calibrations are based on island colonisations by small nectarivorous birds, namely sunbirds (Nectariniidae; Warren et al. 2003), and Hawaiian honeycreepers (Drepaniidae; Tarr and Fleischer 1993, Fleischer et al. 1998). It could be that estimates based on these taxa are biased, since both sunbirds and honeycreepers are small passerine birds with short generation times and high metabolic rates (generation times and metabolic rates have been proposed as a source of heterogeneity in evolutionary rates; see e.g. Kohne 1970, Sibley et al. 1988, Martin and Palumbi, 1993). It could also be that island populations are more divergent from ancestral populations than neighbouring continental populations are because of their unique population histories, undergoing genetic bottlenecks and founder effects (Carson and Templeton 1984, Thorpe et al. 1994), which would result in an overestimation of the clock calibration. Also, as it is difficult to pinpoint the actual island colonisation event, and given the high vagility of birds, it is often assumed that islands are colonised as early as they emerge, resulting in a minimal estimate of JOURNAL OF AVIAN BIOLOGY 35: 1 /4, 2004
Read full abstract
Feb 17, 2024
Leo A Featherstone + 3
Open Access
