JACOBS, David K: Operational and biological factors bias molecular rate calculation leading to spurious early divergence dates for the metazoan radiation.

Many have employed an assumption of a molecular clocks combined with statistical methods to extrapolate from known tie points in the fossil record to infer divergence times. These methods applied to amino acid sequence data have recently been used to infer the timing of the metazoan radiation leading to substantially older divergence dates than implied by fossil evidence. However, aspects of taxon sampling influence sequence alignment and phylogeny reconstruction – steps that are required to calculate rates. Imbalance between well-sampled clades used to infer rates and less well sampled outgroups leads to systematic biases that result in erroneously large basal branch lengths. In addition to such operational issues, recent work relating genome size to deletion processes that eliminate DNA suggesting that genome size and rate of disruption of the genome by mutation are inversely correlated. Thus rapid development and small genome size, desirable characteristic of model systems, are associated with high rates of molecular evolution. Much invertebrate data used to calculate divergence dates is from model systems such as fruit flies and nematodes that are high rate, whereas vertebrate calibration data are not. These two sorts of biases are likely to act in concert leading to spuriously ancient dates of divergence. Given that these biases operate in the same direction leading to older dates, they may account for much or all of the apparent disparity between divergence times based on fossil and molecular data. Until these classes of bias have been addressed in these data, molecular clocks will not be useful in constraining the timing of the metazoan radiation.