BIASES AND BIOLOGY PRODUCE SPURIOUS EARLY MOLECULAR DIVERGENCE DATES FOR THE METAZOAN RADIATION – A TEST BY OUTGROUP COMPARISON

Many have assumed a molecular clocks and extrapolated 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 the nematode C. elegans 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. Genes known in basal Metazoa were compared in a pairwise manner with Mouse and Fly homologues and then with Mouse and Nematode homologues. This analysis demonstrates that the Mouse/Vertebrate genes evolve at a slower rate than the invertebrate model taxa. Thus rate effects are pervasive features of lineages, not the problem of a few genes with anomalous rates.