Deciphering the histories of lineages as they have unfolded over billions of years, and placing them in the context of the known fossil record and Earth’s biogeochemical events, is a challenging task in evolutionary biology. Shortly after discovering the molecular structure of DNA and determining the first amino acid sequences of proteins, it was recognized that the evolutionary distances among lineages could be estimated from the changes observed in nucleotide and amino acid sequences (substitutions). Combined with the hypothesis that steady accumulation of substitutions over time is analogous to the “ticking” of a clock and hence the moniker “molecular clock” (1), the relative divergence times among lineages could be estimated from the genetic distances among them. Estimating the times of the major evolutionary events in the history of Earth’s biosphere could be made when the branches of the evolutionary tree are calibrated with available fossil evidence, allowing conversion of the relative divergence times to the absolute estimates (2). Realization that the molecular clock “ticks” unevenly within and across genes, among different lineages and over time, as well as a scarce and often controversial fossil record, ultimately led to a development of a new class of Bayesian “relaxed clock” models (3). These models do not require all lineages to evolve at the same rate and allow incorporation of heterogeneous (as well as vague) prior information from multiple genes, a large number of taxa, and multiple calibration points (3). Despite these methodological advances, dating various past evolutionary events has continued to produce conflicting estimates and wide uncertainty intervals. In PNAS, Shih and Matzke (4) introduce a clever approach to decrease the uncertainty associated with the inferred dates via additional constraints provided by anciently duplicated gene families.
展开▼
