Estimating divergence times among animal phyla has been confounded by an incomplete understanding of phylogenetic relationships within the kingdom. In recent years, a new phylogeny of animals has been proposed, based predominantly on small subunit ribosomal RNA sequence analyses, that divides the bilaterally-symmetric animals (bilaterians) into three main groups: deuterostomes, edysozoans, and lophotrochozoans (Aguinaldo et al., 1997; de Rosa et al., 1999; Mallatt et al., 2004). Despite a lack of strong statistical support for Ecdysozoa, this new animal phylogeny has had a large influence on studies of metazoan evolution and development, and has led some researchers to suggest that the last common ancestor of the bilaterians may have been a complex organism (Balavoine and Adoutte, 2003). Other studies using larger numbers of genes have supported and refuted certain aspects of this new phylogeny (Blair et al., 2002; Wolf et al, 2004; Philip et al, 2005; Philippe et al., 2005). Perhaps most controversial has been the position of nematodes (round worms) and platyhelminths (flatworms). These two phyla lack true body cavities and are traditionally placed basal to most other bilaterian phyla. Molecular evidence has mostly supported the elevation of platyhelminths (excluding Acoela) into protostomes (specifically within Lophotrochozoa), but there is currently no consensus as to the position of nematodes.

A number of studies over the past four decades have used molecular clocks to time divergences among animal phyla (e.g., Brown et al, 1972; Runnegar, 1982a; Wray et al, 1996; Hedges et al, 2004; Blair et al, 2005). Such analyses have consistently indicated deep origins for animal phyla (~800-1200 Ma), much earlier than predicted by the fossil record (i.e., Cambrian Explosion, ~520 Ma). Recently, some studies have proposed substantially younger molecular time estimates (Aris-Brosou and Yang, 2002, 2003; Douzery et al, 2004; Peterson et al, 2004; Peterson and

Butterfield, 2005). These studies claimed that through careful consideration of potential biases in both rate modelling and calibration, they produced molecular divergence times that were consistent (or more so) with the fossil record. However, upon closer inspection, these studies suffer from methodological biases that cast doubt on their results.

Most recent molecular clock analyses (e.g., Douzery et al., 2004; Hedges et al., 2004; Blair and Hedges, 2005a; Blair et al., 2005) have used sequences concatenated from multiple genes, thus avoiding potential statistical biases from averaging multiple single-gene estimates (the "mean of the ratios" problem). However, the criticism (Rodriguez-Trelles et al., 2002) that previous studies (e.g., Wang et al., 1999) were biased in that manner is incorrect, because those studies addressed asymmetry in distributions of time estimates by using medians and modes, and eliminating outliers. Even without such corrections, the simulations of Rodriguez-Trelles et al. (2002) showed that there was relatively little bias under most normal conditions (parameters). Also, the results from concatenated-gene studies (e.g., Hedges et al., 2004) corroborated the results of those earlier studies (e.g., Wang et al, 1999), indicating that such statistical biases are not responsible for old (~1 Ga) divergence time estimates among animal phyla.

Differences in how fossil calibrations are applied probably explain most of the variation in time estimates among studies. As discussed above (section 3.2), one study estimating animal divergence times (Douzery et al, 2004) used fossil time constraints that were substantially younger than the fossils themselves, producing artificially younger time estimates. In some cases (Peterson et al, 2004; Peterson and Butterfield, 2005), younger divergence times among animals were attributed to the use of calibrations from the invertebrate fossil record, rather than from vertebrates. However, other studies have also used invertebrate fossil calibrations and did not recover such young divergence times among animal phyla (Hedges et al, 2004; Pisani et al., 2004; Blair and Hedges, 2005a). Also, the young times found by Peterson and Butterfield (2005) are likely the result—in large part—of their decision to use molecular differences for timing without any statistical correction for hidden substitutions (multiple hits).

A related methodological issue involves the estimation of evolutionary rates among lineages. Two recent studies claimed that evolutionary rates were higher during the time of the Cambrian Explosion, which when accounted for in rate models allowed for younger divergence times to be recovered (Aris-Brosou and Yang, 2002, 2003). However, simulations (Ho et al, 2005) have suggested that this higher rate was an artefact of the particular method used in those two studies. Other problems associated with rate modelling in those studies have been discussed elsewhere (Blair and Hedges, 2005a).

Finally, although we have noted (above) possible explanations for why these recent molecular studies (Aris-Brosou and Yang, 2002, 2003; Douzery et al, 2004; Peterson et al., 2004; Peterson and Butterfield, 2005) have erred in their analyses, we wish to draw attention to a simple criticism that applies generally. They fail a basic test of consistency because they yield time estimates that are contradicted by the fossil record. For example, the Douzery et al. (2004) study estimated that the origin of various groups of algae (e.g., red, green) was hundreds of millions of years after their first fossil occurrences (see Section 3.2). When additional taxa were added to the data set of Aris-Brosou and Yang (2003), the divergence of animals and plants was found to be 671 Ma, nearly a half-billion years younger than the fossil constraint for that divergence (1200 Ma). In the other studies (Peterson et al., 2004; Peterson and Butterfield, 2005), the relevant data were not assembled by those authors to conduct such a consistency test. Considering this, the relatively small size of that data set, and especially the lack of statistical corrections for multiple substitutions, these results likewise are placed in question.

The consistency test demonstrates that those studies are biased to produce young dates and therefore those time estimates are unreliable. Thus if any of these young time estimates for animal evolution are to be seriously considered, it is incumbent upon those authors to explain why their results are not consistent with other aspects of the fossil record. Not considering these aberrant results, molecular clocks continue to support a long history of animal evolution in the Proterozoic (Fig. 3).


4.1 Complexity

It is logical to assume and expect that life begins in a simple state of organization and, through natural selection, develops greater complexity. For several reasons it is of interest to astrobiologists to know if there is any general and predictable pattern to this rise in complexity, because it would bear on our expectations of the existence of complex life (e.g., animal life) elsewhere in the Universe. For example, if the rise in complexity occurs quickly and easily, the probability that complex life occurs elsewhere is much higher than if it takes billions of years to develop complex life. Ward and Brownlee (2000), using this logic (in part), concluded that complex life is rare in the universe even though simple (prokaryote-like) life may be common.

The conclusion that complex life takes a long time to develop was based on a literal reading of the fossil record, which shows that most animal phyla first appeared in the earliest Phanerozoic, the Cambrian Explosion, nearly four billion years after the Earth was formed. However, most molecular phylogenies and timescales in the last four decades have indicated a deeper (Proterozoic) origin for the major groups of animals, as discussed above (Section 3.5). Also, an earlier origin of plants, fungi, and the major lineages of protists has been estimated (Fig. 3). But how does this new information bear on the rise in complexity? Biological complexity can be defined in many ways, including shape, size, number of cells, and number of genes, among many possibilities. However, the most common measure used to compare complexity across all of life is the number of cell types (Bonner, 1988; Bell and Mooers, 1997; McShea, 2001). Using this measure, and by estimating the number of cell types of common ancestors in the timetree of life (Figs. 1-3), it is possible to construct a contour for the rise in complexity of life on Earth (Hedges et al., 2004) (Fig. 4). This shows that complexity began to rise much earlier in time, roughly 2-2.5 billion years after the Earth was formed. The animal grade of complexity then rose more rapidly (>10 cell types) between 1000-1500 Ma, not 500-600 Ma as predicted by a literal reading of the fossil record.

Figure 4. Increase in the maximum number of cell types throughout the history of life (after Hedges et al., 2004). Data points are from living taxa (time zero) and common ancestors (earlier points) estimated with squared-change parsimony (solid circles) and linear parsimony (hollow circles) using a molecular timetree (Hedges et al., 2004). The dashed line shows an alternative interpretation based on uncertainty as to the level of complexity of ancestors of early branching eukaryotes.

Figure 4. Increase in the maximum number of cell types throughout the history of life (after Hedges et al., 2004). Data points are from living taxa (time zero) and common ancestors (earlier points) estimated with squared-change parsimony (solid circles) and linear parsimony (hollow circles) using a molecular timetree (Hedges et al., 2004). The dashed line shows an alternative interpretation based on uncertainty as to the level of complexity of ancestors of early branching eukaryotes.

Perhaps a key factor in this rise of complexity was the major rise in oxygen in the early Proterozoic (~2300 Ma) (Holland, 2002). The mitochondrion appeared soon thereafter (Hedges et al., 2001; Hedges et al, 2004), allowing eukaryotes to gain much more energy in cellular respiration compared with glycolysis. This symbiotic event (Great Respiration Event) is estimated to have occurred 2300-1800 Ma (Fig. 4).

That additional energy source may have provided the fuel for the rise in complexity, feeding the associated energy requirements (e.g., cell-signalling, mobility, etc.). The addition of the plastid at 1500-1600 Ma, through a symbiotic event with a cyanobacterium, then gave eukaryotes the ability to produce oxygen. This was the beginning of eukaryotic algae and almost certainly led to an increase in eukaryotic diversity and biomass. The parallel diversification of animals and fungi after 1500 Ma are likely related to this Great Algification Event.

Returning to the original question, these new insights from molecular clocks increase the probability that complex life exists elsewhere in the universe because the time required for complexity—in our single example on Earth—is less than previously thought. If the rise in complexity is tied to an energy source such as oxygen, as suggested, then a further consideration must be the time required to evolve the biological machinery for producing that energy (oxygenic photosynthesis or some other process).

Was this article helpful?

0 0

Post a comment