It is now well known that the explosion of genome data in recent decades has made its own unique contribution to the ever-growing list of falsified evolutionary predictions. High conservation of functionally unconstrained sequences, nonsensical evolutionary trees, molecular clocks that do not run right, and phylogenies that do not resolve are all contributing to a reevaluation of tree thinking. From superoxide dismutase to glycerol-3-phosphate dehydrogenase, incongruities are common. And novel designs in similar species, once assumed to have arisen only once and then passed on via common descent, now must be assumed to have evolved multiple times. So it is hardly surprising that a substantial molecular study, recently reported in Science magazine [310:1933], revealed yet more curious results.
The study, encompassing 50 genes in 17 different species, was designed to resolve metazoan evolutionary relationships. But despite the plethora of data, clear results were not produced. Both maximum likelihood and maximum parsimony methods failed to resolve a statistically significant picture of early evolutionary branching, though of course the broad outlines of relationships between very different groups was evident.
So what went wrong? The researchers explored several possible reasons, including the choice of taxa, particular data issues, and mutational saturation. But none of these seemed to pose a problem. So the study was left with only one conclusion: there must have been "a radiation compressed in time." In other words, the new species appeared in rapid fire sequence.
The problem is that even though a substantial quantity of molecular data was collected, there nonetheless is a small level of uncertainty. What if the species split off from each other close in time? If these cladogenetic events were tightly spaced, then the reconstruction would be left uncertain as to specific ordering of events.
The remedy is to use yet more data, but the authors note that this may not be feasible. While species have many thousands of genes, these evolutionary studies are limited to orthologs--genes that can be identified as having an evolutionary relationship across the taxa (yes, there is a bit of circularity in all this). That is not easy, especially when studying a large number of taxa, and unfortunately there may not be many more orthologs available. So this particular story ends with a rain check that, conveniently, may never be cashed in.
