Hello, Anole Annals readers,
I work for the Howard Hughes Medical Institute’s Science Education Department. To support the anole film that Jonathan Losos talked about on this post, we developed a classroom activity to explore the morphology and the phylogeny of Caribbean anoles using photographs and DNA sequences. The resources are available from our BioInteractive.org website.
Students are given color photographs to sort different anole species into ecomorphs. Having grouped the species, they use an online phylogeny tool to build a phylogenetic tree from the DNA sequences from the same species. The results show that different species from the same islands cluster together, independent of the ecomorphs, suggesting that the ecomorphs are examples of convergent evolution. I’m sure Anole Annals readers are well-versed in this, but we tried to make the research accessible to high school teachers and students. Teachers are always looking for evolution resources that use actual DNA sequences.
Anole trading cards, used in the classroom exercises.
- Anole Phylogeny Activity for High School/College Level - May 22, 2014
Kirk Fitzhugh
Mapping characters on to a phylogenetic hypothesis inferred from other characters is not a scientifically correct way to discern character evolution, much less conclude instances of homoplasy (convergence).
Luke Mahler
Sure it is, provided that you have reason to believe that the former character set is a reliable indicator of phylogenetic relationships. This is pretty standard fare in the scientific literature. The inferred pattern of evolution of the second character set is regarded as conditional on the phylogenetic hypothesis, and I don’t think any scientist would claim otherwise.
In Anolis, the major features of the “DNA phylogeny” are corroborated by many independent lines of evidence including different types of DNA (many independent nuclear loci, as well as mtDNA), protein electrophoresis, and to some extent karyological and allozyme data. This is not the case for morphometric measurements of traits important for habitat use, such as relative limb and tail lengths or the number of subdigital lamellae. Such traits tend to be discordant with the range of phylogenetic patterns inferred from DNA data, except when close relatives (as inferred by DNA) happen to also be ecologically similar to one another. There are many reasons to think that nucleotides are reliable phylogenetic markers in Anolis (e.g., much variation is plausibly neutral; many independent loci corroborate one another), and there’s nothing scientifically incorrect about using such DNA-based phylogenetic hypotheses as the basis for (conditional) models of evolution for other characters.
Kirk Fitzhugh
A ‘phylogeny’ is a set of hypotheses specifically inferred to causally account for those characters used to infer that ‘phylogeny.’ As such, whatever additional characters one might have, they are irrelevant to the hypotheses originally inferred since that ‘phylogeny’ doesn’t explain the second set of characters; it only explains the characters for which it was inferred. Nothing whatsoever can be said about other characters relative to those hypotheses. Now, if one really does want to offer evolutionary explanations for additional characters, then what is needed is a new inference in which both sets of data are used as premises to infer a new set of hypotheses. This is a basic requirement of rationality when it comes to non-deductive reasoning. The problem that currently exists is that speaking of a ‘phylogeny’ has become a matter of reification.
A DNA phylogeny [sic] cannot be corroborated [sic] by other character data. No proper hypothesis testing has occurred to which corroboration can refer. This notion of testing is one that developed within systematics, with no real connection to the actual mechanics of testing as applied in other fields of science. Re reliability, that is only meaningful in the case one can point to successful tests of hypotheses. The problem, however, is that phylogenetic hypotheses aren’t tested – certainly not in the accepted manner of seeking predicted consequences, test evidence, from the hypotheses that specifically address the causal claims implied by a ‘phylogeny.’
Luke Mahler
Hi Kirk,
I recognize that not every systematist agrees to a single definition, but I don’t think the definition you offered describes what most workers in the field would consider a phylogeny.
What I mean by “phylogeny” is an estimate of the history of evolutionary divergence (the order and timing of branching events) for a group of species. I think this is what most others mean (systematists as well as members of the public), as evidenced by a quick look at textbooks in systematics, evolutionary biology, and general biology, as well as various online dictionary definitions.
A phylogenetic estimate can be generated for data under a given model of character evolution, and support for alternative estimates can be compared statistically using a variety of techniques. This approach of model-based inference is robust and has yielded countless tremendously useful and interesting insights in systematics, including many that have been validated experimentally (e.g., through direct observation of phylogenetic diversification in rapidly evolving organisms such as viruses) or that have received additional support from other sources of information about the history of life (e.g., fossil discoveries). Model-based inference of historical process is also not by any means unique to systematics, and does not preclude additional hypothesis testing. Consider the recent discovery of gravity waves, which were previously unknown, but were predicted by the “big bang” model of inflation in the universe. The predictions of phylogenetic models can always be tested with new data. Phylogenetic models make all sorts of predictions, and these are routinely tested through discovery and analysis of new data, or application of novel analytical techniques.
Martha Muñoz
Dear Satoshi,
This is a great classroom exercise! I hope to use this in the future. Thanks for the work you’ve done and for sharing it with us.
Best,
Martha