My Research: Morphological Evolution of Cypriot Arthropods and its Predictability

An animal’s morphology is determined by many factors, from genetic regulation to cellular migration and tissue growth to the environmental conditions where the embryo is growing, and all these levels are chained together through feedback loops. These alone can explain the differences between snakes and birds and whales and humans and cockroaches and sponges and jellyfish.

My interest here is at a much smaller scale, explaining the difference in morphology between sister species rather than entire animal clades. As an example, take this 2004 study on Darwin finches by Abzhanov et al.. They found that the well-known variation in Darwin finch beaks is down to changes in the expression of one protein, Bmp4, in the developing beak. This leads to the differently-shaped beaks and underlies the finches’ ecological specialisations that helped inspire Darwin to his theory of natural selection.

In this project, I will be finding out what expression changes control the development of the diagnostic characters for the endemic arthropod species of Cyprus.

This information will then be incorporated into a spatial model, to check whether a species’s ecological range overlaps significantly with the ophiolitic areas of Cyprus. This is important because my core hypothesis is that the geology of Cyprus is a significant driver of endemism; see here.

If this core hypothesis turns out to have some truth to it, then that will open up a new research stage, developing a computational model that will predict the morphological possibilities.

These three steps involve different skillsets, and each is a giant project on its own.

Step 1 requires a complete taxonomic database of Cypriot arthropods, to allow the identification of the endemics and to choose some good model organisms. I am currently concentrating on pholcid spiders and buprestid beetles as models, although there are many more candidates (I have my eye set on wasps). I only chose pholcids because they are evolutionary masterminds and buprestids because almost half of Cypriot buprestids are endemic to the island.

The endemic species are investigated from a taxonomic perspective to identify the truly diagnostic characteristics. This also entails studying species from the entire (sub)genus that occur in the general area (Middle East, North Africa, Anatolian Turkey), to identify the species that the Cypriot endemic is most closely related to, possibly even descended from. I do this from an old-school morphological perspective because I’m not satisfied with molecular phylogenetics just yet.

Once all this is done, the truly difficult (and costly) work begins: identifying the bases of the differences between the endemic and the sister species. Nurture and culture the beasts, and find out which proteins are expressed in the relevant areas, what controls setal count and length, what determines the number of antennomeres, and what happens when you mess around with that control (knock genes out, silence or supercharge protein expressions, whatever else your sick mind can think of). Basically all the crucial labwork.

By the end of that, I will have a very precise idea of what led to the morphological changes in the endemic species. With the addition of geographical and ecological data, I can trace whether any of these changes could be correlated to serpentine soils and heavy metals, which could pave the way for computational models for predictive morphological evolution.

This is still an idea I’m developing in my head, but a recent Science paper by Sheth et al. (2012) convinced me it’s a possibility, if there is a pattern to the morphological variation in response to expression changes. In an ideal scenario, I could make a model that computes regular within-species variation, which will make taxonomic work easier when trying to determine the status of an unknown specimen that could potentially be a new species. Preliminary tests with a small Ruby program and fake data yielded good results, for whatever that’s worth (nothing at all). Less importantly, the model could be used to predict all possible phenotypes, but I don’t see any use for that (video games?).

Leave a Reply