Teaching computers about comparing anatomy across species, part 1
Recently at PZ’s place, there was discussion of animal modeling, and in the comments, it emerged that we still have a lot of communication to do about why it’s necessary for understanding disease and developing treatments. I don’t know any researchers who enjoy having to use animals for research; I think most or all would prefer that it be unnecessary. And it is true that some egregious violations have occurred; it’s for that reason that protocols and oversight boards have been set up to oversee the research process. They are intended to ensure that no unnecessary suffering is inflicted on research animals, and apply to all researchers at accredited institutions.
Still, in spite of these measures, there is a lot of misinformation about the necessity for animal research. Human volunteers and computer models, while they can be appropriate in a small number of cases, cannot yet take the place of vertebrate and invertebrate animals in most research, and there are many cases where they never will be able to. In some cases, through refining the computer models, the use of animal models can be made more efficient or possibly can be minimized. As an animal lover, and a biological realist who recognizes the necessity for animal models in research, it is my hope that my comparative anatomy information system can ultimately help make the use of animal models more streamlined, both by providing access to previous results (to discourage unnecessary duplication), and by highlighting gaps in our knowledge (to indicate more productive possible hypotheses). If either or both of these goals were able to be even partially realized, it would mean that the necessity for using animals in research was channeled in a way that cut down on dead ends and inefficiencies, and,by optimizing it, made the necessary animal research count for more.
That is quite an ambitious goal I have set in the paragraph above, and so it is now my task to connect the dots for you–to explain how such a process could take place. Toward that end, this is the first of a series of posts on the comparative anatomy information system developed for my dissertation research–what I did, what its meaning is, and the vast amount of work remaining to be done.
To start with, let’s make sure that we’re all on the same page about terminology. First of all, a comparative anatomy information system is a computer system that allows users to compare canonical phenotypes of corresponding (i.e., homologous) anatomical structures across medically-relevant species at varying levels of detail, and that returns responses to queries about those comparisons.
By canonical, which is an adjective that can apply to “anatomy” and “phenotype”, as well as to any named organism or anatomical structure, I am taking my definition straight from Rosse’s definition . By it, I mean, as does Rosse, “a synthesis of generalizations based on qualitative observations, and sanctioned implicitly by accepted usage among domain experts”. Rosse contrasts canonical anatomy with “instantiated” anatomy, which is your anatomy, my anatomy, etc. In contrast to your particular instantiated kidney, or my particular instantiated lung, or whatever, the canonical kidney or lung is the generalized standard which is taught in anatomy education. Similarly, my cat or your iguana is an instantiated animal, but when we talk about the canonical vertebrate, we are talking about a generalization that biologists agree upon (for the most part, if not in every single detail). By animal model, I am referring to any animal which is studied for medical purposes as a surrogate for another species, usually (but not always) human, and it’s a subset of biological model. For example, the metastasis of prostate cancer is studied in the rat model.
Phenotype means the outward attributes of an organism–what do we see when we look at it (at any level)? So we can talk about normal observable variation between organisms as phenotypes, as well as outward manifestation of diseases, such as “the prostate cancer phenotype”. Here’s an example of the latter usage:
“Our findings suggest several regions that may contain genes which, when mutated, predispose men to develop a more aggressive prostate cancer phenotype. This provides a basis for attempts to identify these genes, with potential clinical utility for men with aggressive prostate cancer and their relatives.”
We’ll delve into more detail about corresponding; and homology; in later posts; for now, we’ll just define them as having a shared evolutionary relationship. For example, the mouse heart and the human heart are homologous organs, and so we consider them to correspond in the comparative anatomy information system (CAIS) we’ll talk about.
Now that we’re talking about the same things in the same terms, we’ll continue to explore these issues in the next post, Part 2.
 Rosse C, Mejino JL, Modayur BR, Jakobovits R, Hinshaw KP, Brinkley JF. Motivation and organizational principles for anatomical knowledge representation: the Digital Anatomist symbolic knowledge base. J Am Med Inform Assoc. 1998 Jan-Feb;5(1):17-40
 Schaid DJ; Investigators of the International Consortium for Prostate Cancer Genetics. Pooled genome linkage scan of aggressive prostate cancer: results from the International Consortium for Prostate Cancer Genetics. Hum Genet. 2006 Aug 25; [Epub ahead of print].