“This is very beautiful. It is neat, it is modern technology, and it is fast. I am just wondering very seriously about the biological validity of what we are doing with this machine.” – Melvin Moss, 1967*
“This machine” to which Moss referred nearly 50 years ago was not a contraption to clone a Neandertal or a Godzilla-like MechaGodzilla, but a computer. Along these lines, a paper came out recently describing a new, automated method for analyzing (biological) shapes, and while I think the method is pretty sweet, I think future researchers employing it should keep Moss’s monition in mind.
Doug Boyer and colleagues (2011) present “Algorithms to automatically quantify the geometric similarity of anatomical surfaces.” It seems the main goals of the study were to make shape analysis [1] faster and [2] easier for people who don’t otherwise study anatomy (such as geneticists), making it possible [3] to amass large phenotypic datasets comparable to the troves of genetic data accumulated in recent years. Using some intense math that’s way over my head, the computer algorithm takes surface data (acquired through CT or laser scans) of a pair of specimens and automatically fits these forms with a “correspondence map” linking geometrically (and not necessarily biologically) homologous features between the two. It then uses the map to fit landmarks (a la geometric morphometrics) which are used to calculate the shape difference metric between individuals in the pairings.
See at the right just how pretty it is! The authors posit that this technique could be used with genetic knock-out studies to assess how certain genes affect the development of bones and teeth, or to model the development of organs. That certainly would be useful in biomedical and evo-devo research.
But while I appreciate the automated-ness of the procedure, I don’t think we can simply write off the role of the biologist in determining what features are homologous, in favor of a computer. The paper itself illustrates this nicely. The authors state that there is debate about the origins of a cusp on the molar tooth of the sportive lemur (Lepilemur) – is it the same as the entoconid of the living mouse lemur, or the enlarged metaconid of the extinct “koala lemur”? Their automated algorithm can map the sportive lemur’s mystery cusp to match either alternative scenario. It is the external paleontological and phylogenetic evidence, not the intrinsic shape information, that renders the alternative scenario more plausible.
So let me reiterate that I think this paper presents an important step for the study of the biology of form, or the form of biology. Automating the analysis of form will certainly expedite studies of large datasets (not to mention freeing up the time of hordes of research assistants). But I hope that researchers employing this procedure will have a little Mossian Angel (poor play on “guardian angel,” sorry) on their shoulders, reminding them that the algorithm won’t necessarily show them homology better than their own experience. And I hope all biologists have this Mossian Angel there, reminding them that even though this method is “neat … modern technology, and … fast,” it may not be the most appropriate method for their research question.
References
Boyer, D., Lipman, Y., St. Clair, E., Puente, J., Patel, B., Funkhouser, T., Jernvall, J., & Daubechies, I. (2011). Algorithms to automatically quantify the geometric similarity of anatomical surfaces Proceedings of the National Academy of Sciences, 108 (45), 18221-18226 DOI: 10.1073/pnas.1112822108
*This quote comes from a discussion at the end of a symposium: Cranio-Facial Growth in Man (1967). RE Moyers and WM Krogman, editors. New York: Pergamon Press.
Lawn Chair Anthropology 
"It is the external paleontological and phylogenetic evidence, not the intrinsic shape information, that renders the alternative scenario more plausible"And here is the reason why I think that most studies using geometric morphometrics are mostly a waste of time.I can see the reason why geometric morphometrics are useful in regards of studying ontogenetic or functional aspects, but besides from that you don't get much more information then with "oldschool" methods. More then that I think in some cases it's even a step backwards, because people start to concentrate too much on descriptive and phenomenological studies instead of stuff which generates testable hypothesises.On the other hand: It looks like this is where the money is so maybe I should forget about my concernes and join the party.P.S.: Keep in mind that I only have basic knowledge about geometric morphometrics so my scepticisms might be simply because I don't understand the method well enough.
I wouldn't go so far as to say most GM studies have been wastes of time, but I do agree that there are many papers out there that seemed to do GM just simply to do GM. GM is certainly a powerful tool for examining shapes, but you're right that often people simply ran the analyses, looked at plots of PC1 vs 2, PC2 vs 3, etc and made inferences from there, or argued that a plot supported a given argument.I really like Paul O'Higgins' GM work, though, as he and his colleagues always set up clear hypotheses, and then actually test them with GM and multivariate tests. Good science!Anyway, though, I am interested to see this automated procedure of Boyer and colleagues implemented. I am especially intrigued at their suggestion that this technique could be used to help understand how genes (in KO experiments) affect embryonic development.That said, one of the goals of the new method was to amass phenotypic data on the scale of genetic data. But it should be borne in mind that genetic and shape data are vastly different qualitatively: genes are sequences of As, Cs, Gs and Ts. But these biological shape data cannot simply be quantified with a number, but the quantification of form is always in reference to another form, be it a group mean or another individual in a pair-wise comparison. So I don't know how useful "phenomic" datasets could be outside their original contexts.
I had the pleasure to hear Paul O' Higgins in September while I was in Leipzig and I completely agree with you. He knows about the strengths and weaknesses of GM, but unfortunately most other people don't.Judging from the talks I heard in Leipzig most people use GM as a simple substitute for classic morphometrics and so far I have yet to realise why I should need GM to determine whether or not Skull A looks different from Skull B.And judging from a cladistic perspective I think using GM is really harmful in terms of classification, because it's purely phenetic.But I think the reason why GM is often used the wrong way, is because of a general Problem of physical anthropology which always has been a little bit obsessed with data acquisition and descritption rather then being driven by hypothesis testing.Right now all I can see from this new method is people generating huge amounts of mostly useless data. Or maybe I'm just a little bit too cynical right here.P.S.: I agree with you that there might be an area where this method is really helpful.
I should add to my original comment, that pairwise comparisons are actually very important in genetics – such comparisons were crucial in the recent studies showing that different people have different amounts of Neandertal and "Denisovan" ancestry. I'm also relying heavily on them in my dissertation. That is all.