Full text
| |
| Author(s): |
Guilherme Garcia
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Instituto de Biociências (IBIOC/SB) |
| Defense date: | 2016-04-07 |
| Examining board members: |
Gabriel Henrique Marroig Zambonato;
Paulo Roberto Guimaraes Junior;
Tábita Hünemeier;
Blanche Christine Pires de Bitner Mathe Leal;
João Alves de Oliveira
|
| Advisor: | Gabriel Henrique Marroig Zambonato |
| Abstract | |
Modularity is a characteristic property biological systems exhibit regarding the distribution of interactions between their composing elements; in this context, a module is a subset of elements which interact more among themselves than with other subsets. Regarding morphological systems, such property usually refers to the structure of the linear component of the genotype/phenotype map; however, the genetic, developmental, and functional interactions that produce phenotypes are often best described by non-linear dynamics, and a full appreciation of the complexity of such interactions is necessary for understanding phenotypic variational properties. Furthermore, given methodological advances in the field of morphometrics, one may choose different ways to represent morphological variation, and differences between representations may impact inferences made regarding variational properties. The present dissertation aims at exploring the relationship between morphometric representations and the characterization of variational properties, focusing on the comparative analysis of such properties on a macroevolutionary timeframe; Anthropoid Primates are used as a model lineage, given the availability of a large database of skull measurements. This relationship was evaluated under three different perspectives. First, an estimation of the error rates associated with tests for hypothesis that describe modularity patterns related to three different morphometric representations; such evaluation is also associated with an exploration of a subset of the database used here, considering the dynamical properties of developmental interactions that produce the Anthropoid skull. The results of this chapter imply that one of such representations, Procrustes residuals, fails to capture modularity patterns in this setting, considering its particular mathematical underpinnings. Other two representations, interlandmark distances and local shape variables, produce similar results which are directly associated with developmental dynamics, and the differences they exhibit are consistent with their different properties; error rates for tests over both representations are also acceptable. The next chapter deals with comparing these two representations with respect to these different properties, focusing on estimating allometric relationships over local shape variables and the relationship between such estimates and modularity patterns estimated for interlandmark distances. The results found stress out that modularity patterns observed in interlandmark distances are a consequence of allometry; lineages such as Homo and Gorilla, which exhibit distinct modularity patterns in interlandmark distances are associated with substantial changes in allometric relationships for skull traits. The last chapter explores the phylogenetic structure of changes in phenotypic variational properties across Anthropoid diversification, considering local shape variables alone, since this chapter also aims at reinforcing previous results obtained from interlandmark distances, considering a different type of morphometric representation. This chapter shifts the focus from testing a priori-defined modularity patterns to estimating the uncertainty related to covariance matrix structure decomposed over the Anthropoid phylogeny. The results obtained demonstrate that changes in covariance structure on this lineage are localized in the same skull regions across the entire evolutionary history of Anthropoidea, while other regions maintain stable associations. Thus, when one considers the different properties of morphometric representations carefully, inferences made from such representations regarding variational properties are in fact compatible (AU) | |