Geometric Morphometric Methods for Functional Morphological Studies

Suny At Stony Brook, Stony Brook NY

Investigators

Abstract

This project has four objectives related to the application of geometric morphometric methods. These methods are concerned with new mathematical and statistical methods for analyzing variation in the shapes of biological structures and their covariation with other variables (such as size, environmental variables, etc.). This approach allows one to treat the shape of an entire configuration of landmarks and/or a set of outlines as a single mathematical object that can be analyzed statistically. The specific objectives are as follows: (1) A study of the comparative method (statistical methods that take into account an expected lack of independence of observations due to phylogenic relatedness). The generalized least-squares technique will be used to extend several types of standard multivariate statistical analyses (e.g., two-block partial least-squares and principal components analysis) and the application to reticulate evolution. The phylogenetic autocorrelation method will also be studied. (2) Because morphometric data is sometimes incomplete, methods for the imputation of missing coordinate data will be investigated. These will include mean-value substitution, regression estimation, spline relaxation, and hot-decking. (3) Work will start on building a database on the morphometric diversity of Near Eastern human populations. These data will be useful for the two objectives described above. (4) New methods (described above) that we find most useful will be implemented in the tps series and the morpheus software so that these techniques will be freely available to others. Software for phylogenetic autocorrelation analysis will also be developed.