For analyzing shape differences between classes of observations, it is useful to compute the differences between the average shapes of those classes within multiple groups. For instance, a study of the evolution of sex dimorphism of shape can compute the difference between the average shapes of males and females within multiple species and then conduct further analyses of the resulting estimates of sex dimorphism (e.g. Gidaszewski et al. 2009).
MorphoJ creates a new dataset with the differences, which can be used for further analyses.
To use this type of analysis, the data must be organised in groups, for example species, and within each these groups, there are classes that are used to define the disfferences, for example the sexes. Differences can only be computed for a group if both classes are represented (e.g., to calculate sexual dimorphism in a species, both males and females are required). There is no need for the data to be perfectly balanced, but of course data where the numbers of specimens in the different classes and in all groups are about the same tend to yield the best estimates of differences.
For computing differences, MorphoJ requires a dataset that contains two separate classifier variables as the criteria for determining the groups and the classes (e.g. species and sex, respectively). It stores the resulting differences in a new dataset.
To start, select the dataset in the Project Tree window and then choose Differences of Averages by Groups... from the Preliminaries menu. A dialog box like the following will appear:
At the top, there is a text field for the name of the new dataset.
Below that is a box that lists all the classifiers in the original dataset that was selected in the Project Tree. From them, the user should select one or more classifiers that are to be used for defining the groups within which the differences are calculated. In the example, there are two classifiers, named "species" and "sex", and "species" has been selected as the criterion for groups. As a result, the differences will be calculated within species in this example.
The lower two boxes are for specifying how the differences are to
be computed within each group. In the left box, all the
classifiers except for those used in defining groups are listed.
From these classifiers, select the one that is to be used for
defining the classes between which the differences are to be
calculated. In the example, only the single classifier "sex" is
available and it has been selected here.
Once the classifier is selected, the values that it takes are
shown in the box to the right. If there are only two values, these
are automatically used as the classes between which the
differences then are computed. If there are more values, the user
can choose the appropriate ones. For instance, if there are males,
females and specimens for which sex has not been identified, it
makes sense to use males and females for computing sex dimorphism,
but to omit the unidentified specimens.
Click the Execute button to start, or the Cancel button to abort the procedure.
The result is a new dataset, in which there are differences between the selected classes in all groups. The differences are computed for all available data types, including shape variables, centroid size and covariates.
Gidaszewski, N. A., M. Baylac, and C. P. Klingenberg. 2009. Evolution of sexual dimorphism of wing shape in the Drosophila melanogaster subgroup. BMC Evolutionary Biology 9:110.