how to compare FA and DA between populations

31 views
Skip to first unread message

Audrey Friestad

unread,
Mar 31, 2025, 2:55:55 PMMar 31
to geomorph R package
Hi all!

I am working with a set of lizard cephalic scale shape data from three populations. I ran bilat.symmetry() for all three populations separately and found significant "side" effects in some and "individual*side" in others. I was wondering if there is a way I can test if the DA and FA levels are more than what is to be expected by chance, and if they differ between populations.

If someone could pass along some sample code, that would be much appreciated!

This is what my code looks like for each population:
bilat.sym.pharm <- bilat.symmetry(Pharm.all.gpa, land.pairs = pairs,
                                ind = muralis188pharm$indiv, replicate = pharmrep,
                                object.sym = TRUE, RRPP = TRUE)


Thank you very much!
-Audrey Friestad

Adams, Dean [EEOB]

unread,
Mar 31, 2025, 3:40:45 PMMar 31
to geomorph R package
Audrey,

To answer your first question (is the DA or FA more than expected by chance). Yes.  In the ANOVA table from bilat.symmetry, the 'side' component is a test of DA, and the 'ind:side' is a test of FA.  Each is tested relative to expectation using an appropriate permutation test.  Note however that to statistically evaluate FA (ind:side) you must have replication. In other words, each observation needs to be digitized at least twice. Otherwise, there is no variation against which the FA can be tested.

Your second question: can one determine whether these patterns differ across populations? Yes one can do this, but the procedure is a bit more complex to perform, as there is not a standardized function for it presently. Basically, what you are asking for is a comparison of, say the DA in population 1 vs. DA in population 2.  What you have in each is an effect size (Z-score) for the DA term in the respective ANOVA table. That Z-score is a permutation-based Z-score. We have developed statistical procedures for comparing effect sizes for test statistics in other contexts. Examples include comparisons of modularity or integration across datasets (compare.pls and compare.modularity functions). We had papers describing the math underlying these procedures in papers in 2016 and 2019 (see help files of those functions or my web page).

It turns out that the statistical procedure is general, and not restricted to integration and modularity only. Thus in theory, one may follow that algebra to develop the two-sample test to compare Z-scores for the DA component.  

Dean
--
Dr. Dean C. Adams
Distinguished Professor
Department of Ecology, Evolution, and Organismal Biology
Iowa State University

From: geomorph-...@googlegroups.com <geomorph-...@googlegroups.com> on behalf of Audrey Friestad <acfri...@gmail.com>
Sent: Monday, March 31, 2025 1:55 PM
To: geomorph R package <geomorph-...@googlegroups.com>
Subject: [geomorph-r-package] how to compare FA and DA between populations
 
--
You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/geomorph-r-package/d4ec612b-c9dd-4c7d-adb4-09df3acd744fn%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages