interpretation of phylogenetic signal Kmult statistical test

[Juan Esteban Vrdoljak]

Hi all.

I am interested in Kmult statistical test but I have some doubts. I understand the interpretation of Kmult = 1, >1, or <1, but I don’t understand the meaning of the statistical test. For example, I performed this test in my dataset and obtained Kmult = 0.063; the null distribution ranged between 0.03 and 0.13. Suppose K is significant, is there some difference in the interpretation of K if the observed K is to the right or to the left of the null distribution?

On the other hand, suppose Kmult close to 1 (for example, 0.95). In this case, is there some different interpretation if this observed K is to the right of the null distribution (for example, a distribution ranged between 0.5-0.8) or to the left (null distribution ranged 0.99-1.5).

Thank you very much!

Best,

Juan

[Adams, Dean [EEOB]]

Juan,

 

The test is relative to a random permutation of tips data with respect to the phylogeny. It is a one-tailed test, comparing the observed to values greater or equal to it (i.e., how ‘large’ is the observed relative to the distribution). Following Blomberg et al. 2003, a significant result indicates that there is greater phylogenetic signal than expected.

 

For a more detailed explanation, see: Adams 2014, and parts of Adams & Collyer 2019 (the section on phylogenetic signal).

 

Dean

 

Dr. Dean C. Adams (he/him)

Distinguished Professor of Evolutionary Biology

Director of Graduate Education, EEB Program

Department of Ecology, Evolution, and Organismal Biology

Iowa State University

https://faculty.sites.iastate.edu/dcadams/

phone: 515-294-3834

--
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 on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/12f9b5ea-5f0b-4a5e-a80f-ded2d273f63dn%40googlegroups.com.

[Juan Esteban Vrdoljak]

Hi Dean, thanks a lot for your quick response!

Ok, I understood that. However, in the biological example, Adams 2014 found K=0.957 and p=0.014 and suggested that "closely related species were more similar to one another in head shape than was expected under a Brownian motion model of evolution." Why is this sentence correct? I understand that Kmult <1 implies that taxa resemble each other phenotypically less than expected under Brownian motion.

Thanks again!

Juan

[Adams, Dean [EEOB]]

K = 1.0 is not a hard-and fast cut-off. The value, combined with the significance test, helps to make the determination. See also the comments of Blomberg et al. 2003 in the univariate proposal of K and how to interpret tests based on it.

 

This is a challenging issue to be sure. We have more to say about phylogenetic signal and comparison tests of it, in an upcoming paper in MEE (Collyer et al.).

To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/796c4a93-94d7-4c11-8d85-13e05480009an%40googlegroups.com.

[Juan Esteban Vrdoljak]

Thanks again for the quick response! I'll be waiting for that paper!

cheers!