PhiST with AMOVA

[Kristopher Wojtusik]

Hello All,

We used poppr to do an amova for our study and a reviewer has called out the use phiST as a measure of differentiation and says it is designed to use %basepair differences amond DNA sequences, which obviously SNP data do not have. 

We know that poppr (or pegas since that's the method we used), adjusts the data to Euclidian distance and then runs the AMOVA on that, but we were wondering if anyone had resources about this, or a more thorough explanation of this. Thank you.

-Kris

[Zhian Kamvar]

Meirmans and Liu 2018 gives a pretty good summation of the issue at hand in the section titled "The AMOVA Framework": https://www.frontiersin.org/articles/10.3389/fevo.2018.00066/full

The details of how poppr implements AMOVA is found in https://grunwaldlab.github.io/poppr/reference/poppr.amova.html#details

hope that helps.

Zhian

--
You received this message because you are subscribed to the Google Groups "poppr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to poppr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/poppr/9d0628e6-31b4-43e5-854c-4e7a081a489dn%40googlegroups.com.

[Kristopher Wojtusik]

Thank you for the resources Zhian! They are helpful and at the very least will give us something to cite in our rebuttal.

-Kris 

[Gabriella Scatà]

Hi,

I ran poppr::amova with correction = NULL (method = pegas), because I provided my own geentic distance matrix based on Nei's genetic distances, and so I did not want it to be corrected.

I read that if you use the ade4 implementation of amova, you need to provide an euclidean distance, so you need to correct you distance matrix, but if you use the pegas implementation, would any genetic distance be ok?

Thank you!

Best

Gabriella