Dear Udita, To complete Krzysztof's message, there is an implementation of phylogenetic PCA for multiple models in the mvMORPH package (see the mvlgs.pca function). This model is simpler than some of the other multivariate versions you may find in the package, but it allows departure from the BM assumption used in most implementations.
All the best, Julien ________________________________ De : R-sig-phylo <[email protected]> de la part de Krzysztof Bartoszek via R-sig-phylo <[email protected]> Envoy� : mardi 16 d�cembre 2025 23:03 � : [email protected] <[email protected]> Objet : Re: [R-sig-phylo] Phylogenetic ordination for proportion data with zeroes (Udita Bansal) Dear Udita, While I do not have a complete answer to your question there are two things that could be useful. You mention that you hav emany structural zeroes. What do you mean ny this? That the trait is not present in the particular species? In this case the PCMFit (on GitHub) or glinvci (on CRAN) packages could be an option. They allow for estimation under a wide family of phylogenetic Gaussian models and permit missing traits directly without the need to add some tiny values (as this has a completely different mathematical meaning). You also mention that you want to do a PCA. While one can do a phylogenetic PCA under a Brownian motion model of evolution, if you change the model to, e.g., an Ornstein-Uhlenbeck model (which allows for stabilizing selection), then I do not think that there is any phylogenetic PCA implemented. Best wishes Krzysztof Bartoszek _______________________________________________ R-sig-phylo mailing list - [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/[email protected]/ [[alternative HTML version deleted]]
_______________________________________________ R-sig-phylo mailing list - [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/[email protected]/
