:::: MENU ::::
Browsing posts in: Phylogenetics

New paper and Shiny app

New paper on systematics of Cyclotella and resurrection of a long forgotten name Lindavia published in Phytotaxa. Thanks to Wilson, Matt, Ed and Andy for their contributions!


New R Shiny app to monitor and present data from experiments on diatom growth at different salinities. Two experiments with eurihaline Cyclotella species included. See it here Diatoms and Salinity.

More on the interactive diatom phylogeny

Its been a while since I wrote anything. Its also been a while since I did much work on the diatom phylogeny browser. For all those eager to learn about the phylogenetic relationships among diatoms, the browser is your friend. You can select large clades (e.g. raphid or pennate diatoms), plot custom clades based on two terminal taxa and even plot a sub-tree based on a genus of interest. Browse away.

In the last couple of weeks I changed the phylogeny. Instead of the now very old tree from Theriot et al. (2010) the browser uses the more recent and much larger 282-taxon phylogeny from our analysis of the evolution of habitat and coloniality across diatoms.

In addition to subsetting the tree in various ways, one can now plot some important characters and their ancestral reconstructions. Specifically, I added three binary traits: Planktonic-Benthic, Solitary-Colonial and Marine-Freshwater. The former two are exactly as they appear in the paper, whereas for “Salinity” I coded the taxa based on where we collected them or the culture conditions they grow in. The reconstructions are performed separately for each selected sub-tree using the “all rates different” model in the R package corHMM with “marginal” reconstructions for each node. Due to this, reconstructions in the app will be slightly different from the original character histories shown in our paper. Note also, that some reconstructions in the paper were made with hidden-rates models not used in the app.

The final minor improvement of the browser is that the plot height is now scaled automatically based on the number of taxa in the selected sub-tree. This works much better than the previous version where sub-trees of e.g. 100 and 10 taxa would have been “stretched” to the same plotting height.

The tree, character data, and code are available from the app’s GitHub repository.