In cromanpa94/phruta: Phylogenetic Reconstruction and Time-dating

Table of Contents

1. Installing RAxML
2. Installing PATHd-8 and treePL
3. Phylogenetic inference with phruta and RAxML
4. Tree dating in phruta

Installing RAxML

In MacOS, RAxML can be easily installed to the PATH using one of the two lines below in conda:

conda install -c bioconda/label/cf201901 raxml 

conda install -c bioconda raxml

For other OS (Windows, Linux), please follow the instructions listed in the official RAxML website

Once RAxML has been installed to your computer, open R and make sure that the following line doesn't throw an error.

system("raxmlHPC")

Depending on how RAxML was installed, you may want to check if RAxML is called from the terminal using raxmlHPC or raxmlHPC. This string needs to be passed to tree.raxml() using the argument raxml_exec. Please note that this argument corresponds to the exec argument in ips::raxml().

Finally, note that RStudio sometimes has issues finding stuff in the path while using system(). If you're using macOS, try starting RStudio from the command line by running the following line:

open /Applications/RStudio.app

In other OS, it might be better to simply avoid using RStudio if you're interested in running the phylogenetic functions in phruta.

Installing PATHd-8 and treePL

There are excellent guides for installing PATHd-8 and treePL. Here, I summarize two potentially relevant options.

First, you can use Brian O'Meara's approach for installing PATHd-8 in MacOs and linux. I summarize the code in the following link. For Windows users, please use the compiled version of the software provided in the following link.

Second, you can use homebrew to install treePL (Windows, MacOS, and Linux), thanks to Jonathan Chang.

brew install brewsci/bio/treepl

Please check the following link) if you're interested in running brew from Windows and Linux.

Phylogenetic inference with phruta and RAxML

Phylogenetic inference is conducted using the tree.raxml() function. We need to indicate where the aligned sequences are located (folder argument), the patterns of the files in the same folder (FilePatterns argument; "Masked_" in our case). We'll run a total of 100 boostrap replicates and set the outgroup to "Manis_pentadactyla".

tree.raxml(folder='2.Alignments', 
           FilePatterns= 'Masked_', 
           raxml_exec='raxmlHPC', 
           Bootstrap=100, 
           outgroup ="Manis_pentadactyla")

The trees are saved in 3.Phylogeny. Likely, the bipartitions tree, "RAxML_bipartitions.phruta", is the most relevant. 3.Phylogeny also includes additional RAxML-related input and output files.

Finally, let's perform tree dating in our phylogeny using secondary calibrations extracted from Scholl and Wiens (2016). This study curated potentially the most comprenhensive and reliable set of trees to summarize the temporal dimension in evolution across the tree of life. In phruta, the trees from Scholl and Wiens (2016) were renamed to match taxonomic groups.

Tree dating in phruta

Tree dating is performed using the tree.dating() function in phruta. We have to provide the name of the folder containing the 1.Taxonomy.csv file created in sq.curate(). We also have to indicate the name of the folder containing the RAxML_bipartitions.phruta file. We will scale our phylogeny using treePL.

tree.dating(taxonomyFolder="1.CuratedSequences", 
            phylogenyFolder="3.Phylogeny", 
            scale='treePL')

Running this line will result in a new folder 4.Timetree, including the different time-calibrated phylogenies obained (if any) and associated secondary calibrations used in the analyses. We found only a few overlapping calibration points (family-level constraints):

Here's the resulting time-calibrated phylogeny. The whole process took \~20 minutes to complete on my computer (16 gb RAM, i5).

cromanpa94/phruta documentation built on July 7, 2024, 4:58 p.m.