This repository contains supplemental results, data, and scripts for Kramer et al. 2023
The subfolders are outlined below.
1_make_starting_tree: This folder contains scripts to produce a filtered global phylogeny, the "starting tree" in the manuscript. Samples from the starting tree are used in 3_real_data_experiments to infer trees from real data.
2_optimize_starting_tree: The starting tree from the previous folder is optimized with various methods, and the best tree is chosen as the "ground truth" over which to simulate sequences (in folder 4_simulated_data_experiments)
3_real_data_experiments: This folder contains scripts and results of tree inference methods on real SARS-CoV-2 data. Each method is compared by log likelihood.
4_simulated_data_experiments: Scripts and results of inference methods on simulated SARS-CoV-2 data. Comparisons between methods are done by computing tree distances to a ground truth phylogeny.
The scripts in this repository use the following programs:
- Python 3
- The UShER suite (UShER, matOptimize, matUtils)
- FastTree 2
- IQ-TREE 2
- RAxML-NG
- TreeCmp
Most of the above can be installed with Conda:
# Create and activate a new environment
conda create -n parsimony
conda activate parsimony
# Set up channels
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge
# Install packages (versions are those used in our experiments except where stated otherwise)
conda install usher=0.4.8
conda install raxml-ng=1.1.0
For IQ-TREE 2, download version 2.1.3 Linux multicore edition:
wget https://github.com/iqtree/iqtree2/releases/download/v2.1.3/iqtree-2.1.3-Linux.tar.gz
For FastTree 2, install the double-precision executable:
wget http://www.microbesonline.org/fasttree/FastTreeDbl
For TreeCmp, follow the instructions at https://github.com/TreeCmp/TreeCmp
For more details on the UShER suite, see the wiki.