https://doi.org/10.5061/dryad.bg79cnph0

Data, Code, and Tables for “CASTER: Direct species tree inference from whole-genome alignments.”

To benchmark the performance of CASTER against alternative methods, we simulate aligned sequences under Hudson model (species tree & recombination). Particularly, we simulate 200 ingroup species + 1 outgroup species (SR201 dataset) to benchmark the performance of CASTER under different mutation rates, population sizes, and ploidies (unphased). Next, we subsample 20 ingroup species + 1 outgroup species but increase the number of individuals per species to benchmark the performance of CASTER with up to 20 individuals. The scripts for those simulations are also provided (simulate_SR201_10X_population.py, simulate_SR201_most_conditions.py, simulate_SR21.py).

We also simulate a lot more alignments under SR201 condition for scalability studies (scale) which are not included here because:

• The total size exceeds the 300GB limit
• Those who are interested in this data set can easily simulate them using simulate_scale.py
• We are only interested in the running time, rather than the accuracy

We also apply CASTER to biological datasets. We upload a copy of the alignments used here, unless they are easily accessible (which we provide links).

Data

Alignments

Quartet dataset (FASTA format)

quartet.zip: all condition (100 replicates)

SR201 simulated dataset (FASTA format)

• Alignment_SR201_default_diploid_part[1-5].tar.bz2: default + diploid condition (50 replicates)
• Alignment_SR201_0.1X_mutation_part[1-2].tar.bz2: 0.1X mutation rate condition (50 replicates)
• Alignment_SR201_10X_mutation.tar.bz2: 10X mutation rate condition (20 replicates)
• Alignment_SR201_10X_population_part[1-2].tar.bz2: 10X population size condition (50 replicates)

Note:
Alignment_SR201_default_diploid_part[1-5].tar.bz2 contain two sequences (phased) per species. The default condition (haploid) can be obtained by extracting every other sequence.

SR21 simulated dataset (FASTA format)

Alignment_SR21_part[1-5].tar.bz2: all condition (50 replicates)

Note:
Alignment_SR21_part[1-5].tar.bz2 contain 20 sequences per species. Conditions with 1/2/5 individuals per species can be obtained by subsampling every 20/10/4 sequences.

S101 dataset

We reused S101 dataset from ASTRALIII (ASTRAL-III: polynomial time species tree
reconstruction from partially resolved gene trees). You can access the S101 dataset (S101.tar.gz) here.

Placental mammal dataset

See: https://cglgenomics.ucsc.edu/data/cactus/

Avian dataset

Also see: https://cglgenomics.ucsc.edu/data/cactus/

82 mammal dataset

mammal82.zip

See: https://datadryad.org/stash/dataset/doi:10.5061/dryad.bp462

Ratite dataset

See: https://datadryad.org/stash/dataset/doi:10.5061/dryad.v72d325

Ruminant dataset

See: https://datadryad.org/stash/dataset/doi:10.5061/dryad.52213gc

Oakleaf butterfly dataset

Alignment_butterfly.bz2

Birch dataset

Alignment_birch.tar.bz2

Lizard dataset

Alignment_lizard.phylip.bz2

Gene trees

gene_trees.zip: Gene trees for wASTRAL inferences (Newick format)

Folder structure

• SR201_default_condition: SR201 dataset default condition (50 replicates)
• SR201_0.1X_mutation: SR201 dataset 0.1X mutation rate condition (50 replicates)
• SR201_10X_mutation: SR201 dataset 10X mutation rate condition (20 replicates)
• SR201_10X_population: SR201 dataset 10X population size condition (50 replicates)
• SR21_1_ind: SR21 dataset 1 individual (50 replicates)
• SR21_5_ind: SR21 dataset 5 individual (50 replicates)

Species trees

• species_trees.zip: Species tree files and logs
• subsample.zip: Species tree files and logs for subsample experiment
• score_true.zip: Comparision of CASTER scores between true and reconstructed phylogenies

Folder structure (species_trees.zip)

• biological: CASTER inferred trees on biological datasets
• truth: True species tree for SR201 and SR21 datasets
• estimated/SR201_default_condition: Estimated species trees for SR201 dataset default condition (50 replicates)
• estimated/SR201_unphased_diploid: Estimated species trees for SR201 dataset unphased diploid condition (50 replicates)
• estimated/SR201_0.1X_mutation: Estimated species trees for SR201 dataset 0.1X mutation rate condition (50 replicates)
• estimated/SR201_10X_mutation: Estimated species trees for SR201 dataset 10X mutation rate condition (20 replicates)
• estimated/SR201_10X_population: Estimated species trees for SR201 dataset 10X population size condition (50 replicates)
• estimated/SR21_1_ind: Estimated species trees for SR21 dataset 1 individual (50 replicates)
• estimated/SR21_2_ind: Estimated species trees for SR21 dataset 2 individual (50 replicates)
• estimated/SR21_5_ind: Estimated species trees for SR21 dataset 5 individual (50 replicates)
• estimated/SR21_20_ind: Estimated species trees for SR21 dataset 20 individual (50 replicates)
• estimated/scale: Estimated species trees for scale study dataset

Simulated files (estimated/*)

• *.caster-site files: CASTER-site species trees
• *.caster-pair files: CASTER-pair species trees
• *.raxml.bestTree files: RaXML species trees
• *.svdquartets files: SVDQuartets species trees
• *.gtrees.wastral files: wASTRAL-hybrid species trees
• *.log: log files

Biological files (biological/*)

• mammal/whole_genome.site: Mammalian whole-genome CASTER-site tree
• mammal/chr/chr*.site: Mammalian per-chromosome CASTER-site trees
• ruminant/Ruminant.WGA.phy.caster-site: Ruminant whole-genome CASTER-site tree
• ruminant/Ruminant.WGA.phy.caster-pair: Ruminant whole-genome CASTER-pair tree
• butterfly/38indiv_4Dec.recode_scaffold_1to32.phylip.caster-site: Oakleaf butterfly SNP tree
• birch/all.ploidy.phylip.caster-pair: CASTER-pair tree for all 47 birch species
• birch/20_diploid.phylip.caster-site: CASTER-site tree for 20 diploid birch species
• birch/20_diploid.phylip.caster-pair: CASTER-pair tree for 20 diploid birch species
• lizard/all.phylip.na.caster-{site,pair}: CASTER-site/pair tree for lizard nucleotide alignments with 3-rd codon positions
• lizard/all.phylip.na12.caster-{site,pair}: CASTER-site/pair tree for lizard nucleotide alignments without 3-rd codon positions
• lizard/all.phylip.aa.recode.caster-pair: CASTER-pair tree for lizard amino-acid alignments
• *.log: log files

Code

CASTER software

Users can find up-to-date version of CASTER here.

For reproducibility of our results, we also upload CASTER v1.15.0.0 as “ASTER.zip” file.

Sliding window code

This is the code we used in sliding-window analysis. Compile and run it without input parameters to see help messages.

caster-site_sliding_window.cpp

Proof checks

CASTER-Proof.nb is a Mathmatica script to verify the correctness of Proposition 2.

Simulation

Scripts to simulate alignments in SR201 and SR21 datasets:

• simulate_SR201_10X_population.py
• simulate_SR201_most_conditions.py
• simulate_scale.py
• simulate_SR21.py

Analyzing sliding window outputs

• stat_mammalian_sliding_window.py: script for mammalian sliding windows
• stat_SR201_sliding_window.py: script for SR201 sliding windows

GitHub repository for data and scripts used for plotting figures

See: https://github.com/chaoszhang/CASTER-data

• CASTER-data-main.zip: A permanent archieve of the GitHub repository

Change log

October 2024

Uploading gene trees

• gene_trees.zip: Annotated gene tree files for wASTRAL and ASTRAL

Combining tables and scripts into a Github repository

• CASTER-data-main.zip: Combined tables and scripts for plotting figures
• Deleting individual table and script files

August 2024

Additional biological dataset

• mammal82.zip: 2nd mammal dataset by Liu et al

Additional simulation experiments

• quartet.zip: New quartet (4-species) dataset
• subsample.zip: Additional subsampling experiment complementary to SR201 dataset
• score_true.zip: Additional experiment comparing CASTER scores of reconstructed and true phylogenies complementary to SR201 dataset

Obsolete experiments removed

• bpp_mcmc.zip: BPP experiment on SR21 dataset was removed due to convergence issue
• snapper_mcmc.zip: SNAPPER experiment on SR21 dataset was removed due to convergence issue