Statistical test

1. simulate data using simulate_all_data.sh

Use simulation/simulate_all_data.sh. Need to specify your own path to ms and seqgen in simulation/model.py The simulation process is described below:

1. generate homogeneous gene trees and markers
2. generate heterogeneous gene trees and markers
3. infer gene trees using iqtree and obtain estimated gene tree errors
4. randomly resolve short branches

2. Infer species tree for full network PhyloNet

Use command Infernetwork_ML in Phylonet to infer network prepare_ML_nex.sh, run_ML.sh are two scripts may help

3. Analysis

Use summarize_result.sh

3.1 compute expected gene tree probabilities

You can use CalGTProb Command in PhyloNet for each gene tree topology

3.2 compute test statistics on full network

1. write gt probs to a csv: summarize_probs.py
2. compute test statistics- approximate: summarize_stats_5taxa.py
3. plot figures with summarize_plot.py

3.3 triplets multitest

1. summarize_triplet_probs.py
2. summarize_triplet_stats.py
3. summarize_triplet_multitest.py
4. plot figures with summarize_triplet_plot.py

4. Comparison

D3

The scripts to run D3 are: D3_test/run_D3.sh and D3_test/run_D3_gt.sh

QuartetGoodnessFit

1. Use run_quartet_probs.sh to prepare input files.
2. Put mydata.jl and run_mydata.sh in to directory of QuartetGoodnessFit and run
3. summarize results with summarize_pval.py

Supporting locus-specific substitution rates in MCMC_SEQ

This is implemented in the code base of PhyloNet.

Download: https://github.com/NakhlehLab/PhyloNet/releases

Nexus line:

MCMC_SEQ -cl 20000000 -bl 10000000 -sf 5000 -sd 12345 -pl 24 -tm <C:C_0;L:L_0;R:R_0;Q:Q_0>; 
MCMC_SEQ -cl 20000000 -bl 10000000 -sf 5000 -sd 12345 -pl 24 -tm <C:C_0;L:L_0;R:R_0;Q:Q_0> -murate; 

The -murate switch is to enable the sampling of varying substitution rates.