doc/README.md

In richelbilderbeek/razzo: The Error if Nature is MBD

razzo

Folder structure

In razzo_project, there are two folders:

• data: the simulated data
• results: the interpreted data, e.g. tables and figures

The data folder

In data, there are folders named after their parameters, e.g. 0.2-0.15-1-0.1. In each of these folders, there are folders named after their seeds, e.g. 1. In each of these folders, there are folders named after their site and clock models, e.g. strict-jc69.

In each of these folders, there are:

Filename |Description -----------------------------|---------------------------------------------------------------------------------------------------------------------- parameters.RDa |the parameter file mbd.tree |the true MBD tree mbd.fasta |the true MBD alignment mbd_gen.xml |the BEAST2 input file, using the generative model mbd_gen.trees |the posterior trees from the true alignment, using the generative model mbd_gen.log |the posterior parameter estimates from the true tree, using the generative model mbd_gen_evidence.trees |the trees to estimate the marginal likelihood, using the generative model mbd_gen_evidence.log |the parameter estimates to estimate the marginal likelihood, using the generative model mbd_gen.xml.state |the final posterior state from the true tree, using the generative model mbd_best.xml |the BEAST2 input file, using the best candidate model mbd_best.trees |the posterior trees from the true tree, using best candidate model mbd_best.log |the posterior parameter estimates from the true tree, using the best candidate model mbd_best.xml.state |the final posterior state from the true tree, using the best candidate model mbd_best_[x]_evidence.trees|the trees to estimate the marginal likelihood, for the xth candidate model (x == 1 for first model) mbd_best_[x]_evidence.log |the parameter estimates to estimate the marginal likelihood, for the xth candidate model (x == 1 for first model) mbd_marg_lik.csv |the evidences (aka marginal likelihoods) all models have for the true alignment mbd_nltts_gen.csv |the error between true tree and its posterior for the generative model mbd_nltts_best.csv |the error between true tree and its posterior for the best candidate model

Filename |Description -------------------------|--------------------------------------- mbd_twin.tree |the twin MBD tree mbd_twin.fasta |the twin MBD alignment mbd_gen_twin.trees |the posterior trees from the twin alignment, using the generative model mbd_gen_twin.log |the posterior parameter estimates from the twin tree, using the generative model mbd_best_twin.trees |the posterior trees from the twin tree, using best candidate model mbd_best_twin.log |the posterior parameter estimates from the twin tree, using the best candidate model mbd_marg_lik_twin.csv |the evidences (aka marginal likelihoods) all models have for the twin alignment mbd_nltts_gen_twin.csv |the error between twin tree and its posterior for the generative model mbd_nltts_best_twin.csv|the error between twin tree and its posterior for the best candidate model

See a razzo_project build log to see such a data folder.

The results folder

In results, there are only files:

Filename|Description|Created by -------------|---------------------------------------|-------------------------- esses.csv|all ESSes|create_esses_file marg_liks.csv|all marginal likelihoods|create_marg_liks_file nltt_stats.csv|all nLTT statistics, in short form|create_esses_file figure_1.png|figure 1|create_fig_1_file figure_2.png|figure 2|create_fig_2_file

See a razzo_project build log to see such a results folder.

Function overview

• Data creation: produces parameter files and output
• Results creation: creates figures and tables from the data

Data creation

Produces parameter files and output

File handling functions

All functions that end with _file or _files. These functions are friendly to scripted use and require only filenames as arguments.

Status|Description|Function name ---|---|--- ?|Create parameter files|create_parameters_files ?|Create all output files|run_razzo_from_file

Data handling functions

These functions do the actual work. They work on parsed data and are not intended to be called from a script

Status|Description|Function name ---|---|--- ?|1. Create parameters|create_parameters ?|2. Run the experiment|run_razzo

Results creation

Creates figures and tables from the data.

Data handling

All these functions take the project folder's name as an argument and return a tidy data frame or figure. They are not intended to be called from a script

Status|Description|Function name ---|---|--- ?|Collect marginal likelihoods|collect_marg_liks ?|Collect effective sample sizes|collect_esses ?|Collect nLTT statistics|collect_nltt_stats ?|12. Create figure 1|create_fig_1

File handling

All functions that end with _file or _files. These functions are friendly to scripted use and require only the project folder's name as an argument.

Status|Description|Function name ---|---|--- ?|Collect marginal likelihoods|create_marg_liks_file ?|Collect effective sample sizes|create_esses_file ?|Create nLTT statistics file|create_nltt_stats_file ?|Create figure 1|create_fig_1_file

How to create the dependency graph from the .dot file?

dot -Tps dependencies.dot -o dependencies.ps; convert dependencies.ps dependencies.png

richelbilderbeek/razzo documentation built on March 10, 2021, 1:17 p.m.