Skip to contents

cod: Inference of Coalescent ODds in Phylogenetic Trees

License: MIT

Overview

The cod R package provides tools for inferring variation in coalescent rates across branches in phylogenetic trees. It provides methods to identify growing lineages, adjust for biased sampling, compute phylogenetic clusters, optimize cluster thresholds, and summarize rate variations.

For worked examples and detailed usage, see package documentation and vigettes at https://emvolz.github.io/cod/.

Installation

You can install cod from GitHub using devtools:

# install.packages("devtools")
devtools::install_github("emvolz/cod")

Dependencies

The package requires several dependencies: - Matrix (>= 1.7) - ape (>= 5.0) - glue (>= 1.0.0) - mgcv (>= 1.9-1) - parallel

Optional packages for visualization: - ggtree (>= 3.12.0) - ggplot2 (>= 3.0.0)

Key Features

Estimating Coalescent Odds

The main function codls() estimates coalescent odds using weighted least squares:

# Load a phylogenetic tree
tr <- ape::read.tree("your_tree_file.nwk")

# Fit the COD GMRF model
f <- codls(tr)

# View summary
summary(f)

# Plot the results (requires ggtree)
plot(f)

Phylogenetic Clustering

Identify clusters in phylogenetic trees based on coalescent odds:

# Find optimal clustering threshold
chis <- chindices(f, clths = seq(0.1, 1.5, length = 20))

# Compute clusters
clusterdf <- computeclusters(f, clth = chis$threshold[which.max(chis$CH)])

# Plot clusters (requires ggtree)
plotclusters(f, clusterdf)

Adjusting for Sampling Bias

Correct for biased sampling using the sample reweighting method:

# Identify tips to reweight
reweight_tips <- c("tip1", "tip2", "tip3")

# Automatically determine optimal reweighting
reweightedfit <- autoreweight(f, reweight_tips)

Documentation

For detailed documentation on functions and their parameters, use the R help system and vignettes:

?codls
?computeclusters
?autoreweight

Citation

A manuscript describing the methods in this package is currently in preparation.

License

This package is available under the MIT License. See the LICENSE file for more information.