accessoryglands

Code and data for analyzing the evolution of reproductive accessory glands across fishes. Specifically, we test whether the evolution of accessory glands is driven by other reproductive traits, namely parental care and spawning mode.

running the project

getting started

A recent version of R (i.e. 4.2.0 or higher) and updating your packages is strongly recommended (update.packages()). You will probably need to have compilation tools installed on your system, for installing R packages from source.

All the code should run, but compiling everything (including PDF version of the supplementary material) needs a few additional tools: a working LaTeX installation, the optipng program, ...

if (!require("remotes")) install.packages("remotes")
detach("package:remotes")
source("R/pkg_install.R")
install_pkgs()

running the workflow

These instructions will run all the code and build the supplementary material.

within R

To run single-threaded:

library(targets)
tar_make()

Or multi-threaded:

library(targets)
library(future)
plan("multicore")
tar_make_future()

via make

• make all if you have make installed.

To load existing products (e.g. for debugging):

targets::tar_load(everything())

To visualize the workflow:

targets::tar_visnetwork()

figures

• figure 1 (flow diagram): mk_flowfig.R (needs LaTeX, imagemagick for PDF to PNG conversion)
• figure 2 (phylogeny): used diversitree package, phytools for stochastic mapping; overlaid in PowerPoint
• figure 3: run figures.R

workflow

• implemented in the targets package
• for different cases (data, settings, etc.), we need to be able to:
  • fit a basic corHMM model
  • possibly: compute Wald and/or likelihood profile confidence intervals
  • specify priors and do MCMC computation
  • plot/interpret results

contents

documents

• ag_bayesdiag.rmd: Bayes diagnostics for MCMC fits (link)
• ag_supp.rmd: supplementary material (link)
• ag_tech.rmd: tech overview document (????)
• ag_model.rmd: mostly obsolete
• corHMM.bib: bibliography for paper/supp

input files

• contr*.csv: definitions of contrast matrices for different cases. Use source("R/utils.R"); tar_load(<R_object>); iplot(<R_object>) to view them.
  • contr_full.csv: contrast matrix for full (24-parameter) model (R object contrast_mat_full)
  • contr.csv: 12-parameter model, including contrasts for net gain of AGs (gain - loss) (R object contrast_mat_0)
  • contr_invertible.csv: 12-parameter model with gain, loss, and including simple (1-to-1) contrasts for transitions in spawning mode and male care

data files (data/ directory)

• accessTree.R: extract phylogenies via fishtree package, given data
• cleanTraitData.csv: trait data in full format
• binaryTraitData.csv: trait data collapsed to binary (0/1) for traits
• treeBlock.rds: tree block (100 imputed trees)
• treeSingle.rds: phylogeny only for complete genetic data
• dataReferencecs.csv: list of references for each species
• referenceList.rtf: full reference list corresponding to the above csv

other folders

• bayestraits: explorations of BayesTraits
  • test_btw.R: preliminary tests of btw package (BayesTraits R interface)
  • BTrefs.xls: papers using BayesTraits with some added information about what priors were used in the papers, how inference was done, etc..
• old: Miscellaneous old explorations etc.
• twotraits: code for two-trait analyses (comparisons with LE's thesis)

google drive link

BayesTraits review

BTrefs.xls has all articles from 2020-2022 referencing Pagel & Meade (99 articles)

installing BayesTraits

Linux (from source, V3)

git clone ://github.com/bbolker/BayesTraitsV3
cd BayesTraitsV3/src
## sudo apt install libnlopt-dev
make threaded
mv BayesTraitsv3_threaded ../../BayesTraitsV3

from binary:

download.file("http://www.evolution.reading.ac.uk/BayesTraitsV4.0.0/Files/BayesTraitsV4.0.0-Linux.tar.gz", destfile = "btv4.tgz")
untar("btv4.tgz")
unlink("btv4.tgz")

test:

library("btw")
bayestraits(data = primate.discrete1,
	tree = primate.tree100,
	command = c("1", "1", "run"))

git clone git@github.com:bbolker/btw.git