In development.
This package is initially developed as a simplified implementation of Tracer v1.7 in R. It can take an advantage of ggplot2 to create high quality images, and also make batch processing of multiple logs easier.
Through the development, it is growing to a package not only used as
a simple version of Tracer,
but also aiming to provide more post-analysis methods to understand
the posterior distributions from the Bayesian phylogenetic inference using MCMC.
In addition, coverage tests for a model validation are recently added.
- For example,
readState
can extract the summary of operator proposals from BEAST 2*.state
log file.
readState("data/star.beast.state")
- Summarising tree statistics from the posterior trees logged by BEAST 2, such as the total branch length, tree height, etc.
tre.sta.df <- readTrees("data/RSV2long.trees")
tre.sta <- analyseTreeStats(tre.sta.df)
- The 5-step pipeline to summarise the coverage tests for validating Bayesian phylogenetic models.
- For Tracer, please cite:
Rambaut A, Drummond AJ, Xie D, Baele G and Suchard MA (2018),
Posterior summarisation in Bayesian phylogenetics using Tracer 1.7,
Systematic Biology. syy032.
doi:10.1093/sysbio/syy032
- For the coverage-test pipeline, please cite:
Drummond AJ, Xie D, Mendes F (), LinguaPhylo: a probabilistic model specification language for reproducible phylogenetic analyses, in preparation.
You can use the devtools install_github() function to install the latest development version directly from the GitHub.
library("devtools")
remove.packages("TraceR")
devtools::install_github("walterxie/TraceR")
library("TraceR")
Or install from TraceR_*.tar.gz
.
setwd("~/WorkSpace/TraceR")
install.packages("TraceR_?.?.?.tar.gz", repos = NULL, type = "source")
library("TraceR")
To see all exported functions:
help(package = "TraceR")
Try the following commands:
# read MCMC log
mcmc.log <- readMCMCLog("data/star.beast.log")
# get traces and remove burn in
traces <- getTraces(mcmc.log, burn.in=0.1)
# get stats
stats <- analyseTraces(traces)
Please note: you can use ggplot2 instead of gg1L
package for visualisation.
gtTraces(...)
is the only function depending on gg1L,
and the gg1L code will be removed from the future dependecy,
after it is replaced by the ggplot2 code.
devtools::install_github("walterxie/gg1L")
library("gg1L")