Important changes to this project will be documented in this file.
We try to follow Keep a Changelog, and we use Semantic Versioning.
- Update
calibrations.R
with functions to calibrate cummulative and quantile functions for joint and upper calibration bounds.
- Add function
mcmc2multiphylo
to generate a list of trees from the MCMC output of MCMCtree or BPP.
- Changed function
write.morpho
to include a scaling constant, so that all the traits are rescaling before writing to file. Useful to rescale the mean rate of the morphological partition.
- Update
DESCRIPTION
file:- Remove
Morpho
andRdpack
fromImport
.
- Remove
- Update
Morpho.R
script:- Check whether users have
Morpho
installed when running functionproc2MCMCtree
, stop if not -> error and warning solved. - Update
.checkData
and.checkArr
functions so that checks withclass()
do not raise warnings -> warning solved.
- Check whether users have
- Update
calibrations.R
file:- Remove additional parenthesis in line 45 (i.e. from
pv <- pL(qv, tL=1))
topv <- pL(qv, tL=1))
) -> error solved.
- Remove additional parenthesis in line 45 (i.e. from
- Update
dBD.R
script:- Define parameter
x
as@param
-> warning solved.
- Define parameter
- Update
mcmc2densitree
:- Change
cex.labels
tocex.lab
(line 18,@param
section) as the latter is used in the function -> warning solved.
- Change
- Update
DESCRIPTION
file:- Upgrade
RoxygenNote
to 7.3.1. - Upgrade minimum R version to
(>= 3.5.0)
.
- Upgrade
- Update
Morpho.R
script:- Add
\dontrun{}
for examples inproc2MCMCtree
that require packageMorpho
in case users do not have this package installed.
- Add
- Update
mcmc2anc.R
script:- Add
\dontrun{}
for examples section to avoid warnings when usingrgl
. - Change lines 45-46 so that lines are not wider than 100 characters as before they were truncated in the PDF manual.
- Add
- Update vignette
Reproduce_Carnivora_analysis.Rmd
:- Add
eval=FALSE
in all code snippets. PackageMorpho
is not imported and running these examples raises errors. - Define code snippets from section 4 onward using
{r}
. - Add
mcm3r::
before calling a function within this package in the code snippets. - Replace some
=
with<-
as it is encouraged in R. - Fix some typos.
- Add
- Update
carnivores.R
script:- Fix typo in surname.
- Function
mcmc.sum
to calculate summary statistics on MCMCtree's output.
- Functions
pL
andqL
to calculate the probability and quantile functions for the minim bound (truncated-Cauchy) calibrations.
- Function
mcmc2anc
for ancestral character reconstruction from an MCMC sample. - Directory
misc/ape4s
, and moved all ape4s files into it.
- The analysis of carnivores vignette,
Reproduce_Carnivora_analysis.Rmd
, to streamline the text and add an example of ancestral reconstruction.
- Function
mcmc2densitree
to plot densitrees using the MCMC output from either MCMCtree or BPP. microcebus
andhominid
BPP A00 MCMC datasets.- Carnivora data to
misc/carnivora
, suitable for reproducing MCMC sampling of divergence times using molecular and morphometric data with MCMCtree.
- The analysis of carnivores vignette,
Reproduce_Carnivora_analysis.Rmd
, to simplify the text and to add MCMC sampling using MCMCtree. Added filesfigtree.jpg
anddensitree.jpg
to illustrate the vignette.
- An error was returned in several instances of testing input where class(x)
was used to check matrices due to a new behaviour of classes in matrices from
R > v4.x. Now
inherits
is used to test matrix classes instead ofclass
.
- A bug in the
dBD
example (missing comment character).
- Function
gauss.quad
so it now accepts ase
argument (as forstepping.stones
). This solves a bug when runninggauss.quad.boot
.
- Function
gauss.quad.boot
to calculate marginal likelihood on bootstrap replicates using the Gaussian quadrature method. Documentation forgauss.quad.boot
andstepping.stones.boot
are merged.
- Function
dBL
to calculate the kernel density for the birth-death process with species sampling.
- Functions
dL
,dB
anddU
to calcute the fossil calibration densites used in MCMCtree (useful for plotting the densities).
- An error in
block.boot
that caused generation of empty likelihood block samples.
- Functions
block.boot
andstepping.stones.boot
to perform block (stationary) bootstrap resampling to estimate the standard error for the log-marginal likelihood under the stepping stones method. Suitable for short MCMC runs for which the delta approximation to estimate the standard error may not work well.
- Functions
stepping.stones
andgauss.quad
so that they work correctly with incomplete mcmctree MCMC output files.
- Function
bayes.factors
so that bootstrap CI's are printed with models as rows and 95% CI as columns.
A collection of functions for working with continuous morphological data. These are the functions used in Álvarez-Carretero et al. (2018, bioRxiv, 441105).
- Function
bayes.factors
so that it now calculates bootstrap CIs for posterior probabilities. Prior probabilities can now be used.
Tags: Added, Changed, Deprecated, Removed, Fixed, Security