PalantiR
PalantiR simulates and visualizes complete substitution
histories — every substitution event, with its time, codon, and fitness
context — under time-heterogeneous mutation–selection codon substitution
models and other advanced models, including site-specific fitness
profiles, epistasis between sites, and changes in effective population
size or fitness along a phylogeny.
For installation notes, see this page.
Details on particular models and simulation types can be found on the following pages:
Nucleotide Models
Codon Models
Temporal Heterogeneity
What’s new in 1.2
Following a full source review: reproducible simulations via
set_palantir_seed(); an exact time-change alternative to
the segment rescaler (rescale_method = "exact"); corrected
CoEvolution and Markov-modulated model semantics; input
validation throughout; and repaired datasets and documentation. See NEWS.md
for the complete list, including breaking changes.
Engine validation
The simulation engine is validated by a rerunnable notebook that tests branch-length calibration under each scaling mode, the stationary distribution, and the transient between two equilibria in a time-heterogeneous simulation, each against an exact expectation with a pre-stated pass criterion. Read the current validation report (all tests pass, version 1.2.1), or rerun it from a clone of the repository:
# in validation/, with PalantiR installed
rmarkdown::render("engine_validation.Rmd", params = list(seed = 1))For a larger event-level demonstration, the GY94 history-validation report
simulates 5,000 codons on a 50-taxon tree, reconstructs Goldman–Yang
dN/dS from all recorded substitutions, and visualizes the
branch-specific histories. The executable Jupyter notebook is retained
in the repository under validation/.
Substitution histories under mutation–selection with epistasis
|
A fitness shift under a time-heterogeneous mutation–selection
model
|