This vignette shows some examples how to set up different inference
models in babette
.
For all examples, do load babette
:
All these examples check that BEAST2
is installed at the
default location at /github/home/.local/share/beast/lib/launcher.jar. If
this is not the case, we will use some fabricated output:
posterior <- create_test_bbt_run_output()
posterior$anthus_aco_sub_trees <- posterior$anthus_aco_trees
names(posterior)
#> [1] "estimates" "anthus_aco_trees" "operators"
#> [4] "output" "anthus_aco_sub_trees"
All examples read the alignment from a FASTA file (usually
my_fasta.fas
).
Instead of a full run, the MCMC chain length is shortened to 10K states, with a measurement every 1K states:
We will re-create this MCMC setup, as doing so initializes it with new filenames for temporary files. These temporary files should not exist before a run and should exist after a run. Sure, there is the option to overwrite…
Using all default settings, only specify a DNA alignment.
if (is_beast2_installed()) {
inference_model <- create_inference_model(
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
All other parameters are set to their defaults, as in BEAUti.
An alternative is to date the node of the most recent common ancestor of all taxa.
Create the MCMC:
if (is_beast2_installed()) {
inference_model <- create_inference_model(
mcmc = mcmc,
mrca_prior = create_mrca_prior(
taxa_names = sample(get_taxa_names(fasta_filename), size = 3),
alignment_id = get_alignment_id(fasta_filename),
is_monophyletic = TRUE,
mrca_distr = create_normal_distr(
mean = 15.0,
sigma = 0.025
)
)
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
Here we use an MRCA prior with fixed (non-estimated) values of the mean and standard deviation for the common ancestor node’s time.
if (is_beast2_installed()) {
inference_model <- create_inference_model(
site_model = create_jc69_site_model(),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
if (is_beast2_installed()) {
inference_model <- create_inference_model(
clock_model = create_rln_clock_model(),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
if (is_beast2_installed()) {
inference_model <- create_inference_model(
tree_prior = create_bd_tree_prior(),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
if (is_beast2_installed()) {
inference_model <- create_inference_model(
tree_prior = create_yule_tree_prior(
birth_rate_distr = create_normal_distr(
mean = 1.0,
sigma = 0.1
)
),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
Thanks to Yacine Ben Chehida for this use case
if (is_beast2_installed()) {
inference_model <- create_inference_model(
site_model = create_hky_site_model(
gamma_site_model = create_gamma_site_model(prop_invariant = 0.5)
),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
Thanks to Yacine Ben Chehida for this use case
if (is_beast2_installed()) {
inference_model <- create_inference_model(
clock_model = create_strict_clock_model(
clock_rate_param = 0.5
),
mcmc = mcmc
)
beast2_options <- create_beast2_options()
posterior <- bbt_run_from_model(
fasta_filename = fasta_filename,
inference_model = inference_model,
beast2_options = beast2_options
)
bbt_delete_temp_files(
inference_model = inference_model,
beast2_options = beast2_options
)
}
Thanks to Paul van Els and Yacine Ben Chehida for this use case.