Skip to contents

Simulate host and symbiont trees with known co-diversification

Source: R/simulate_codiv_data.R
simulate_codiv_data.Rd

Builds a host tree and a symbiont tree assembled from several clades, each with known ground truth, all mapping to the same set of hosts. Use it for a fixed sanity-check or to generate many random datasets to study how tree features affect the scan.

Usage

simulate_codiv_data(
  n_hosts = 12,
  clades = NULL,
  n_clades = NULL,
  prop_codiversifying = 0.5,
  congruence_range = c(0.5, 1),
  n_per_host_range = c(1, 4),
  bl_noise = 0.1,
  host_tree = NULL,
  seed = NULL
)

Arguments

n_hosts

Number of host tips (ignored if host_tree is supplied)

clades

Optional list of clade specifications, each a list with codiversifying (logical) and optionally congruence and n_per_host. If NULL, n_clades random clades are generated.

n_clades

Number of random clades to generate when clades is NULL

prop_codiversifying

Probability a random clade is co-diversifying

congruence_range

Range to draw random co-diversifying congruence from

n_per_host_range

Integer range to draw random n_per_host from

bl_noise

Multiplicative branch-length jitter (0 = none)

host_tree

Optional host tree (phylo); if NULL a random coalescent tree of n_hosts tips is used

seed

Optional random seed

Value

A list with:

  • host_tree: the host tree (phylo)

  • symbiont_tree: the combined symbiont tree (phylo)

  • links: data frame of Host-Symbiont associations

  • truth: per-symbiont ground truth (clade, codiversifying, congruence, n_per_host)

Details

Each clade is one monophyletic block of the symbiont tree, controlled by:

  • codiversifying: whether its topology tracks the host tree

  • congruence: for co-diversifying clades, how closely it tracks the host (1 = identical branching order; lower = more scrambled)

  • n_per_host: symbionts sampled per host (1 = no within-host pseudoreplication; > 1 introduces it)

Examples

# fixed four-clade sanity check
sim <- simulate_codiv_data(
  n_hosts = 10,
  clades = list(
    list(codiversifying = TRUE,  congruence = 1.0, n_per_host = 1),
    list(codiversifying = TRUE,  congruence = 0.6, n_per_host = 4),
    list(codiversifying = FALSE, n_per_host = 1),
    list(codiversifying = FALSE, n_per_host = 4)
  ),
  seed = 1
)
sim$host_tree
#> 
#> Phylogenetic tree with 10 tips and 9 internal nodes.
#> 
#> Tip labels:
#>   H1, H2, H3, H4, H5, H6, ...
#> 
#> Rooted; includes branch length(s).
head(sim$links)
#>   Host Symbiont
#> 1   H1    c1_h1
#> 2   H2    c1_h2
#> 3   H3    c1_h3
#> 4   H4    c1_h4
#> 5   H5    c1_h5
#> 6   H6    c1_h6

# many random datasets (vary seed) to study tree-feature effects
rand <- simulate_codiv_data(n_hosts = 15, n_clades = 6, seed = 42)