Skip to contents

Molecular-clock corroboration of co-diversification

Source: R/molecular_clock.R
molecular_clock.Rd

For the most strongly co-diversifying clades, regresses symbiont divergence against host clade age. A positive relationship corroborates that symbiont clades diversified contemporaneously with their hosts, and the slope is an estimate of the symbiont molecular rate (substitutions per unit host time). This follows the approach of Sanders et al. (2023) and Sprockett et al. (2025).

Usage

molecular_clock(
  codiv_results,
  host_time_tree,
  statistic = NULL,
  stat_threshold = 0.95,
  pvalue = NULL,
  p_threshold = 0.01,
  symbiont_divergence = NULL
)

Arguments

codiv_results

A codiv result (or data frame) from codiv()

host_time_tree

A time-calibrated host tree (phylo, ultrametric) whose branch lengths are in time units (e.g. millions of years)

statistic

Name of the test-statistic column. NULL (default) uses Hommola_r, the collapsed Hommola correlation that codiv() always writes when the Hommola method runs

stat_threshold

Only clades with a statistic above this value are used; default 0.95 (strong co-diversification)

pvalue

Name of the p-value column. NULL (default) resolves alongside statistic

p_threshold

Only clades with a p-value below this value are used; default 0.01

symbiont_divergence

Optional override for the symbiont-divergence summary. Either a named numeric vector keyed by Node_ID, or a function taking a symbiont subtree (phylo) and returning a single number. If NULL (default), the clade's crown depth is used (mean root-to-tip distance in the symbiont tree's branch-length units)

Value

A list with:

  • model: the fitted lm of symbiont divergence on host age

  • slope, intercept, r_squared, p_value: regression summaries

  • slope_ci: 95% confidence interval for the slope

  • per_clade: data frame with one row per clade used (Node_ID, n_hosts, host_age, symbiont_divergence)

Details

Host clade age is the crown age of the clade's hosts (the age of their most recent common ancestor in host_time_tree). The default symbiont divergence is a simplification: it is the clade crown depth in substitution units, not a dedicated absolute date. For rigorous absolute dating supply your own estimates via symbiont_divergence. Nested clades are non-independent; the regression treats each passing node as a point, so consider restricting to non-nested clades for formal inference.

Examples

# \donttest{
sim <- simulate_codiv_data(n_hosts = 14, clades = list(
  list(codiversifying = TRUE, congruence = 1, n_per_host = 3)), seed = 1)
res <- codiv(sim$host_tree, sim$symbiont_tree, sim$links, span_fraction = 0.9,
             methods = "hommola", permutations = 199, verbose = FALSE)
#> Warning: One or both trees are not fully bifurcating; resolve polytomies with ape::multi2di() for correct tree-shape metrics
# sim$host_tree is ultrametric, so it stands in for a time-calibrated tree
mc <- molecular_clock(res, sim$host_tree, stat_threshold = 0.75)
mc$slope
#> [1] 1.010627
mc$per_clade
#>   Node_ID n_hosts  host_age symbiont_divergence
#> 1 Node_29       5 0.1622165           0.1642994
#> 2  Node_4       6 0.4981807           0.5164868
#> 3  Node_3       9 0.7521906           0.7612640
#> 4  Node_2      14 1.9897942           2.0152759
# }