Skip to contents

Topology-congruence test with permutation significance

Source: R/topology_wf.R
topology_wf.Rd

Tests whether a symbiont tree's branching order matches a host tree's, ignoring branch lengths. Each host is reduced to one representative tip (its largest single-host clade, tie-broken by smallest total clade branch length); hosts whose symbionts are entirely scattered have no defined position and are excluded. Both trees are then pruned to the shared hosts and compared with normalized mutual clustering information (1 = identical branching order). The null distribution is built by shuffling the host labels.

Usage

topology_wf(
  host_tree,
  symbiont_subtree,
  host_to_symbiont_df,
  permutations,
  seed
)

Arguments

host_tree

A phylo object; the host (sub)tree

symbiont_subtree

A phylo object; the symbiont (sub)tree

host_to_symbiont_df

A data frame with "Host" and "Symbiont" columns

permutations

Number of permutations for the null; at least 99 recommended

seed

Random seed for reproducibility; if NA, no seed is set

Value

A list with:

  • stat: normalized mutual clustering information (0-1; 1 = identical branching order), or NA when fewer than 3 hosts have a defined position

  • pvalue: one-tailed permutation p-value

  • hosts_used: number of hosts with a defined position

  • fraction_kept: fraction of symbiont tips represented after reducing each host to its dominant clade (low values flag heavy pruning)

Details

This complements the distance-based methods (Hommola, PACo, ParaFit): it is insensitive to branch-length differences, so it can detect congruent topology even when evolutionary rates differ between the trees.

Examples

# \donttest{
h <- ape::rcoal(8)
s <- h; s$tip.label <- paste0("s", seq_along(s$tip.label))
hs <- data.frame(Host = h$tip.label, Symbiont = s$tip.label)
topology_wf(h, s, hs, permutations = 99, seed = 1)
#> $stat
#> [1] 1
#> 
#> $pvalue
#> [1] 0.01
#> 
#> $hosts_used
#> [1] 8
#> 
#> $fraction_kept
#> [1] 1
#> 
# }