Skip to contents

Differential centrality between two observed networks

Source: R/diff_centrality.R
diff_centrality.Rd

Compute per-node, per-statistic differences in centrality between two observed networks (reported as group2 - group1) and assess significance using a null generated by null_ggm() via either permutation or bootstrap.

Usage

diff_centrality(
  obs_networks,
  dat,
  group_col = "phenotype",
  null_networks = NULL,
  alternative = c("two.sided", "greater", "less"),
  inference_method = c("D-S_NW_SL", "B_NW_SL"),
  shuffle_method = c("permutation", "bootstrap"),
  shuffle_iter = 100,
  balanced = FALSE,
  filter = c("pval", "fdr", "none"),
  threshold = 0.05,
  n_cores = NULL,
  seed = NULL
)

Arguments

obs_networks

A *named* list of length 2 containing two igraph objects. The names (for example c("healthy", "disease")) must match values present in dat[[group_col]].

dat

A data frame passed to null_ggm() to construct the null.

group_col

Character scalar; name of the column in dat indicating group labels. Default is "phenotype".

null_networks

A list of null networks from null_ggm().

alternative

Character; one of "two.sided", "greater", or "less" defining the alternative hypothesis.

inference_method

Character; forwarded to null_ggm() (for example "D-S_NW_SL" or "B_NW_SL").

shuffle_method

Character; one of "permutation" or "bootstrap" specifying how the null is generated by null_ggm().

shuffle_iter

Integer; number of null replicates to generate. Default is 100.

balanced

Logical; forwarded to null_ggm().

filter

Character; one of "pval", "fdr", or "none", forwarded to null_ggm().

threshold

Numeric; forwarded to null_ggm().

n_cores

Integer or NULL; forwarded to null_ggm().

seed

Optional seed; forwarded to null_ggm().

Value

A data.frame with rows corresponding to nodes (row names taken from the centrality matrices) and columns interleaved by statistic:

  • If shuffle_method == "permutation": \(p x (2k)\) with columns <stat>_diff, <stat>_pval.

  • If shuffle_method == "bootstrap": \(p x (3k)\) with columns <stat>_diff, <stat>_pval, <stat>_95pct_CI. (The CI level is fixed at 95 percent in the current implementation.)