Network diffusion class

This class helps to do diffusion of seed nodes in a single or tied version over a network, measure the ability of these diffusion vectors to recall against a gold standard set of nodes and do community detection within the subset of the network that received the the most heat from the initial seed genes.

Usage

NetworkDiffusions(edge_data_frame, weighted, directed)

Arguments

edge_data_frame

data.table that contains the edge information. It is expected to have the columns 'from' and 'to'.

weighted

Boolean. Is the graph weighted. If set to TRUE, the edge_data_frame needs to have a weight column.

directed

Boolean. Shall the graph be stored as directed.

Value

Returns the NetworkDiffusions class for further operations.

Properties

graph

igraph. The underlying graph.

diffusion_res

Numeric vector. Contains contains the single or tied diffusion results.

params

A (nested) list that will store all the parameters of the applied function.

final_results

data.table. Contains final results.

Examples

# unweighted, undirected diffusion class from an edge table
set.seed(42)
g <- igraph::sample_pa(15, directed = FALSE)
edges <- data.table::setDT(igraph::as_data_frame(g))[, `:=`(
  from = sprintf("node_%i", from),
  to = sprintf("node_%i", to)
)]
object <- NetworkDiffusions(edges, weighted = FALSE, directed = FALSE)
class(object)[1]
#> [1] "bixverse::NetworkDiffusions"