reduce_solution_network_double_thresh

Extracts a subnetwork using a two-tier absolute-score threshold, then restricts to only paths connecting upstream_input seeds to level 0 nodes, with recursive consistency filtering to ensure edge signs match node activities.

Usage

reduce_solution_network_double_thresh(
  decoupleRnival_res,
  meta_network,
  primary_thresh,
  secondary_thresh,
  upstream_input,
  RNA_input = NULL
)

Arguments

decoupleRnival_res

A data.frame with columns `source`, numeric `score`, and integer `level`.

meta_network

A data.frame with columns `source`, `target`, and `interaction`.

primary_thresh

Numeric. Absolute score cutoff for primary node selection.

secondary_thresh

Numeric. Absolute score cutoff for secondary node restriction.

upstream_input

A named numeric vector of upstream seed nodes.

RNA_input

Optional named numeric vector of differential expression values; merged into ATT.

Value

A list with: - SIF: data.frame of filtered edges (`source`,`target`,`interaction`). - ATT: data.frame of node attributes (`source`,`score`,`level`,`type`,`RNA_input`).

Examples

dec_res <- data.frame(
  source = paste0("G", 1:6),
  score  = c(2.5, 1.2, 0.8, -2.2, 1.5, -0.5),
  level  = c(0, 0, 1, 0, 1, 1)
)
meta_net <- data.frame(
  source      = c("G1","G1","G2","G3","G4","G5"),
  target      = c("G2","G3","G4","G5","G2","G6"),
  interaction = c(1, -1, 1, 1, -1, 1)
)
upstream_input <- c(G1 = 1, G4 = -1)
RNA_input <- c(G1 = 0.5, G2 = -0.2, G4 = 1.1)
dbl_net <- reduce_solution_network_double_thresh(
  decoupleRnival_res = dec_res,
  meta_network       = meta_net,
  primary_thresh     = 2,
  secondary_thresh   = 1,
  upstream_input     = upstream_input,
  RNA_input          = RNA_input
)
print(dbl_net$SIF)
#> [1] source      target      interaction
#> <0 rows> (or 0-length row.names)
print(dbl_net$ATT)
#> [1] source    score     level     type      RNA_input
#> <0 rows> (or 0-length row.names)