Extracts a subnetwork from a decoupleRnival result and a prior knowledge network, enforcing score thresholds, neighbourhood distance, and structural constraints: - Pure children (no outgoing edges) must be level 0 in decoupleRnival_res. - Pure parents (no incoming edges) must be among the provided upstream_input nodes. - All nodes must have |score| > cutoff.
Usage
reduce_solution_network(
decoupleRnival_res,
meta_network,
cutoff,
upstream_input,
RNA_input = NULL,
n_steps = 10
)Arguments
- decoupleRnival_res
A data.frame with columns `source`, `score`, and `level` from decoupleRnival().
- meta_network
A data.frame with columns `source`, `target`, `interaction` for signed directed edges.
- cutoff
Numeric. Absolute score threshold for node filtering.
- upstream_input
A named numeric vector of upstream seed nodes and their activity values.
- RNA_input
Optional named numeric vector of differential expression values; merged into ATT.
- n_steps
Integer. Maximum number of steps away from any upstream_input node.
Value
A list with: - SIF: data.frame of filtered edges (`source`, `target`, `interaction`, `consistency`). - ATT: data.frame of node attributes (`nodes`, `score`, `level`, `RNA_input`).
Examples
# Example input data
upstream_input <- c("A" = 1, "B" = -1, "C" = 0.5)
downstream_input <- c("D" = 2, "E" = -1.5)
meta_network <- data.frame(
source = c("A", "A", "B", "C", "C", "D", "E"),
target = c("B", "D", "D", "E", "D", "B", "A"),
interaction = c(-1, 1, -1, 1, -1, -1, 1)
)
RNA_input <- c("A" = 1, "B" = -1, "C" = 5, "D" = 0.7, "E" = -0.3)
# Run decoupleRnival to generate scores
dec_res <- moon(upstream_input, downstream_input, meta_network,
n_layers = 2, n_perm = 100)
#> Warning: NaNs produced
# Extract solution network
sol_net <- reduce_solution_network(dec_res, meta_network,
cutoff = 0.4, upstream_input,
RNA_input, n_steps = 3)
#> Warning: No upstream_input nodes found in network; returning empty network.
# View outputs
print(sol_net$SIF)
#> data frame with 0 columns and 0 rows
print(sol_net$ATT)
#> data frame with 0 columns and 0 rows