compute_null_dnp: function to compute null distribution of dnp

In DavisWeaver/crosstalkr: Analysis of Graph-Structured Data with a Focus on Protein-Protein Interaction Networks

function to compute null distribution of dnp

Description

compute_null_dnp calculates a null distribution for the change in network potential for for each node in a cell signaling network.

Usage

compute_null_dnp(
  cache = NULL,
  df,
  ppi = "biogrid",
  n,
  n_genes = 50,
  experiment_name,
  ncores = 4,
  min_score = NULL
)

Arguments

cache	user-provided filepath for where to store data etc
df	output of compute_dnp()
ppi	should we use biogrid or stringdb for the PPI
n	number of permutations
n_genes	integer describing number of genes per sample that we will compute the null distribution for
experiment_name	name of the experiment for saving output.
ncores	number of cores to use for calculations
min_score	if ppi is stringdb, which mininum score should we use to filter edges?

Details

The input for this function will be the output of compute_dnp(). To compute the null distribution, the nodes in the provided cell signaling network will be randomly permuted n times, with dnp computed or each new cell signaling network. The mean and standard error of dnp for this set of random networks will constitute the null model that we will use for comparison. Be warned that this operation is extremely expensive computationally. It is recommended to either use a high-performance cluster or limit the computation of the null distribution to a small number of nodes. To distribute the workload over multiple cores, just specify ncores.

Value

df, also saves to cache if specified

See Also

compute_dnp() and compute_np()

DavisWeaver/crosstalkr documentation built on May 6, 2024, 6:09 p.m.